Description
To access a financial institution's online banking facility, a customer having personal Internet access must register with the institution for the service, and set up some password (under various names) for customer verification.
Impact of Web Banking Usability on Community Bank Performance: A Heuristic Evaluation of Community Bank Website Homepages
Srinivasa Rao Lingam Master of Computing Studies Division of Computing Studies Arizona State University [email protected]
Abstract—This research study attempts to study the impact of usability of community banks’ Web banking efforts on their performance through heuristic evaluation of their Web site homepages. The homepages were evaluated and a measure of their usability is developed which we term usability index. This usability index is regressed against bank performance indicators along with other explanatory variables. Based on the sample, the results show that (1) the asset size of banks has no impact on the usability of the homepages of Web sites, and (2) the impact of usability of bank Web site homepages on bank performance is insignificant as a result of which no meaningful and reliable conclusions can be drawn. It may be that Web usability in community banking sector is not mature enough to have significant impact on performance. However, on cross examining the data it is found that more efficient banks tend to maintain quality Web sites that have higher usability index though higher usability index is not leading to higher efficiency. Key Words—Community Banks, Web Banking, Web Usability, Heuristic Evaluation, Profit Efficiency.
community banks’ implementing Web banking as a critical success factor in the banks’ business model and revealed that Web banking activities have significant positive impact on the banks’ return on assets (ROA) and return on equity (ROE) and improves the asset quality by identifying non performing loans. Further, increasing Information Technology (IT) costs reduce ROA and ROE but improve the loan quality. This research study is a sequel to our previous work [3] – [6] and attempts to study the impact of usability (user experience derived from the usability) of community banks’ Web banking efforts, particularly homepages, on their performance through heuristic evaluation of their Web site homepages. A. Statement Of The Problem This project intends to evaluate the impact of the usability of homepages of bank Web sites on their financial performance. To this effect, the problem statement has two parts: (1) Do banks having greater asset size have more usable homepages (or Web sites)? This helps determine if the investment of resources in the usability of Web sites is driven by asset size. (2) Do banks having more usable homepages (or Web sites), as indicated by a higher compliance rate to usability design guidelines, show more profitability or perform better than competing banks with less usable homepages because of the usability of their homepages (or Web sites)? Alternately, does usability have an effect on financial performance of a bank? This helps determine if there is a payoff from the investment of resources in the usability of Web sites and their homepages. B. Research Objectives The specific research objectives of this study are to: · Conduct heuristic evaluation of the Web site homepages of the community banks under study to assess their compliance with the homepage usability design guidelines and use the data to develop a measure that represents the overall usability of the homepages? · Collect relevant financial performance data for the sample community banks from the Consolidated Report of Condition and Income (“Call Reports”) filed by the banks with the Federal Deposit
I. INTRODUCTION Community banks historically had a strategic advantage over large regional and national banks in providing loans to small and mediumsized businesses especially in nonmetro areas. However, recent innovations in the use of communication and information technology, the deregulation of the banking sector, and an increasing use of electronic banking products (Internet banking) have altered the community bankerlocal merchant relationship. These developments have increased the competition from large commercial banks in the small business lending markets, which were mainly served by the community banks [1], [2]. As a result, community banks started implementing electronic commerce (EC) activities, particularly Internet (online or Web) banking applications, as a strategic measure to withstand competition to remain viable market players and gain market share. Community banks started implementing Web sites around 1997 as simple information tools but a majority of them implemented online banking applications on their Web sites during the threeyear period 2001 to 2003. Our previous work [3] – [6] assessed the strategy of the
1
·
Insurance Corporation (FDIC)? and Analyze the bank level performance with appropriate econometric modeling techniques using data pertaining to the usability measure developed in step 1 as an explanatory variable of primary interest, along with other explanatory variables formed from financial performance data collected in step 2, in an attempt to assess the impact of usability of homepages on the specific financial performance measures of banks.
C. Scope The study includes sample community banks insured by FDIC in the upper Midwest and Southwest regions of U.S. covering Arizona, Iowa, Minnesota, Montana, North Dakota, South Dakota and Texas due to the preponderance of traditional unit banks and community banks in these regions. This study covers the usability of homepages alone as evaluated with a limited set of homepage usability design guidelines, but not the entire Web sites. 1) Assumptions: The main assumptions in this research concern the benefits of Web banking activities for community banks. Web banking is assumed to be a critical success factor for community banks in withstanding competition to stay in business. Web banking is assumed to bring new markets and an emerging customer base to the community banks and help retain customers as opposed to the traditional brickand mortar banking model. Above all, to successfully sustain the Web banking activities in the longrun, the Web sites and their homepages must be usable for the masses. 2) Limitations: The use of EC in community banking is continuously evolving and design of the Web sites and Web banking applications and features on the Web sites change from time to time. However, this study takes a snap shot of the usability of the design and content of homepages of community banks at just one point of time. As a result, the usability of homepages of bank Web sites could only be evaluated against the financial performance data for the current observation period, rather than the entire time period of Web banking implementation. II. REVIEW OF LITERATURE EC technology is the new delivery channel in banking designed to reach bank customers efficiently. For community banks, using technology in relationship banking assumes greater importance as these local banks have traditionally based their business model on relationship banking. This section reviews relevant literature and previous research pertaining to the role played by information and communication technology in the evolution of community banking as well as the importance of usability of these EC initiatives (especially Web banking) to reap the real benefits that accrue from their implementation in the longrun.
A. EC Technology In Community Banking My previous work [6] presented the usage and role of EC technology in the history and evolution of community banking over the last three decades in the midst of deregulation in the banking industry, innovations in financial services, technological advances, increased competitive rivalry, changing industry structure, and changing business models and business processes which have profoundly affected the size and health of the U.S. community banking sector and the availability and quality of banking products and services. The banking and nonbanking activities that were implemented using EC, especially Web banking, and the benefits accrued have been discussed. The vital importance, role, and usage of technology in improving efficiency and productivity in the traditional relationship banking model of community banks has been discussed. The issues of privacy, transaction and information security, customer satisfaction, loyalty, and trust in the online relationship banking environment have been discussed. B. Impact Of Web Banking Activities Except for the de novo (i.e., new) banks, banks in all size categories offering Internet banking tend to rely less on traditional banking (interestyielding activities and deposits) than do nonInternet banks, and outperform nonInternet banks in terms of profitability. Internet de novo banks tend to be less efficient and less profitable than nonInternet de novo banks. For most banks, the low percentage of customers using Internet banking and the relatively modest cost of setting up an Internet banking Web site makes it unlikely that Internet banking is having a noticeable impact on firm profits. However, the largest and smallest banks may be exceptions to this. Large commercial banks have a disproportionately sizeable share of Internet banking usage. For small banks with less than $100 million assets and de novo banks, Internet banking could be a primary reason for their lack of profitability if they rely more on an Internetbased strategy as the costs of offering Internet banking may be significant, if not prohibitive, for these banks [7]. Though bank profitability is strongly correlated with Internet banking, the impact of Internet banking on bank profitability is not statistically significant. Apparently banks that adopted Internet banking already had higher profitability, accounting efficiency, and scale than other banks [8] – [10]. Our previous work assessed the strategy of the community banks’ implementing Web banking as a critical success factor in the banks’ business model and revealed that Web banking activities have significant positive impact on the banks’ ROA and ROE and improves the asset quality by identifying non performing loans. Further, increasing IT costs reduce ROA and ROE but improve the loan quality. Their results show that community banks that provide extensive online banking services tend to perform better (approximately 3 percent more profit efficient) than competitors who lag behind. These
2
results indicate that online banking is an important strategic option for competitive positioning of community banks. The implementation of a wide array of Webbased products by community banks allows them to compete for customers that may traditionally be outside the “local” market. As the level of demand and customer affinity to Webbased banking services increases, the community banks should design a product mix that improves profit efficiency while engendering enhanced customer service quality. Such a strategic direction will expand the customer base of the firm in general and the market segment that is requiring robust online services in particular. However, the impact of Web sites and Web banking applications on the banks’ performance can be expected to be more pronounced in the longrun if these Web sites are more usable for the masses because more usable Web sites attract and retain more customers in the longrun thereby increasing revenues, reducing customer support costs [11], and increasing profits. The current research study is a sequel to our previous work [3] – [6] and attempts to study the impact of usability (user experience derived from the usability) of community banks’ Web banking efforts on their performance through heuristic evaluation of their Web site homepages. C. Usability Of Web Sites And Web Applications Now that most banks have already established themselves on the Web, competition on the Web is bound to increase and banks will have to focus on the usability of their Web sites and Web banking applications in order for their Web activities to sustain competition on the Web and payoff in the longrun. At the strategic level, usability is a prerequisite for EC success among other things as sites cannot sell unless people can find easily what they are looking for [12]. The Web sites should provide a truly usable environment that supports users’ goals and maximizes companies’ return on investments [13], [14]. Usability is defined as a multidimensional property of a user interface and is traditionally associated with the five quality attributes: (1) Learnability. The system should be easy to learn so that users can rapidly start getting some work done with the system? (2) Efficiency. The system should be efficient to use so that a high degree of productivity is possible once users learned the system? (3) Memorability. The system should be easy to remember so that a casual user can return to the system after some period of not having used it, without having to learn everything all over again? (4) Errors. The system should have low error rate so that users make few errors when they use the system and if they do make errors they can easily recover from them. (5) Satisfaction. The system should be pleasant to use so that users are satisfied when using it [11], [14]. From this definition it is clear that usability and usercentered design are so critical that Web sites design warrants a lifecycle approach, with appropriate steps taken in each development phase of the sites, to ensure
their usability [11]. Web usability is more important than ever before because of increased competition. For the Web sites, usability is critical for success because customers cannot purchase unless they find what they want [14]. For the customers, the Web is an empowering environment where one can buy or leave for competitor sites with just a click. Moreover, competition on the Web is not limited to a single industry because of cross selling and competition for the users’ time and attention. Further, Web users form their expectations for usability based on the usability of the best of all the other sites that they visit [15]. This is true even in the banking industry after the deregulation act as many types of financial institutions are aggregating an array of overlapping services, especially on their Web sites, in an attempt to attract and retain customers. In the network economy, the Web site is a firm’s primary interface to the customer [15]. Web design needs to cater to the masses that have less experience on the Web. Rarely can a site be successful if it is aimed at more advanced Web users [14], [16]. Design flaws and nonstandard interaction design lead to lost customers and sales [14], [15], [17]. Branding and unique design of Web sites may give a temporary advantage, but usability is a must for having sustained advantage. As more sites recognize the need for simplicity and usercentered design to earn the users’ business and loyalty to be successful, the usability barrier will be lowered [14]. Community banks do not have the scale and resources to match large commercial banks in their IT investments and product offerings. Fortunately, usability has the potential to act as a great equalizer of competition [18] putting community banks on equal footing with large commercial banks to compete for nonlocal customers. Therefore, community banks should emphasize on the usability of their Web sites and the limited Web banking applications that they offer. D. Usability Of Homepages Of Web Sites The homepage is the most important page on a Web site acting as a company’s face to the world and getting more page views than any other page. Firms spend millions of dollars in designing homepages and the impact of the homepage on a company’s profits is beyond simple measures of revenues. The homepage must make a sound impression to draw the attention of visitors and communicate where they are and where they can go in the site. At the same time, Web design is interaction design, and the role of homepages in the user experience that follows after entering the site is the key. The most critical role of the homepage is to communicate what the company is, the value the site offers over the competitors and the physical world, and the products/services offered. Users are often overwhelmed by homepages that do not clearly convey their options. If they cannot understand a homepage in about 10 seconds, they abandon the site and turn
3
to sites that are more usable [12]. Such is the importance of homepage usability that the abundance of choice and the ease of going elsewhere with just a click demand to make it extremely easy to enter a site [19]. Web banking customers are no exception to this as the cost of switching online to a new bank is relatively low as many banks and financial institutions provide a broad array of services, which covers almost everything an average customer needs, over the Web. A homepage serves as the portal to the site. Because the primary purpose of a homepage is to facilitate navigation in the site, it is critical that users be able to find the appropriate navigation area effortlessly, differentiate between the choices, and understand what lies beneath the links. The navigation area should also reveal the most important content of the site so that users understand the content of the toplevel categories. So, the challenge is to design a homepage that allows access to all important features without cramming them onto the page itself which can be achieved through focus, clarity, and understanding of users’ goals [13]. Therefore, homepage design decisions have impact on their usability and better usability of homepage is indispensable. Jakob’s Law of the Internet User Experience [14] states that Web users spend most of their online time on other sites than on any one site. This law implies the need for Web sites to follow design conventions to be more usable because users generally visit a large number of other Web site homepages and form a general mental model of the way homepages should work based on their experiences on these other sites. Because the decision to stick or leave a Web site is often made at the homepage, the homepage needs to communicate immediate value and enable users to find things relevant to what they are looking for within 25 to 35 seconds which implies that the homepage has to follow standard user interface design conventions [13], [14]. Strategically, most community banks have a clickand mortar business model as they view the Internet as a delivery channel that complements traditional brickandmortar branches, ATMs, and telephone banking call centers in delivering banking services. Such a hybrid EC strategy has synergistic opportunities if the online and offline operations collaborate [12]. By following appropriate usability guidelines on the homepage to communicate the hybrid EC strategy, as opposed to branding the Web site as a different entity from the bank, it is possible to present a unified face of the bank, build trust, and exploit the synergistic opportunities [13]. The existing literature talks about the importance of the usability of the Web sites and homepages. This research study intends to quantify the impact of the usability of homepages on the performance of community banks. III. HOMEPAGE USABILITY DESIGN GUIDELINES The Web is dynamic and designing Web sites for usability has no established axioms. The design features, either good or bad for usability, implemented by Web sites visited by 4
majority of people tend to become accepted guidelines within a particular industry. Based on extensive user tests on homepages observing what makes the homepages pass or fail user scrutiny, Jakob Nielsen and Marie Tahir [13] proposed 113 design guidelines to ensure usability of homepages of EC Web sites. However, not all of these guidelines are applicable to Web sites operating in the banking industry. Because of time and resource constraints, a subset of the applicable homepage usability design guidelines to banking industry has been chosen to evaluate the homepages of the sample community banks under study. This subset containing 32 usability guidelines fall under five categories namely Communicating the site’s purpose, Communicating information about the company, Content writing, Revealing content through examples, and Archives and accessing past content as discussed below. These five categories have been chosen because of their greater strategic importance to any Web site to be successful in the longrun and, as a first step in the immediate short term, to quickly help users identify what the site is about and make them stick to the homepage for a little longer if the site matches the users current needs thereby increasing the chance of a sale. The remaining usability guidelines appear to be more of tactical nature having implementation importance to help users complete their tasks in an easy and efficient way. A. Communicating The Site’s Purpose Communicating the site’s purpose is of utmost strategic importance as the homepage is the face of the Web presence of a company. In a short glance the homepage must communicate where users are, what the company does, and what products/services are available over the site. In order to communicate well, homepages must give appropriate emphasis to both branding and highpriority tasks. The homepage must also have a memorable and distinct look, so that users can recognize it as their starting place when coming from any other part of the site [13]. The usability design guidelines evaluated under this category are: · Show the company name and/or logo in a reasonable size and noticeable location. · Include a tag line that explicitly summarizes what the site or company does. · Emphasize what your site does that's valuable from the user's point of view, as well as how you differ from key competitors. · Emphasize the highest priority tasks so that users have a clear starting point on the homepage. · Clearly designate one page per site as the official homepage. · On your main company Web site, don't use the word “Web site” to refer to anything but the totality of the company's web presence. · Design the homepage to be clearly different from all the other pages on the site.
B. Communicating Information About The Company Business Web sites providing products/services need to provide a clear way to find information about their company. Providing information about the company gives credibility to the site and establishes trust with the users/customers who sometimes visit Web sites with the sole purpose of getting information about the company [13]. Trust is a very important issue in banking and financial services industry as it deals with the whole financial security of the customers. The usability design guidelines evaluated under this category are: · Group corporate information, such as About Us, Investor Relations, Press Room, Employment and other information about the company, in one distinct area. · Include a homepage link to an “About Us” section that gives users an overview about the company and links to any relevant details about your products, services, company values, business proposition, management team, and so forth. · If you want to get press coverage for your company, include a “Press Room” or “News Room” link on your homepage. · Present a unified face to the customer, in which the Web site is one of the touchpoints rather than an entity unto itself. · Include a “Contact Us” link on the homepage that goes to a page with all contact information for your company. · If you provide a “feedback” mechanism, specify the purpose of the link and whether it will be read by customer service or the webmaster, and so forth. · Don't include internal company information (which is targeted for employees and should go on the intranet) on the public Web site. · If your site gathers any customer information, include a “Privacy Policy” link on the homepage. C. Content Writing Content is the king on the Web as people visit Web sites for their content and not design. Therefore, effective content writing is one of the most critical aspects of Web design. Most users scan online content, rather than carefully read linebyline. So, online content must be optimized for scannability and drafted to convey maximum information in minimum words. Scannable content is especially important for homepages where the greatest number of topics have to be represented in a single and short page in an effective way and still capture and hold the users’ interest [13]. The usability design guidelines evaluated under this category are: · Use customerfocused language. Label sections and categories according to the value they hold for the customer, not according to what they do for your company. 5
· ·
· · · ·
·
·
· · ·
Avoid redundant content. Don't use clever phrases and marketing lingo that make people work too hard to figure out what you're saying. Use consistent capitalization and other style standards. Don't label a clearly defined area of the page if the content is sufficiently selfexplanatory. Avoid singleitem categories and singleitem bulleted lists. Use nonbreaking spaces between words in phrases that need to go together in order to be scannable and understood. Only use imperative language such as “Enter a City or Zip Code” for mandatory tasks, or qualify the statement appropriately. Spell out abbreviations, initialisms, and acronyms, and immediately follow them by the abbreviation, in the first instance. Avoid exclamation marks. Use all uppercase letters sparingly or not at all as a formatting style. Avoid using spaces and punctuation inappropriately, for emphasis.
D. Revealing Content Through Examples Showing examples of the site’s content on the homepage is an effective way of communicating the actual content of interest to the users. Example content can help instantly communicate what the site is about and whether the site can meet the current needs of the user. The primary benefit of using examples is that they can help users successfully differentiate between categories and navigate because they show the actual content beneath the abstract category names. Using examples effectively, a site can reveal the breadth of products or content offered. Specific examples can stimulate users’ interest more than the abstract category names provided in the main navigation apparatus [13]. The usability design guidelines evaluated under this category are: · Use examples to reveal the site's content, rather than just describing it. · For each example, have a link that goes directly to the detailed page for that example, rather than to a general category page of which that item is a part. · Provide a link to the broader category next to the specific example. · Make sure it's obvious which links lead to followup information about each example and which links lead to general information about the category as a whole. E. Archives And Accessing Past Content It is important and useful to have an archives section to store content that has recently been moved off the homepage.
Homepage is the place where the most important information in the Web site is featured. Often, important developments at the bank such as introduction of new products/services, promotions are featured on their homepage. Therefore, having an archive section for the past homepage content helps the existing as well as new customers, who can get to know what the bank has offered or communicated to the customers in the past that is still valid and relevant to them [13]. The usability design guideline evaluated under this category is: Make it easy to access anything that has been recently featured on your homepage, for example, in the last two weeks or month, by providing a list of recent features as well as putting recent items into the permanent archives. The homepages of the sample community banks were evaluated to find the compliance rate of each bank homepage to the above mentioned 32 usability design guidelines from the five categories. This usability compliance rate, termed Usability Index, is a measure of the degree of usability of the bank Web site homepages. Web site features, products, and services provided by firms vary from industry to industry as revealed by [20]. Within an industry, certain features are more valuable than others to the customers and have more impact on the performance of the Web site as a delivery channel. Therefore, the usability guidelines to provide and evaluate these Web features, products, and services also differ with industry. The usability guidelines selected in this study are applicable to banking Web sites. The following methods section addresses the methods by which the data has been collected for the homepage usability guidelines in this study. IV. METHODS The objective of this study is to examine the usability of the homepages of the sample community bank Web sites as defined by their compliance to the subset of the usability design guidelines mentioned in the previous section and to determine the impact of the usability of bank homepages on the bank’s performance. To address the objectives of this study, a sample of community banks has to be selected and data on the usability of their homepages as well as their financial performance need to be collected. A. Selecting The Sample Banks A list of all currently active banks in the states of Arizona, Iowa, Minnesota, Montana, North Dakota, South Dakota, and Texas has been obtained from the FDIC Web site. Banks from these states have been selected for two reasons. One, our previous work [3] – [6] that investigated the influence of Web banking features, products, and services on the performance of community banks used the population of banks in these particular states. Since this research study is a sequel to that study, a sample of banks has been chosen from the same states. Two, excepting Arizona, these states have a long and unique history in the concentration and evolution of
community banking in the U.S. From this list, only banks belonging to bank classes N (commercial bank, national/federal charter and Fed member, supervised by the Office of the Comptroller of the Currency), SM (commercial bank, state charter and Fed member, supervised by the Federal Reserve), and NM (commercial bank, state charter and Fed nonmember, supervised by the FDIC) are retained in the sample as they have the full mandate for regular banking services as opposed to bank classes SB (savings banks, state charter, supervised by the FDIC), SA (savings associations, state or federal charter, supervised by the Office of Thrift Supervision), and OI (insured U.S. branch of a foreign chartered institution) that are not regular banks and have limited and different function compared to commercial banks. As in our previous study [3] – [6], to set criteria for selecting the sample frame, a community bank is defined as any commercial bank with less than $1 billion in assets [21], [22], [23]. This classification is done based on 2005 year end assets as disclosed in FDIC call reports. Banks with more than $1 billion in assets were excluded from the sample. Within the community banks, those that fall in the asset category $500 million to <$1 billion were chosen as the final sample expecting that they fall with in the definition of a community bank and still may have reasonably well designed Web sites to conduct usability studies through heuristic evaluation. B. Collecting Usability Data Usability analysis of bank homepages is conducted using heuristic evaluation method with the homepage usability design guidelines proposed in [13] as mentioned in the guidelines section. Heuristic evaluation relies on evaluating a user interface relative to a known set of usability principles called the “heuristics” [11], [24]. Compared to a decade earlier, the field of Web usability has now matured sufficiently that specialized guidelines to codify the best design practices for specific components of a Web site have been developed through extensive research thereby making heuristic evaluation a valid and reliable research method.
TABLE 1: DATA RANGE CODING SCHEME USED FOR THE USABILITY GUIDELINES 0 0.25 0.50 0.75 1.0 Guideline is not followed. Guideline is partially followed, but not satisfactory. Guideline is partially followed, but neutral in terms of satisfaction. Guideline is partially followed, but satisfactory. Guideline is fully and consistently followed.
The data collection and observation process in the evaluation of bank homepages for compliance to usability design guidelines is consistent with the simple methodology proposed in [13]. However, the data range or scale of evaluation was extended from the proposed 0 – 0.5 – 1 scale to a 0 – 0.25 – 0.5 – 0.75 – 1 scale, as shown in Table 1, because this study is evaluating a sample of 58 bank homepages instead of a single bank homepage and it is 6
necessary to differentiate among the banks based on how well they implemented the suggested design guidelines. This data collection was done in 2006 Q3. C. Collecting Financial Performance Data The financial performance data for the sample banks were collected from the FDIC Call Reports and Performance and Condition Ratios reports for Q4 2005 which reflects consolidated reports for the year 2005. D. Assumptions and Limitations There are some assumptions and limitations in this study and the data collection process that need to be mentioned. Some bank performance studies impose age cutoff on banks while selecting the sample to help focus on mature banking institutions by excluding younger, de novo banks that typically operate with low efficiency and tend to under perform their competitors until they are about nine years old [25], [26]. This study does not impose any age cutoff restriction, however, out of the 58 sample banks, 55 are more than 8 years old, 2 are 5 years old and 1 is 3 ½ years old thereby satisfactorily mitigating the problem of de novo banks. This study does not take into account when a particular bank initiated their Web site because it was not mandatory for community banks to report their Web sites and Web activities in the call reports until 2002. Some banks may have started their Web activities very recently, in which case they need more time to realize the benefits of Web banking. However, all the 58 sample banks had Web site in 2003, but we do not know for how long they had their Web sites before 2003. The following section explains the models proposed for financial performance measures. V. MODEL DEVELOPMENT This research has two components: (1) analyzing the homepages of sample bank Web sites to evaluate their usability and develop a measure of usability of each bank’s homepage? and (2) modeling the financial performance of banks as a function of the usability index to explain the effect of the usability of the homepages of banks upon their financial performance. To address the first component, a usability index will be developed using the usability data collected for each bank which represents the usability of the homepage of the bank Web site. To address the second component, an econometric model will be proposed that contains a multiple regression specification to model select financial performance measures of community banks. The usability index will be integrated into the econometric model in an attempt to explore the effect of the usability of bank Web site homepages on the financial performance measures. A. Developing Usability Index Each bank homepage was evaluated for the 32 guidelines discussed in the guidelines section based on the scale 7
mentioned in the methods section. The data collected is aggregated for each bank and the aggregate score is divided by the total number of guidelines evaluated, which is 32, and multiplied by 100 to arrive at a percentage figure that represents compliance rate to home page usability design guidelines which is termed Usability Index. The usability index values for the sample banks ranged from 40.62% to 82.81% with an average of 66.11%. Web sites should conform to 80% of design guidelines to be considered usable [13], [27]. This index is used in the econometric model as an explanatory variable to examine the impact of usability of homepages on two reported bank performance measures ROE and ROA. B. Econometric Model A multiple regression modeling approach is chosen to explain the financial performance of community banks for the following reasons: (1) multiple regression analysis is robust and powerful analytical tool designed to explore all types of dependence relationships, (2) this approach is a straightforward dependence technique that can provide both prediction and explanation/description of the relationships among two or more intervally scaled variables, (3) a regression scheme can be used to examine the incremental and total explanatory power of many variables simultaneously [28], (4) this model design is consistent with existing literature that has applied multiple regression analysis to model the financial performance of banks [26], [29] – [31], and (5) a time series specification may have been better suited to model the performance trend of banks over the entire period of implementation of Web banking. However, the constraint that only a single data point is available for the usability index precludes the use of a timeseries econometric model. Two financial performance measures ROE and ROA are identified to be important in community banking [9], [26], [29], [30] – [34]. Most community banks are locally owned with a relatively high equity to assets ratio which implies that ROE is an important performance measure in evaluating a community bank. ROA is a standard measure of efficiency in utilization of the assets. Also, ROA and ROE are standard indicators of earnings for any firm. Specific models have been developed to ascertain the impact of select financial variables on these two financial performance measures. The usability index (Uindx) developed for each bank is used in the econometric model specifications for both the performance measures, ROE and ROA, as an explanatory variable of primary interest to predict and explain the effect of the usability of the homepages upon the bank’s performance. A community bank’s performance can be defined as a function of factors associated with assets, liabilities, employees, loan structure, interest margins, and the income expense structure. Therefore, the other explanatory variables used are business loans to total loans ratio (Blon), consumer
loans to total loans ratio (Clon), fixed assets to total asset ratio (Faset), equity capital to asset ratio (Eqcap), average rate of growth in assets (Agr), total assets per employee (Aemp), employment growth rate (Egr), noninterest income to total expenses ratio (NIex), level of inefficiency as measured by noninterest expenses to total revenue (NErev), liabilities to asset ratio (Last), and net interest margin (Imgn). These variables have been applied in prior studies [9], [26], [29], [32], [33], [35] – [37]. The models for ROE and ROA are hypothesized applying a subset of the above mentioned variables. The average asset growth rate and the average employment growth rate of the bank are calculated using the past three years’ data. The data for all the remaining financial variables have been downloaded for Q4 2005 from the FDIC Web site and used in estimating the models. The model for ROE is: Yi = ? + ?1Faset + ?2Blon + ?3Clon + ?4NIex + ?5NErev + ?6Last + ?7Aemp + ?8Imgn + ?9Agr + ?10Egr + ?11Uindx + ?i where Yi, ?, ?, and ?i represent the predicted ROE value, the intercept of regression line on yaxis, estimated regression coefficient, and error in prediction of ROE respectively. The model for ROA is: Yi = ? + ?1Faset + ?2Blon + ?3Clon + ?4Eqcap +?5NIex + ?6NErev + ?7Last + ?8Aemp + ?9Imgn + ?10Agr + ?11Egr + ?12Uindx + ?i where Yi, ?, ?, and ?i represent the predicted ROA value, the intercept of regression line on yaxis, estimated regression coefficient, and error in prediction of ROA respectively. The primary objective of this study, corresponding to part 2 of the problem statement, is to determine whether community banks providing more usable Web site homepages than their peers have a higher efficiency and profitability resulting from usability of Web banking in general and homepages in particular. The related hypothesis based upon the statistical significance of the model findings will attempt to measure both the magnitude and direction of the impact of usability index upon ROE and ROA. The secondary objective of this study, corresponding to part 1 of the problem statement, is to determine whether community banks having greater assets size provide more usable Web sites in general and in particular more usable homepages than their peers having lower asset size. For this, a simple regression model is proposed where asset size of the bank is used as explanatory variable and regressed against usability index. The related hypothesis based upon the statistical significance of the model findings will attempt to measure both the magnitude and direction of the impact of asset size upon usability index. The following section presents the results and discussions for the proposed models in detail.
VI. RESULTS AND DISCUSSION The descriptive statistics pertaining to all the financial variables and the usability index used in the econometric models are reported in Table 2. The results of the econometric model specifications with estimated parameters and corresponding tvalues are reported in Table 3. The regression output is provided in the Appendix.
TABLE 2: DESCRIPTIVE STATISTICS ON FINANCIAL VARIABLES USED IN ECONOMETRIC MODEL Sample average 14.8 1.3 1.8 24.9 8.1 27.4 62.8 91.1 3.7 4.0 13.6 7.8 8.9 66.1 Standard deviation 7.4 0.8 0.8 11.7 7.7 18.3 14.1 1.8 1.6 0.8 20.3 21.8 1.8 8.9
Variable description Return on equity Return on assets Fixed asset ratio Business loan to total loan Consumer loan to total loan Noninterest income over expenses Noninterest expense over revenue Liability asset ratio Assets per employee ($million) Net interest margin Asset growth rate Employment growth rate Equity asset ratio Usability index
Data Source: FDIC Online database (SDI) downloaded from http://www2.fdic.gov/sdi/main.asp and author’s estimates.
TABLE 3: REGRESSION RESULTS FOR RETURN ON EQUITY AND RETURN ON ASSETS
Return on equity Variable Coefficient Intercept Business loans to total loans Consumer loans to total loans Equity to assets Fixed assets to total assets Noninterest income to total expense Noninterest expenses to revenue Liabilities to assets Assets per employee Net interest margin Average asset growth rate 1.1310 0.2040** 0.2570** 1.3950** 0.1830 3.1560** 0.1720* 1.296 4.698 4.661 4.408 0.333 4.014 2.420 107.778** 0.0450 0.0170 t Value 3.570 0.930 0.204
Return on assets Coefficient 0.5900 0.0078 0.0086 0.0632* 0.1450 0.0245** 0.0250** t Value 0.620 1.670 1.044 2.082 1.735 5.882 4.745
0.0463 0.4130** 0.0210**
0.878 5.482 3.087
8
Average employment growth rate Usability index
0.1280 0.0740
1.882 1.187
0.0169** 0.0090
2.593 1.493
NOTE. *, ** Denote statistical significance at 5 and 1 percent levels respectively. A. ROE Model Results The econometric model for ROE has significant model fit. Some important results include the positive impact of non interest income to total expense ratio, the negative impact of noninterest expenses to revenue ratio (an inefficiency ratio), the positive impact of liabilities to assets ratio, and the positive impact of net interest margin which all show statistical significance at <1% level. This pvalue of <1% indicates that there is less than 1% probability for the null hypothesis, stating that the coefficients of these variables will be equal to zero, to occur. Also, the absolute t values, from the ttests of the null hypothesis that the coefficients of these variables are equal to zero in the population, are above the critical value of 3.291 thereby rejecting the null hypothesis and indicating that all these coefficients are statistically significant at <0.1% level. The impact of average asset growth rate is negative and statistically significant at the <5% level with an absolute tvalue 2.42 which is above the critical limit of 1.96 for 5% significance level. Usability index, however, shows a theoretically inexplicable and unexpected negative impact on ROE but statistically very insignificant with a pvalue of 0.241 indicating that there is 24.1% probability for the coefficient of usability index to be equal to zero which is heartening to see. The absolute tvalue for usability index is 1.87 which is below the critical limit of 1.96 for 5% significance level thereby accepting the null hypothesis that its coefficient is equal to 0. The coefficient of 2 determination (R ) for the model is 0.792 indicating that 79.2% of the variance in ROE is explained by the hypothesized regression model which is a good fit. The ANOVA test shows that F value is 15.96 which is above the critical value and pvalue is 0.000 indicating that there is <0.1% probability for the null hypothesis, the assumption that there is no true difference between the variables and any difference (statistically) is due to sampling errors, to occur. This indicates that the model variables as a group are a good predictor of ROE and hence the model is a good one. The variance inflation factor (VIF) is <10 for all the variables indicating that there is no significant collinearity problem between the model variables. B. ROA Model Results The econometric model for ROA has significant model fit. Some important results include the positive impact of non interest income to total expense ratio, the negative impact of noninterest expenses to revenue ratio (an inefficiency ratio), the positive impact of net interest margin, the negative impact of average asset growth rate, and the positive impact of 9
average employment growth rate which all show statistical significance at <1% level except average employment growth rate which is marginally close at 1.3% level. Also, the absolute t values for the coefficients of these variables are above the critical value of 3.291. The impact of equity to asset ratio is positive and statistically significant at the <5% level with an absolute tvalue 2.082. Usability index, again, shows unexpected negative impact on ROA but statistically very insignificant with a pvalue of 0.142 and an absolute tvalue of 1.493 which is again a welcome result. The liability to assets ratio has been excluded from the final model because it had high collinearity with equity capital to assets ratio owing to the fact that these two ratios are calculated using total assets figure and total assets equal liabilities plus equity. The 2 R for the model is 0.870 indicating that 87.0% of the variance in ROA is explained by the hypothesized regression model which is a good fit. The ANOVA test shows that F value is 27.89 which is above the critical value and pvalue is 0.000 or <0.1%. This indicates that the model variables as a group are a good predictor of ROA and hence the model is a good one. The variance inflation factor (VIF) is <10 for all the variables indicating that there is no significant collinearity problem between the model variables. The overall results, except usability index, are consistent with prior studies [9], [26], [33] providing validity to the proposed models. Therefore, the regression models are deemed to have acceptable fit indicating that the models predict and explain the performance measures satisfactorily. C. Hypotheses Results The variable of primary interest, the usability index, has a negative impact upon ROE and ROA. In particular, a one unit increase in the usability index would decrease the ROE by 0.074 units and decrease the ROA by 0.009 units. However, the effect of usability index on ROE and ROA is statistically very insignificant. These findings based on a subset of usability guidelines do not lend support to the hypothesis that community banks that provide better and more usable homepages than their peers have a higher efficiency and profitability resulting from the usability of their homepages. Pertaining to part 1 of the problem statement, the simple regression model for predicting usability index using asset 2 size as explanatory variable shows an R of 0.029 which is too low indicating that asset size is not a good predictor of usability index. Also, the hypothesis test of the coefficient for asset size show a tvalue of 1.294 which is below the critical limit and a pvalue of 0.201 which is unacceptable. The following section presents the summary and conclusions of this study and future direction of research. VII. CONCLUSIONS AND SUMMARY The managerial discussion (practical significance) of the conclusions drawn from the aforementioned results from the
econometric models will be discussed in this section followed by a brief summary of the overall research findings, a discussion of limitations, and suggestions for future research. A. Conclusions Some significant conclusions are drawn from the econometric models for ROE and ROA with respect to various explanatory variables. Many of the findings here are consistent with our previous study [3] – [6] and other existing literature. However, the conclusions for usability Index, the variable of primary interest, differ from the conclusions drawn on Web banking index in our previous study. 1) Usability Index: Our previous work found that, online banking index, a measure of the Web features, products and services provided on the bank Web sites, had a significant positive impact on ROE (p ? 0.001) and ROA (p ? 0.005) implying that online banking applications help increase a bank’s earnings. Banks that provide more effective Web sites and a wider array of online products and services than their peers appear to have higher efficiency and profitability. As all financial transactions are digitized, banks can easily access and analyze this information (both bank information and credit score information) rapidly and more effectively to identify underperforming and nonperforming assets as well as target better sources of revenues. However, in this research study, the usability index has a very insignificant negative impact on ROE and ROA implying that more usable homepages (Web sites) decrease the earnings which is not possible. Even if banks overspend on usability of their IT initiatives in their drive to provide more usable Web applications, the inefficiency ratio should reflect this problem rather than the usability index. The statistical insignificance of this negative impact is a blessing in disguise and a welcome result that does not reject the theory that better usability of Web applications should increase efficiency and profitability of banks. However, while comparing the results of the previous study with this study, one has to note that in the earlier study entire Web sites were evaluated for Web features, products, and services. This study evaluates neither the usability of entire Web sites nor the usability of the specific homepagerelated Web features, products, and services evaluated in the previous study alone. These two projects are as much independent as they are related conceptually. A body of literature pertaining to the period 1998 to 2003 [8] – [10] shows that though bank profitability is strongly correlated with Internet banking, the impact of Internet banking on bank profitability is not statistically significant. Apparently banks that adopted Internet banking already had higher profitability, accounting efficiency, and scale than other banks. However, by the end of 2005 the impact of Internet banking on bank profitability is found to be statistically significant as reported by our previous work [3] – [6] probably because rest of the community banks sensed
competition and opportunity for Internet banking and caught up with the early birds in offering Webbased services. Now, this study found that bank profitability is correlated with usability of Internet banking Web sites, but the impact of usability on bank profitability is not statistically significant. It could be the case that Web site usability in community banking sector is not mature enough yet to have significant impact on performance. However, on cross examining the data it is found that more efficient banks tend to maintain quality Web sites that have higher usability index though higher usability index is not leading to higher efficiency. This clearly indicates a potential trend where it appears that more efficient banks with scale are the first ones to adopt Internet banking and they are also the first ones to ensure the quality of their Web offerings by emphasizing on usability of their Web sites. Following the trend, one could expect in 2 to 3 years from now, less efficient banks will feel compelled by the competition to make their Web sites more usable at which point Web usability might show significant impact on bank profitability. Some of the limitations in the study might explain this unexpected negative and insignificant impact of usability index which differs from the positive and significant impact of the online banking index observed in the earlier study. The sample size of 58 used in this study, as opposed to 1271 used in the previous study, is very small for a powerful modeling technique like regression analysis that generally requires larger sample size especially for multiple regression analysis thus making the results less reliable. The asset category of $500 million to <$1billion used as a criteria to select the sample may have made a difference. The homepages are evaluated against a small set 32 homepage usability design guidelines out of a total of 113 guidelines available. Evaluating the remaining guidelines as well will provide a truer picture of the impact of the homepage usability on performance. The procedure used to develop the usability index in this study is not a sophisticated one capable of providing a measure of usability or user experience of the homepages that is capable of capturing the abstract nature of the impact of usability of homepages or Web sites as a whole on the performance of banks. Indirect measurement techniques or latency models like structural equation modeling may be better suited to develop a measure for the unobservable effects of Web usability that can then be used in regression models against performance measures. Moreover, homepage itself is a subset of the Web site. Data for some of the variables for the sample banks show skewness which might account to some extent for the unexpected results. More importantly, many of the community banks may neither have sufficient financial resources nor the critical mass of potential online customers to justify a reasonable budget for the usability of their Web banking initiatives. Pertaining to part 1 of the problem statement, the simple regression model for predicting usability index using asset
10
size as explanatory variable shows that asset size is not a good predictor of usability index. This concludes that banks with greater asset size need not necessarily build more usable Web sites. Again, the sample size, the asset size of the banks in the sample, the small subset of guidelines used to evaluate the homepages, and the procedure used for developing usability index may have influenced this result. B. Investigating the Joint Impact/Relationship of Web Features Index and Usability Index on Performance The previous study on Web features, products, and services showed that Web features and products/services do matter and they have significant impact on the bank performance. However, this study shows that usability does not matter or that usability has no significant impact on the bank performance though the limitations in this study and the previous study are differing. But existing literature pertaining to general usability studies and return on investment (ROI) studies points out that the Web features and their usability are closely related concepts and that the usability of Web sites does have significant impact on revenue generation. Though this may not imply that usability has an impact on efficiency of revenue generation which is different from simple measure of revenue generation, it is however still important to combine the two aspects of Web features and their usability and conduct a postmortem analysis to investigate the joint impact of Web features and usability on the performance of banks. First, the sample banks have been segregated into high, medium, and low categories based on their Web features/products/services scores from previous study and usability index score in this study. The natural data breaks in the sample data and even distribution of subsample sizes are the two criteria employed in segregating the sample banks. Then, with the high, medium, and low categories of usability index on xaxis and with the high, medium, and low categories of Web features/products/services categories on y axis, a cross tabulation matrix containing 9 cells is plotted using SPSS as shown in section C of Appendix. Of these 9 cells, the cell containing banks that are high on both usability and Web features (which is termed highhigh cell) and the cell containing banks that are low on both usability and Web features (which is termed lowlow cell) are of importance to test the joint impact of Web features and usability on bank performance. The highhigh cell has a sample size of 8 banks and the lowlow cell has a sample size of 7 banks. The same performance measures ROE and ROA, which are modeled in the model development section, are chosen again for analysis. The mean ROE and ROA for the banks in the highhigh cell and the lowlow cell are calculated separately after removing the outlying data points. The mean ROE and ROA for the entire sample of 58 banks are also calculated after removing the outlying data points. The pooled variance ttest for testing the statistical significance of the difference in
two means is employed to check if the averages of the performance measures of banks in highhigh cell are significantly different from the averages of respective performance measures of banks in lowlow cell. To make the tests more robust, the ttests are conducted 3way: (1) between means of performance measures of banks in highhigh cell and lowlow cell, (2) between means of performance measures of banks in highhigh cell and overall sample of 58, and (3) between means of performance measures of banks in lowlow cell and overall sample of 58. The null hypothesis in all these tests states that the means of the performance measures of the populations of highhigh and lowlow banks are equal failing to reject which we cannot prove that banks in highhigh cell perform significantly better than banks in lowlow cell which in turn means that banks in highhigh cell are no more efficient than banks in lowlow cell. The average ROE for overall sample, highhigh cell, and lowlow cell are 14.35, 14.68, and 13.42 respectively which means that the performance of banks in highhigh cell is better than the performance of overall sample of 58 banks which in turn is better than the performance of banks in low low cell. This is in line with theoretical expectation of higher efficiency of banks with better Web banking features and also usability. However, all the 3way ttests failed to reject the null hypothesis showing that the difference in the means of performance of these 3 groups in not significant. The average ROA for overall sample, highhigh cell, and lowlow cell are 1.30, 1.23, and 1.46 respectively which means that the performance of banks in lowlow cell is better than the performance of overall sample of 58 banks which in turn is better than the performance of banks in highhigh cell. But, this is not in line with the theoretical expectation of higher efficiency for banks with better Web banking features and also usability. The potential explanation for this unexpected result is that banks in highhigh cell may have significantly invested in assets for IT applications and these assets are tiedup and cannot produce any immediate returns. It is a fact that investments in IT applications will not give returns in the short term. In such a scenario, this unexpected result is bound to happen with any asset based performance measure such as ROA. However, again all the 3way ttests failed to reject the null hypothesis showing that the difference in the means of performance of these 3 groups in not significant. This clearly shows that the ROA of banks in highhigh cell is not significantly lower than ROA of banks in lowlow cell which is an expected result. The overall conclusion is that assetbased performance measures appear to be insignificantly lower for banks with high Web features and high usability compared to banks with low Web features and low usability while nonasset based performance measures appear to be higher for banks with high Web features and high usability compared to banks with low Web features and low usability probably because of higher investments in IT assets by banks that have high Web
11
features and high usability. We know that all these banks had Web sites from 2003 though we do not precisely know how long they had Web sites before 2003. Probably the ROA is lower for banks in highhigh cell because they have not recovered their investments in IT applications as yet and they may require more time to realize the performance efficiency benefits of OB applications. This trend of the results based on ROE and ROA, and the rejection of the null hypotheses in all the 6 ttests indicates that the banks, especially those in high high cell, may be in a twilight stage in realizing the benefits of Web banking and the usability of Web banking. This prediction is based on the fact that ROE, a measure closely related to assetbased performance measures, though not a direct assetbased performance measure in itself, is showing a higher performance for banks in highhigh cell compared to those in lowlow cell while ROA, a direct assetbased performance measure, is showing an insignificant lower performance for banks in highhigh cell compared to those in lowlow cell. The implications of these two findings have to be read in conjunction rather than in an insulated manner. However, it has to be noted that the insignificant results and the less precise conclusions are because of the small sample of banks in the highhigh and lowlow cells of the cross tabulation matrix and the overall sample size. Future research endeavors must make sure of sufficient sample size. C. Summary This research analyzed the impact of usability of community bank Web site homepages on the performance of community banks that are implementing electronic commerce with Internet banking. A usability index is developed as a measure of compliance to the homepage usability design guidelines to proxy the usability of homepages and this index is included in econometric models as an explanatory variable to examine whether it explains differences in community bank performance. Based on the sample, the results show that (1) the asset size of banks has no impact on the usability of the homepages of Web sites, and (2) the impact of usability of bank Web site homepages on bank performance is insignificant as a result of which no meaningful and reliable conclusions can be drawn. It could be the case that Web site usability in community banking sector is not mature enough yet to have significant impact on performance. However, it is also found that more efficient banks tend to maintain quality Web sites that have higher usability index though higher usability is not leading to higher efficiency. An important point to note is that usability is measured relative to certain users and certain tasks. Usability measurement therefore starts with the definition of a representative set of test tasks, relative to which the different usability attributes can be measured. Certain tradeoffs are involved in designing Web sites for usability simultaneously for both novice and expert users and hence usability is in a way a contradiction in itself [11].
D. Future Direction Of Research As community banks continue to implement EC applications, the usability of these applications cannot be ignored if community banks were to withstand the aggressive online competition from competitors in the market which is only bound to increase in the future as almost all banks have gone online and remaining ones are also following suit to catch up the online market. This leads to many future usability research endeavors in the banking sector beyond this project. To better study the effect of the usability of the homepages on financial performance of the banks, it would be interesting to evaluate the homepages with all the 113 homepage usability design guidelines and 40 homepage usability design statistics proposed in [13] based on extensive user testing. However, homepage is only a part of Web site and a heuristic evaluation of the entire Web sites for usability will be more reliable to assess the true and comprehensive impact of the usability of Web activities of banks on their financial performance. REFERENCES
[1] H.Beck, “Banking is essential, banks are not. The future of financial intermediation in the age of the Internet,” Netnomics, vol. 3 (1), pp. 722, 2001. [2] B. Foster, “Do or die time for community banks?” The RMA Journal, pp. 15–19, April 2001. [3] R. N. Acharya, A. Kagan, and S. R. Lingam, “Online banking and profit efficiency: Implications for community banking,” Journal of Money, Credit, and Banking, submitted for publication. [4] A. Kagan, R. N. Acharya, and S. R. Lingam, “Electronic business applications within community banks: A modeling approach,” presented at the Annual Meeting of the Decision Sciences Institute, San Francisco, California, 2005. [5] A. Kagan, R. N. Acharya, L. S. Rao, and V. Kodepaka, “Does Internet banking affect the performance of community banks?” presented at the Annual Meeting of the American Agricultural Economics Association, Providence, RI, 2005. [6] S. R. Lingam, “Performance aspects of community banks implementing electronic commerce with Internet banking,” M.S. thesis, Department of Agribusiness and Resource Management, Arizona State University, Tempe, 2005. [7] K. Furst, W. W. Lang, and D. E. Nolle, “Internet banking: Developments and prospects,” Economic and Policy Analysis Working Paper, Office of the Comptroller of the Currency, pp. 158, September 2000. [8] K. L. Egland, K. Furst, D. E. Nolle, and D. Robertson, “Banking over the Internet,” Quarterly Journal, Office of the Comptroller of the Currency, vol. 17, no. 4, pp. 2530, December 1998. [9] K. Furst, W. W. Lang, and D. E. Nolle, “Internet banking: Developments and prospects,” Program on Information Resources Policy (Jointly sponsored by Harvard University and Center for Information Policy Research), pp. 153, April 2002. [10] A. N. Berger, “The economic effects of technological progress: Evidence from the banking industry,” Journal of Money, Credit and Banking, vol. 35, no. 2, pp. 141176, 2003. [11] J. Nielsen, Usability Engineering. San Diego, California: Academic Press, 1993. [12] J. Nielsen, R. Molich, C. Snyder, and S. Farrell, ECommerce User Experience. Fremont, California: Nielsen Norman Group, 2001. [13] J. Nielsen and M. Tahir, Homepage Usability: 50 Websites Deconstructed. Indianapolis, Indiana: New Riders, 2002. [14] J. Nielsen and H. Loranger, Prioritizing Web Usability. Indianapolis, Indiana: New Riders, 2006. [15] J. Nielsen, Designing Web Usability: The Practice of Simplicity. Indianapolis, Indiana: New Riders, 2000.
12
[16] J. Nielsen. (1998, April 5). Nielsen’s Law of Internet bandwidth. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/980405.html [17] J. Nielsen. (2004, September 13). The need for Web design standards. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/20040913.html [18] J. Nielsen. (2000, August 6). Why Doc Searls doesn’t sell any books. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/20000806.html [19] J. Nielsen. (2000, February 6). Novice vs. expert users. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/20000206.html [20] A. Kagan, G. Post, and C. M. Noel, “Selling the Web: Web features used by retailers,” The Journal of Applied Business Research, vol. 16, no. 1, pp. 1526, 2000. [21] R. DeYoung, W. C. Hunter, and G. F. Udell, “Whither the community bank?” Journal of Financial Services Research, vol. 25:2/3, pp. 8184, 2004. [22] R. DeYoung, W. C. Hunter, and G. F. Udell, “The past, present, and probable future for community banks,” Journal of Financial Services Research, vol. 25:2/3, pp. 85133, 2004. [23] L. I. Nakamura, “Small Borrowers and the Survival of the Small Bank: Is Mouse Bank Mighty or Mickey?” Business Review, Federal Reserve Bank of Philadelphia, pp. 315, November/December 1994. [24] J. Nielsen. (2006, January 23). Ten best intranets of 2006. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/intranet_design.html [25] R. DeYoung, and I. Hasan, “The performance of de novo commercial banks: A profit efficiency approach,” Journal of Banking & Finance, vol. 22, pp. 565587, 1998. [26] K. J. Stiroh, “Do community banks benefit from diversification?” Journal of Financial Services Research, vol. 25:2/3, pp. 135160, 2004.
[27] V. Flanders and D. Peters, Son of Web Pages That Suck: Learn Good Design by Looking at Bad Design. Alameda, California: Sybex, 2002. [28] J. F. Hair, Jr., R. E. Anderson, R. L. Tatham, and W. C. Black, Multivariate Data Analysis, 5th ed. Upper Saddle River, New Jersey: Prentice Hall, 1998. [29] A. N. Berger and L. J. Mester, “Inside the black box: What explains differences in the efficiencies of financial institutions?” Journal of Banking & Finance, vol. 21, pp. 895947, 1997. [30] A. Akhigbe and J. E. McNulty, “The profit efficiency of small US commercial banks,” Journal of Banking & Finance, vol. 27, pp. 307325, 2003. [31] T. J. Yeager, “The demise of community banks? Local economic shocks are not to blame,” Journal of Banking & Finance, vol. xxx, pp. xxxxxx, 2003. [32] Gary C. Zimmerman, “California’s community banks in the 1990s,” Weekly Letter, Federal Reserve Bank of San Fancisco, no. 9604, January 26, 1996a. [33] Gary C. Zimmerman, “Factors influencing community bank performance in California,” Economic Review, Federal Reserve Bank of San Fancisco, no. 1, pp. 2642, 1996b. [34] A. P. Meyer and T. J. Yeager, “Are small rural banks vulnerable to local economic downturns?” Review, Federal Reserve Bank of ST. Louis, pp. 2538, March/April 2001. [35] A. N. Berger and G. F. Udell, “Relationship lending and lines of credit in small firm finance,” The Journal of Business, vol. 68,3, pp. 351381, July 1995. [36] P. E. Strahan and J. P. Weston, “Small business lending and the changing structure of the banking industry,” Journal of Banking & Finance, vol. 22, pp. 821845, 1998. [37] A. N. Berger, A. Saunders, J. M. Scalise, and G. F. Udell, “The effects of bank mergers and acquisitions on small business lending,” Journal of Financial Economics, vol. 50, pp. 187229, 1998.
APPENDIX SPSS output tables for the hypothesized econometric models. A. Output for part 1 of the problem statement
Model Summary Model 1 R R Square a .170 .029 Adjusted R Square .012 Std. Error of the Estimate 8.91589687
a. Predictors: (Constant), Total Assets
b ANOVA
Model 1
Regression Residual Total
Sum of Squares 133.130 4451.620 4584.750
df 1 56 57
Mean Square 133.130 79.493
F 1.675
Sig. a .201
a. Predictors: (Constant), Total Assets b. Dependent Variable: Usability Index
13
a Coefficients
Model 1
Unstandardized Coefficients B Std. Error (Constant) 58.499 5.997 Total Assets 1.109E05 .000
Standardized Coefficients Beta .170
t 9.756 1.294
Sig. .000 .201
a. Dependent Variable: Usability Index
B. Output for part 1 of the problem statement ROE Model:
b Model Summary
Model 1
R R Square a .890 .792
Adjusted R Square .743
Std. Error of the Estimate **********
a. Predictors: (Constant), UINDX, BLON, IMGN, CLON, AGR, FASET, LAST, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROE
b ANOVA
Model 1
Regression Residual Total
Sum of Squares 2489.825 652.378 3142.203
df 11 46 57
Mean Square 226.348 14.182
F 15.960
Sig. a .000
a. Predictors: (Constant), UINDX, BLON, IMGN, CLON, AGR, FASET, LAST, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROE
14
a Coefficients
Model 1
Unstandardized Coefficients B Std. Error (Constant) 107.778 30.200 FASET 1.131 .873 BLON 4.506E02 .049 CLON .017 .085 NIEX .204 .043 NEREV .257 .055 LAST 1.395 .316 AEMP .183 .549 IMGN 3.156 .786 AGR .172 .071 EGR .128 .068 UINDX .074 .063
Standardized Coefficients Beta .127 .071 .018 .505 .488 .344 .040 .337 .472 .376 .090
t 3.569 1.296 .928 .204 4.698 4.661 4.408 .333 4.014 2.420 1.882 1.187
Sig. .001 .201 .358 .839 .000 .000 .000 .741 .000 .020 .066 .241
Collinearity Statistics Tolerance VIF .471 .767 .570 .391 .412 .741 .319 .639 .119 .113 .787 2.125 1.304 1.756 2.555 2.428 1.350 3.138 1.566 8.412 8.832 1.270
a. Dependent Variable: ROE
ROA Model:
b Model Summary
Model 1
R R Square a .933 .870
Adjusted R Square .838
Std. Error of the Estimate **********
a. Predictors: (Constant), EQCAP, IMGN, CLON, FASET, AGR, UINDX, BLON, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROA
b ANOVA
Model 1
Regression Residual Total
Sum of Squares 40.036 6.003 46.039
df 11 46 57
Mean Square 3.640 .131
F 27.890
Sig. a .000
a. Predictors: (Constant), EQCAP, IMGN, CLON, FASET, AGR, UINDX, BLON, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROA
15
a Coefficients
Model 1
Unstandardized Coefficients B Std. Error (Constant) .590 .952 FASET .145 .084 BLON 7.779E03 .005 CLON 8.561E03 .008 NIEX 2.453E02 .004 NEREV .025 .005 AEMP 4.626E02 .053 IMGN .413 .075 AGR .021 .007 EGR 1.692E02 .007 UINDX .009 .006 EQCAP 6.318E02 .030
Standardized Coefficients Beta .135 .102 .074 .501 .394 .083 .365 .477 .410 .090 .129
t .620 1.735 1.670 1.044 5.882 4.745 .878 5.482 3.087 2.593 1.493 2.082
Sig. .538 .089 .102 .302 .000 .000 .384 .000 .003 .013 .142 .043
Collinearity Statistics Tolerance VIF .471 .767 .570 .391 .412 .319 .639 .119 .113 .787 .741 2.125 1.304 1.756 2.555 2.428 3.138 1.566 8.412 8.832 1.270 1.350
a. Dependent Variable: ROA
C. Cross Tabulation Chart Comparing Usability Index and Web Features (Banking) Index:
Crosstabulation of Usability Index and Banking Index Usability Index 1.00 2.00 7 4 50.0% 28.6% 10 6 37.0% 22.2% 2 7 11.8% 41.2% 19 17 32.8% 29.3%
3.00 3 21.4% 11 40.7% 8 47.1% 22 37.9%
Banking Index
1.00 2.00 3.00
Total
Count % within Recoded Bindx Count % within Recoded Bindx Count % within Recoded Bindx Count % within Recoded Bindx
Total 14 100.0% 27 100.0% 17 100.0% 58 100.0%
Note: Codes 1.00, 2.00, and 3.00 for Usability Index (Uindx) and Banking Index (Bindx) correspond to low, medium, and high values respectively for the concerned index.
16
doc_819536545.pdf
To access a financial institution's online banking facility, a customer having personal Internet access must register with the institution for the service, and set up some password (under various names) for customer verification.
Impact of Web Banking Usability on Community Bank Performance: A Heuristic Evaluation of Community Bank Website Homepages
Srinivasa Rao Lingam Master of Computing Studies Division of Computing Studies Arizona State University [email protected]
Abstract—This research study attempts to study the impact of usability of community banks’ Web banking efforts on their performance through heuristic evaluation of their Web site homepages. The homepages were evaluated and a measure of their usability is developed which we term usability index. This usability index is regressed against bank performance indicators along with other explanatory variables. Based on the sample, the results show that (1) the asset size of banks has no impact on the usability of the homepages of Web sites, and (2) the impact of usability of bank Web site homepages on bank performance is insignificant as a result of which no meaningful and reliable conclusions can be drawn. It may be that Web usability in community banking sector is not mature enough to have significant impact on performance. However, on cross examining the data it is found that more efficient banks tend to maintain quality Web sites that have higher usability index though higher usability index is not leading to higher efficiency. Key Words—Community Banks, Web Banking, Web Usability, Heuristic Evaluation, Profit Efficiency.
community banks’ implementing Web banking as a critical success factor in the banks’ business model and revealed that Web banking activities have significant positive impact on the banks’ return on assets (ROA) and return on equity (ROE) and improves the asset quality by identifying non performing loans. Further, increasing Information Technology (IT) costs reduce ROA and ROE but improve the loan quality. This research study is a sequel to our previous work [3] – [6] and attempts to study the impact of usability (user experience derived from the usability) of community banks’ Web banking efforts, particularly homepages, on their performance through heuristic evaluation of their Web site homepages. A. Statement Of The Problem This project intends to evaluate the impact of the usability of homepages of bank Web sites on their financial performance. To this effect, the problem statement has two parts: (1) Do banks having greater asset size have more usable homepages (or Web sites)? This helps determine if the investment of resources in the usability of Web sites is driven by asset size. (2) Do banks having more usable homepages (or Web sites), as indicated by a higher compliance rate to usability design guidelines, show more profitability or perform better than competing banks with less usable homepages because of the usability of their homepages (or Web sites)? Alternately, does usability have an effect on financial performance of a bank? This helps determine if there is a payoff from the investment of resources in the usability of Web sites and their homepages. B. Research Objectives The specific research objectives of this study are to: · Conduct heuristic evaluation of the Web site homepages of the community banks under study to assess their compliance with the homepage usability design guidelines and use the data to develop a measure that represents the overall usability of the homepages? · Collect relevant financial performance data for the sample community banks from the Consolidated Report of Condition and Income (“Call Reports”) filed by the banks with the Federal Deposit
I. INTRODUCTION Community banks historically had a strategic advantage over large regional and national banks in providing loans to small and mediumsized businesses especially in nonmetro areas. However, recent innovations in the use of communication and information technology, the deregulation of the banking sector, and an increasing use of electronic banking products (Internet banking) have altered the community bankerlocal merchant relationship. These developments have increased the competition from large commercial banks in the small business lending markets, which were mainly served by the community banks [1], [2]. As a result, community banks started implementing electronic commerce (EC) activities, particularly Internet (online or Web) banking applications, as a strategic measure to withstand competition to remain viable market players and gain market share. Community banks started implementing Web sites around 1997 as simple information tools but a majority of them implemented online banking applications on their Web sites during the threeyear period 2001 to 2003. Our previous work [3] – [6] assessed the strategy of the
1
·
Insurance Corporation (FDIC)? and Analyze the bank level performance with appropriate econometric modeling techniques using data pertaining to the usability measure developed in step 1 as an explanatory variable of primary interest, along with other explanatory variables formed from financial performance data collected in step 2, in an attempt to assess the impact of usability of homepages on the specific financial performance measures of banks.
C. Scope The study includes sample community banks insured by FDIC in the upper Midwest and Southwest regions of U.S. covering Arizona, Iowa, Minnesota, Montana, North Dakota, South Dakota and Texas due to the preponderance of traditional unit banks and community banks in these regions. This study covers the usability of homepages alone as evaluated with a limited set of homepage usability design guidelines, but not the entire Web sites. 1) Assumptions: The main assumptions in this research concern the benefits of Web banking activities for community banks. Web banking is assumed to be a critical success factor for community banks in withstanding competition to stay in business. Web banking is assumed to bring new markets and an emerging customer base to the community banks and help retain customers as opposed to the traditional brickand mortar banking model. Above all, to successfully sustain the Web banking activities in the longrun, the Web sites and their homepages must be usable for the masses. 2) Limitations: The use of EC in community banking is continuously evolving and design of the Web sites and Web banking applications and features on the Web sites change from time to time. However, this study takes a snap shot of the usability of the design and content of homepages of community banks at just one point of time. As a result, the usability of homepages of bank Web sites could only be evaluated against the financial performance data for the current observation period, rather than the entire time period of Web banking implementation. II. REVIEW OF LITERATURE EC technology is the new delivery channel in banking designed to reach bank customers efficiently. For community banks, using technology in relationship banking assumes greater importance as these local banks have traditionally based their business model on relationship banking. This section reviews relevant literature and previous research pertaining to the role played by information and communication technology in the evolution of community banking as well as the importance of usability of these EC initiatives (especially Web banking) to reap the real benefits that accrue from their implementation in the longrun.
A. EC Technology In Community Banking My previous work [6] presented the usage and role of EC technology in the history and evolution of community banking over the last three decades in the midst of deregulation in the banking industry, innovations in financial services, technological advances, increased competitive rivalry, changing industry structure, and changing business models and business processes which have profoundly affected the size and health of the U.S. community banking sector and the availability and quality of banking products and services. The banking and nonbanking activities that were implemented using EC, especially Web banking, and the benefits accrued have been discussed. The vital importance, role, and usage of technology in improving efficiency and productivity in the traditional relationship banking model of community banks has been discussed. The issues of privacy, transaction and information security, customer satisfaction, loyalty, and trust in the online relationship banking environment have been discussed. B. Impact Of Web Banking Activities Except for the de novo (i.e., new) banks, banks in all size categories offering Internet banking tend to rely less on traditional banking (interestyielding activities and deposits) than do nonInternet banks, and outperform nonInternet banks in terms of profitability. Internet de novo banks tend to be less efficient and less profitable than nonInternet de novo banks. For most banks, the low percentage of customers using Internet banking and the relatively modest cost of setting up an Internet banking Web site makes it unlikely that Internet banking is having a noticeable impact on firm profits. However, the largest and smallest banks may be exceptions to this. Large commercial banks have a disproportionately sizeable share of Internet banking usage. For small banks with less than $100 million assets and de novo banks, Internet banking could be a primary reason for their lack of profitability if they rely more on an Internetbased strategy as the costs of offering Internet banking may be significant, if not prohibitive, for these banks [7]. Though bank profitability is strongly correlated with Internet banking, the impact of Internet banking on bank profitability is not statistically significant. Apparently banks that adopted Internet banking already had higher profitability, accounting efficiency, and scale than other banks [8] – [10]. Our previous work assessed the strategy of the community banks’ implementing Web banking as a critical success factor in the banks’ business model and revealed that Web banking activities have significant positive impact on the banks’ ROA and ROE and improves the asset quality by identifying non performing loans. Further, increasing IT costs reduce ROA and ROE but improve the loan quality. Their results show that community banks that provide extensive online banking services tend to perform better (approximately 3 percent more profit efficient) than competitors who lag behind. These
2
results indicate that online banking is an important strategic option for competitive positioning of community banks. The implementation of a wide array of Webbased products by community banks allows them to compete for customers that may traditionally be outside the “local” market. As the level of demand and customer affinity to Webbased banking services increases, the community banks should design a product mix that improves profit efficiency while engendering enhanced customer service quality. Such a strategic direction will expand the customer base of the firm in general and the market segment that is requiring robust online services in particular. However, the impact of Web sites and Web banking applications on the banks’ performance can be expected to be more pronounced in the longrun if these Web sites are more usable for the masses because more usable Web sites attract and retain more customers in the longrun thereby increasing revenues, reducing customer support costs [11], and increasing profits. The current research study is a sequel to our previous work [3] – [6] and attempts to study the impact of usability (user experience derived from the usability) of community banks’ Web banking efforts on their performance through heuristic evaluation of their Web site homepages. C. Usability Of Web Sites And Web Applications Now that most banks have already established themselves on the Web, competition on the Web is bound to increase and banks will have to focus on the usability of their Web sites and Web banking applications in order for their Web activities to sustain competition on the Web and payoff in the longrun. At the strategic level, usability is a prerequisite for EC success among other things as sites cannot sell unless people can find easily what they are looking for [12]. The Web sites should provide a truly usable environment that supports users’ goals and maximizes companies’ return on investments [13], [14]. Usability is defined as a multidimensional property of a user interface and is traditionally associated with the five quality attributes: (1) Learnability. The system should be easy to learn so that users can rapidly start getting some work done with the system? (2) Efficiency. The system should be efficient to use so that a high degree of productivity is possible once users learned the system? (3) Memorability. The system should be easy to remember so that a casual user can return to the system after some period of not having used it, without having to learn everything all over again? (4) Errors. The system should have low error rate so that users make few errors when they use the system and if they do make errors they can easily recover from them. (5) Satisfaction. The system should be pleasant to use so that users are satisfied when using it [11], [14]. From this definition it is clear that usability and usercentered design are so critical that Web sites design warrants a lifecycle approach, with appropriate steps taken in each development phase of the sites, to ensure
their usability [11]. Web usability is more important than ever before because of increased competition. For the Web sites, usability is critical for success because customers cannot purchase unless they find what they want [14]. For the customers, the Web is an empowering environment where one can buy or leave for competitor sites with just a click. Moreover, competition on the Web is not limited to a single industry because of cross selling and competition for the users’ time and attention. Further, Web users form their expectations for usability based on the usability of the best of all the other sites that they visit [15]. This is true even in the banking industry after the deregulation act as many types of financial institutions are aggregating an array of overlapping services, especially on their Web sites, in an attempt to attract and retain customers. In the network economy, the Web site is a firm’s primary interface to the customer [15]. Web design needs to cater to the masses that have less experience on the Web. Rarely can a site be successful if it is aimed at more advanced Web users [14], [16]. Design flaws and nonstandard interaction design lead to lost customers and sales [14], [15], [17]. Branding and unique design of Web sites may give a temporary advantage, but usability is a must for having sustained advantage. As more sites recognize the need for simplicity and usercentered design to earn the users’ business and loyalty to be successful, the usability barrier will be lowered [14]. Community banks do not have the scale and resources to match large commercial banks in their IT investments and product offerings. Fortunately, usability has the potential to act as a great equalizer of competition [18] putting community banks on equal footing with large commercial banks to compete for nonlocal customers. Therefore, community banks should emphasize on the usability of their Web sites and the limited Web banking applications that they offer. D. Usability Of Homepages Of Web Sites The homepage is the most important page on a Web site acting as a company’s face to the world and getting more page views than any other page. Firms spend millions of dollars in designing homepages and the impact of the homepage on a company’s profits is beyond simple measures of revenues. The homepage must make a sound impression to draw the attention of visitors and communicate where they are and where they can go in the site. At the same time, Web design is interaction design, and the role of homepages in the user experience that follows after entering the site is the key. The most critical role of the homepage is to communicate what the company is, the value the site offers over the competitors and the physical world, and the products/services offered. Users are often overwhelmed by homepages that do not clearly convey their options. If they cannot understand a homepage in about 10 seconds, they abandon the site and turn
3
to sites that are more usable [12]. Such is the importance of homepage usability that the abundance of choice and the ease of going elsewhere with just a click demand to make it extremely easy to enter a site [19]. Web banking customers are no exception to this as the cost of switching online to a new bank is relatively low as many banks and financial institutions provide a broad array of services, which covers almost everything an average customer needs, over the Web. A homepage serves as the portal to the site. Because the primary purpose of a homepage is to facilitate navigation in the site, it is critical that users be able to find the appropriate navigation area effortlessly, differentiate between the choices, and understand what lies beneath the links. The navigation area should also reveal the most important content of the site so that users understand the content of the toplevel categories. So, the challenge is to design a homepage that allows access to all important features without cramming them onto the page itself which can be achieved through focus, clarity, and understanding of users’ goals [13]. Therefore, homepage design decisions have impact on their usability and better usability of homepage is indispensable. Jakob’s Law of the Internet User Experience [14] states that Web users spend most of their online time on other sites than on any one site. This law implies the need for Web sites to follow design conventions to be more usable because users generally visit a large number of other Web site homepages and form a general mental model of the way homepages should work based on their experiences on these other sites. Because the decision to stick or leave a Web site is often made at the homepage, the homepage needs to communicate immediate value and enable users to find things relevant to what they are looking for within 25 to 35 seconds which implies that the homepage has to follow standard user interface design conventions [13], [14]. Strategically, most community banks have a clickand mortar business model as they view the Internet as a delivery channel that complements traditional brickandmortar branches, ATMs, and telephone banking call centers in delivering banking services. Such a hybrid EC strategy has synergistic opportunities if the online and offline operations collaborate [12]. By following appropriate usability guidelines on the homepage to communicate the hybrid EC strategy, as opposed to branding the Web site as a different entity from the bank, it is possible to present a unified face of the bank, build trust, and exploit the synergistic opportunities [13]. The existing literature talks about the importance of the usability of the Web sites and homepages. This research study intends to quantify the impact of the usability of homepages on the performance of community banks. III. HOMEPAGE USABILITY DESIGN GUIDELINES The Web is dynamic and designing Web sites for usability has no established axioms. The design features, either good or bad for usability, implemented by Web sites visited by 4
majority of people tend to become accepted guidelines within a particular industry. Based on extensive user tests on homepages observing what makes the homepages pass or fail user scrutiny, Jakob Nielsen and Marie Tahir [13] proposed 113 design guidelines to ensure usability of homepages of EC Web sites. However, not all of these guidelines are applicable to Web sites operating in the banking industry. Because of time and resource constraints, a subset of the applicable homepage usability design guidelines to banking industry has been chosen to evaluate the homepages of the sample community banks under study. This subset containing 32 usability guidelines fall under five categories namely Communicating the site’s purpose, Communicating information about the company, Content writing, Revealing content through examples, and Archives and accessing past content as discussed below. These five categories have been chosen because of their greater strategic importance to any Web site to be successful in the longrun and, as a first step in the immediate short term, to quickly help users identify what the site is about and make them stick to the homepage for a little longer if the site matches the users current needs thereby increasing the chance of a sale. The remaining usability guidelines appear to be more of tactical nature having implementation importance to help users complete their tasks in an easy and efficient way. A. Communicating The Site’s Purpose Communicating the site’s purpose is of utmost strategic importance as the homepage is the face of the Web presence of a company. In a short glance the homepage must communicate where users are, what the company does, and what products/services are available over the site. In order to communicate well, homepages must give appropriate emphasis to both branding and highpriority tasks. The homepage must also have a memorable and distinct look, so that users can recognize it as their starting place when coming from any other part of the site [13]. The usability design guidelines evaluated under this category are: · Show the company name and/or logo in a reasonable size and noticeable location. · Include a tag line that explicitly summarizes what the site or company does. · Emphasize what your site does that's valuable from the user's point of view, as well as how you differ from key competitors. · Emphasize the highest priority tasks so that users have a clear starting point on the homepage. · Clearly designate one page per site as the official homepage. · On your main company Web site, don't use the word “Web site” to refer to anything but the totality of the company's web presence. · Design the homepage to be clearly different from all the other pages on the site.
B. Communicating Information About The Company Business Web sites providing products/services need to provide a clear way to find information about their company. Providing information about the company gives credibility to the site and establishes trust with the users/customers who sometimes visit Web sites with the sole purpose of getting information about the company [13]. Trust is a very important issue in banking and financial services industry as it deals with the whole financial security of the customers. The usability design guidelines evaluated under this category are: · Group corporate information, such as About Us, Investor Relations, Press Room, Employment and other information about the company, in one distinct area. · Include a homepage link to an “About Us” section that gives users an overview about the company and links to any relevant details about your products, services, company values, business proposition, management team, and so forth. · If you want to get press coverage for your company, include a “Press Room” or “News Room” link on your homepage. · Present a unified face to the customer, in which the Web site is one of the touchpoints rather than an entity unto itself. · Include a “Contact Us” link on the homepage that goes to a page with all contact information for your company. · If you provide a “feedback” mechanism, specify the purpose of the link and whether it will be read by customer service or the webmaster, and so forth. · Don't include internal company information (which is targeted for employees and should go on the intranet) on the public Web site. · If your site gathers any customer information, include a “Privacy Policy” link on the homepage. C. Content Writing Content is the king on the Web as people visit Web sites for their content and not design. Therefore, effective content writing is one of the most critical aspects of Web design. Most users scan online content, rather than carefully read linebyline. So, online content must be optimized for scannability and drafted to convey maximum information in minimum words. Scannable content is especially important for homepages where the greatest number of topics have to be represented in a single and short page in an effective way and still capture and hold the users’ interest [13]. The usability design guidelines evaluated under this category are: · Use customerfocused language. Label sections and categories according to the value they hold for the customer, not according to what they do for your company. 5
· ·
· · · ·
·
·
· · ·
Avoid redundant content. Don't use clever phrases and marketing lingo that make people work too hard to figure out what you're saying. Use consistent capitalization and other style standards. Don't label a clearly defined area of the page if the content is sufficiently selfexplanatory. Avoid singleitem categories and singleitem bulleted lists. Use nonbreaking spaces between words in phrases that need to go together in order to be scannable and understood. Only use imperative language such as “Enter a City or Zip Code” for mandatory tasks, or qualify the statement appropriately. Spell out abbreviations, initialisms, and acronyms, and immediately follow them by the abbreviation, in the first instance. Avoid exclamation marks. Use all uppercase letters sparingly or not at all as a formatting style. Avoid using spaces and punctuation inappropriately, for emphasis.
D. Revealing Content Through Examples Showing examples of the site’s content on the homepage is an effective way of communicating the actual content of interest to the users. Example content can help instantly communicate what the site is about and whether the site can meet the current needs of the user. The primary benefit of using examples is that they can help users successfully differentiate between categories and navigate because they show the actual content beneath the abstract category names. Using examples effectively, a site can reveal the breadth of products or content offered. Specific examples can stimulate users’ interest more than the abstract category names provided in the main navigation apparatus [13]. The usability design guidelines evaluated under this category are: · Use examples to reveal the site's content, rather than just describing it. · For each example, have a link that goes directly to the detailed page for that example, rather than to a general category page of which that item is a part. · Provide a link to the broader category next to the specific example. · Make sure it's obvious which links lead to followup information about each example and which links lead to general information about the category as a whole. E. Archives And Accessing Past Content It is important and useful to have an archives section to store content that has recently been moved off the homepage.
Homepage is the place where the most important information in the Web site is featured. Often, important developments at the bank such as introduction of new products/services, promotions are featured on their homepage. Therefore, having an archive section for the past homepage content helps the existing as well as new customers, who can get to know what the bank has offered or communicated to the customers in the past that is still valid and relevant to them [13]. The usability design guideline evaluated under this category is: Make it easy to access anything that has been recently featured on your homepage, for example, in the last two weeks or month, by providing a list of recent features as well as putting recent items into the permanent archives. The homepages of the sample community banks were evaluated to find the compliance rate of each bank homepage to the above mentioned 32 usability design guidelines from the five categories. This usability compliance rate, termed Usability Index, is a measure of the degree of usability of the bank Web site homepages. Web site features, products, and services provided by firms vary from industry to industry as revealed by [20]. Within an industry, certain features are more valuable than others to the customers and have more impact on the performance of the Web site as a delivery channel. Therefore, the usability guidelines to provide and evaluate these Web features, products, and services also differ with industry. The usability guidelines selected in this study are applicable to banking Web sites. The following methods section addresses the methods by which the data has been collected for the homepage usability guidelines in this study. IV. METHODS The objective of this study is to examine the usability of the homepages of the sample community bank Web sites as defined by their compliance to the subset of the usability design guidelines mentioned in the previous section and to determine the impact of the usability of bank homepages on the bank’s performance. To address the objectives of this study, a sample of community banks has to be selected and data on the usability of their homepages as well as their financial performance need to be collected. A. Selecting The Sample Banks A list of all currently active banks in the states of Arizona, Iowa, Minnesota, Montana, North Dakota, South Dakota, and Texas has been obtained from the FDIC Web site. Banks from these states have been selected for two reasons. One, our previous work [3] – [6] that investigated the influence of Web banking features, products, and services on the performance of community banks used the population of banks in these particular states. Since this research study is a sequel to that study, a sample of banks has been chosen from the same states. Two, excepting Arizona, these states have a long and unique history in the concentration and evolution of
community banking in the U.S. From this list, only banks belonging to bank classes N (commercial bank, national/federal charter and Fed member, supervised by the Office of the Comptroller of the Currency), SM (commercial bank, state charter and Fed member, supervised by the Federal Reserve), and NM (commercial bank, state charter and Fed nonmember, supervised by the FDIC) are retained in the sample as they have the full mandate for regular banking services as opposed to bank classes SB (savings banks, state charter, supervised by the FDIC), SA (savings associations, state or federal charter, supervised by the Office of Thrift Supervision), and OI (insured U.S. branch of a foreign chartered institution) that are not regular banks and have limited and different function compared to commercial banks. As in our previous study [3] – [6], to set criteria for selecting the sample frame, a community bank is defined as any commercial bank with less than $1 billion in assets [21], [22], [23]. This classification is done based on 2005 year end assets as disclosed in FDIC call reports. Banks with more than $1 billion in assets were excluded from the sample. Within the community banks, those that fall in the asset category $500 million to <$1 billion were chosen as the final sample expecting that they fall with in the definition of a community bank and still may have reasonably well designed Web sites to conduct usability studies through heuristic evaluation. B. Collecting Usability Data Usability analysis of bank homepages is conducted using heuristic evaluation method with the homepage usability design guidelines proposed in [13] as mentioned in the guidelines section. Heuristic evaluation relies on evaluating a user interface relative to a known set of usability principles called the “heuristics” [11], [24]. Compared to a decade earlier, the field of Web usability has now matured sufficiently that specialized guidelines to codify the best design practices for specific components of a Web site have been developed through extensive research thereby making heuristic evaluation a valid and reliable research method.
TABLE 1: DATA RANGE CODING SCHEME USED FOR THE USABILITY GUIDELINES 0 0.25 0.50 0.75 1.0 Guideline is not followed. Guideline is partially followed, but not satisfactory. Guideline is partially followed, but neutral in terms of satisfaction. Guideline is partially followed, but satisfactory. Guideline is fully and consistently followed.
The data collection and observation process in the evaluation of bank homepages for compliance to usability design guidelines is consistent with the simple methodology proposed in [13]. However, the data range or scale of evaluation was extended from the proposed 0 – 0.5 – 1 scale to a 0 – 0.25 – 0.5 – 0.75 – 1 scale, as shown in Table 1, because this study is evaluating a sample of 58 bank homepages instead of a single bank homepage and it is 6
necessary to differentiate among the banks based on how well they implemented the suggested design guidelines. This data collection was done in 2006 Q3. C. Collecting Financial Performance Data The financial performance data for the sample banks were collected from the FDIC Call Reports and Performance and Condition Ratios reports for Q4 2005 which reflects consolidated reports for the year 2005. D. Assumptions and Limitations There are some assumptions and limitations in this study and the data collection process that need to be mentioned. Some bank performance studies impose age cutoff on banks while selecting the sample to help focus on mature banking institutions by excluding younger, de novo banks that typically operate with low efficiency and tend to under perform their competitors until they are about nine years old [25], [26]. This study does not impose any age cutoff restriction, however, out of the 58 sample banks, 55 are more than 8 years old, 2 are 5 years old and 1 is 3 ½ years old thereby satisfactorily mitigating the problem of de novo banks. This study does not take into account when a particular bank initiated their Web site because it was not mandatory for community banks to report their Web sites and Web activities in the call reports until 2002. Some banks may have started their Web activities very recently, in which case they need more time to realize the benefits of Web banking. However, all the 58 sample banks had Web site in 2003, but we do not know for how long they had their Web sites before 2003. The following section explains the models proposed for financial performance measures. V. MODEL DEVELOPMENT This research has two components: (1) analyzing the homepages of sample bank Web sites to evaluate their usability and develop a measure of usability of each bank’s homepage? and (2) modeling the financial performance of banks as a function of the usability index to explain the effect of the usability of the homepages of banks upon their financial performance. To address the first component, a usability index will be developed using the usability data collected for each bank which represents the usability of the homepage of the bank Web site. To address the second component, an econometric model will be proposed that contains a multiple regression specification to model select financial performance measures of community banks. The usability index will be integrated into the econometric model in an attempt to explore the effect of the usability of bank Web site homepages on the financial performance measures. A. Developing Usability Index Each bank homepage was evaluated for the 32 guidelines discussed in the guidelines section based on the scale 7
mentioned in the methods section. The data collected is aggregated for each bank and the aggregate score is divided by the total number of guidelines evaluated, which is 32, and multiplied by 100 to arrive at a percentage figure that represents compliance rate to home page usability design guidelines which is termed Usability Index. The usability index values for the sample banks ranged from 40.62% to 82.81% with an average of 66.11%. Web sites should conform to 80% of design guidelines to be considered usable [13], [27]. This index is used in the econometric model as an explanatory variable to examine the impact of usability of homepages on two reported bank performance measures ROE and ROA. B. Econometric Model A multiple regression modeling approach is chosen to explain the financial performance of community banks for the following reasons: (1) multiple regression analysis is robust and powerful analytical tool designed to explore all types of dependence relationships, (2) this approach is a straightforward dependence technique that can provide both prediction and explanation/description of the relationships among two or more intervally scaled variables, (3) a regression scheme can be used to examine the incremental and total explanatory power of many variables simultaneously [28], (4) this model design is consistent with existing literature that has applied multiple regression analysis to model the financial performance of banks [26], [29] – [31], and (5) a time series specification may have been better suited to model the performance trend of banks over the entire period of implementation of Web banking. However, the constraint that only a single data point is available for the usability index precludes the use of a timeseries econometric model. Two financial performance measures ROE and ROA are identified to be important in community banking [9], [26], [29], [30] – [34]. Most community banks are locally owned with a relatively high equity to assets ratio which implies that ROE is an important performance measure in evaluating a community bank. ROA is a standard measure of efficiency in utilization of the assets. Also, ROA and ROE are standard indicators of earnings for any firm. Specific models have been developed to ascertain the impact of select financial variables on these two financial performance measures. The usability index (Uindx) developed for each bank is used in the econometric model specifications for both the performance measures, ROE and ROA, as an explanatory variable of primary interest to predict and explain the effect of the usability of the homepages upon the bank’s performance. A community bank’s performance can be defined as a function of factors associated with assets, liabilities, employees, loan structure, interest margins, and the income expense structure. Therefore, the other explanatory variables used are business loans to total loans ratio (Blon), consumer
loans to total loans ratio (Clon), fixed assets to total asset ratio (Faset), equity capital to asset ratio (Eqcap), average rate of growth in assets (Agr), total assets per employee (Aemp), employment growth rate (Egr), noninterest income to total expenses ratio (NIex), level of inefficiency as measured by noninterest expenses to total revenue (NErev), liabilities to asset ratio (Last), and net interest margin (Imgn). These variables have been applied in prior studies [9], [26], [29], [32], [33], [35] – [37]. The models for ROE and ROA are hypothesized applying a subset of the above mentioned variables. The average asset growth rate and the average employment growth rate of the bank are calculated using the past three years’ data. The data for all the remaining financial variables have been downloaded for Q4 2005 from the FDIC Web site and used in estimating the models. The model for ROE is: Yi = ? + ?1Faset + ?2Blon + ?3Clon + ?4NIex + ?5NErev + ?6Last + ?7Aemp + ?8Imgn + ?9Agr + ?10Egr + ?11Uindx + ?i where Yi, ?, ?, and ?i represent the predicted ROE value, the intercept of regression line on yaxis, estimated regression coefficient, and error in prediction of ROE respectively. The model for ROA is: Yi = ? + ?1Faset + ?2Blon + ?3Clon + ?4Eqcap +?5NIex + ?6NErev + ?7Last + ?8Aemp + ?9Imgn + ?10Agr + ?11Egr + ?12Uindx + ?i where Yi, ?, ?, and ?i represent the predicted ROA value, the intercept of regression line on yaxis, estimated regression coefficient, and error in prediction of ROA respectively. The primary objective of this study, corresponding to part 2 of the problem statement, is to determine whether community banks providing more usable Web site homepages than their peers have a higher efficiency and profitability resulting from usability of Web banking in general and homepages in particular. The related hypothesis based upon the statistical significance of the model findings will attempt to measure both the magnitude and direction of the impact of usability index upon ROE and ROA. The secondary objective of this study, corresponding to part 1 of the problem statement, is to determine whether community banks having greater assets size provide more usable Web sites in general and in particular more usable homepages than their peers having lower asset size. For this, a simple regression model is proposed where asset size of the bank is used as explanatory variable and regressed against usability index. The related hypothesis based upon the statistical significance of the model findings will attempt to measure both the magnitude and direction of the impact of asset size upon usability index. The following section presents the results and discussions for the proposed models in detail.
VI. RESULTS AND DISCUSSION The descriptive statistics pertaining to all the financial variables and the usability index used in the econometric models are reported in Table 2. The results of the econometric model specifications with estimated parameters and corresponding tvalues are reported in Table 3. The regression output is provided in the Appendix.
TABLE 2: DESCRIPTIVE STATISTICS ON FINANCIAL VARIABLES USED IN ECONOMETRIC MODEL Sample average 14.8 1.3 1.8 24.9 8.1 27.4 62.8 91.1 3.7 4.0 13.6 7.8 8.9 66.1 Standard deviation 7.4 0.8 0.8 11.7 7.7 18.3 14.1 1.8 1.6 0.8 20.3 21.8 1.8 8.9
Variable description Return on equity Return on assets Fixed asset ratio Business loan to total loan Consumer loan to total loan Noninterest income over expenses Noninterest expense over revenue Liability asset ratio Assets per employee ($million) Net interest margin Asset growth rate Employment growth rate Equity asset ratio Usability index
Data Source: FDIC Online database (SDI) downloaded from http://www2.fdic.gov/sdi/main.asp and author’s estimates.
TABLE 3: REGRESSION RESULTS FOR RETURN ON EQUITY AND RETURN ON ASSETS
Return on equity Variable Coefficient Intercept Business loans to total loans Consumer loans to total loans Equity to assets Fixed assets to total assets Noninterest income to total expense Noninterest expenses to revenue Liabilities to assets Assets per employee Net interest margin Average asset growth rate 1.1310 0.2040** 0.2570** 1.3950** 0.1830 3.1560** 0.1720* 1.296 4.698 4.661 4.408 0.333 4.014 2.420 107.778** 0.0450 0.0170 t Value 3.570 0.930 0.204
Return on assets Coefficient 0.5900 0.0078 0.0086 0.0632* 0.1450 0.0245** 0.0250** t Value 0.620 1.670 1.044 2.082 1.735 5.882 4.745
0.0463 0.4130** 0.0210**
0.878 5.482 3.087
8
Average employment growth rate Usability index
0.1280 0.0740
1.882 1.187
0.0169** 0.0090
2.593 1.493
NOTE. *, ** Denote statistical significance at 5 and 1 percent levels respectively. A. ROE Model Results The econometric model for ROE has significant model fit. Some important results include the positive impact of non interest income to total expense ratio, the negative impact of noninterest expenses to revenue ratio (an inefficiency ratio), the positive impact of liabilities to assets ratio, and the positive impact of net interest margin which all show statistical significance at <1% level. This pvalue of <1% indicates that there is less than 1% probability for the null hypothesis, stating that the coefficients of these variables will be equal to zero, to occur. Also, the absolute t values, from the ttests of the null hypothesis that the coefficients of these variables are equal to zero in the population, are above the critical value of 3.291 thereby rejecting the null hypothesis and indicating that all these coefficients are statistically significant at <0.1% level. The impact of average asset growth rate is negative and statistically significant at the <5% level with an absolute tvalue 2.42 which is above the critical limit of 1.96 for 5% significance level. Usability index, however, shows a theoretically inexplicable and unexpected negative impact on ROE but statistically very insignificant with a pvalue of 0.241 indicating that there is 24.1% probability for the coefficient of usability index to be equal to zero which is heartening to see. The absolute tvalue for usability index is 1.87 which is below the critical limit of 1.96 for 5% significance level thereby accepting the null hypothesis that its coefficient is equal to 0. The coefficient of 2 determination (R ) for the model is 0.792 indicating that 79.2% of the variance in ROE is explained by the hypothesized regression model which is a good fit. The ANOVA test shows that F value is 15.96 which is above the critical value and pvalue is 0.000 indicating that there is <0.1% probability for the null hypothesis, the assumption that there is no true difference between the variables and any difference (statistically) is due to sampling errors, to occur. This indicates that the model variables as a group are a good predictor of ROE and hence the model is a good one. The variance inflation factor (VIF) is <10 for all the variables indicating that there is no significant collinearity problem between the model variables. B. ROA Model Results The econometric model for ROA has significant model fit. Some important results include the positive impact of non interest income to total expense ratio, the negative impact of noninterest expenses to revenue ratio (an inefficiency ratio), the positive impact of net interest margin, the negative impact of average asset growth rate, and the positive impact of 9
average employment growth rate which all show statistical significance at <1% level except average employment growth rate which is marginally close at 1.3% level. Also, the absolute t values for the coefficients of these variables are above the critical value of 3.291. The impact of equity to asset ratio is positive and statistically significant at the <5% level with an absolute tvalue 2.082. Usability index, again, shows unexpected negative impact on ROA but statistically very insignificant with a pvalue of 0.142 and an absolute tvalue of 1.493 which is again a welcome result. The liability to assets ratio has been excluded from the final model because it had high collinearity with equity capital to assets ratio owing to the fact that these two ratios are calculated using total assets figure and total assets equal liabilities plus equity. The 2 R for the model is 0.870 indicating that 87.0% of the variance in ROA is explained by the hypothesized regression model which is a good fit. The ANOVA test shows that F value is 27.89 which is above the critical value and pvalue is 0.000 or <0.1%. This indicates that the model variables as a group are a good predictor of ROA and hence the model is a good one. The variance inflation factor (VIF) is <10 for all the variables indicating that there is no significant collinearity problem between the model variables. The overall results, except usability index, are consistent with prior studies [9], [26], [33] providing validity to the proposed models. Therefore, the regression models are deemed to have acceptable fit indicating that the models predict and explain the performance measures satisfactorily. C. Hypotheses Results The variable of primary interest, the usability index, has a negative impact upon ROE and ROA. In particular, a one unit increase in the usability index would decrease the ROE by 0.074 units and decrease the ROA by 0.009 units. However, the effect of usability index on ROE and ROA is statistically very insignificant. These findings based on a subset of usability guidelines do not lend support to the hypothesis that community banks that provide better and more usable homepages than their peers have a higher efficiency and profitability resulting from the usability of their homepages. Pertaining to part 1 of the problem statement, the simple regression model for predicting usability index using asset 2 size as explanatory variable shows an R of 0.029 which is too low indicating that asset size is not a good predictor of usability index. Also, the hypothesis test of the coefficient for asset size show a tvalue of 1.294 which is below the critical limit and a pvalue of 0.201 which is unacceptable. The following section presents the summary and conclusions of this study and future direction of research. VII. CONCLUSIONS AND SUMMARY The managerial discussion (practical significance) of the conclusions drawn from the aforementioned results from the
econometric models will be discussed in this section followed by a brief summary of the overall research findings, a discussion of limitations, and suggestions for future research. A. Conclusions Some significant conclusions are drawn from the econometric models for ROE and ROA with respect to various explanatory variables. Many of the findings here are consistent with our previous study [3] – [6] and other existing literature. However, the conclusions for usability Index, the variable of primary interest, differ from the conclusions drawn on Web banking index in our previous study. 1) Usability Index: Our previous work found that, online banking index, a measure of the Web features, products and services provided on the bank Web sites, had a significant positive impact on ROE (p ? 0.001) and ROA (p ? 0.005) implying that online banking applications help increase a bank’s earnings. Banks that provide more effective Web sites and a wider array of online products and services than their peers appear to have higher efficiency and profitability. As all financial transactions are digitized, banks can easily access and analyze this information (both bank information and credit score information) rapidly and more effectively to identify underperforming and nonperforming assets as well as target better sources of revenues. However, in this research study, the usability index has a very insignificant negative impact on ROE and ROA implying that more usable homepages (Web sites) decrease the earnings which is not possible. Even if banks overspend on usability of their IT initiatives in their drive to provide more usable Web applications, the inefficiency ratio should reflect this problem rather than the usability index. The statistical insignificance of this negative impact is a blessing in disguise and a welcome result that does not reject the theory that better usability of Web applications should increase efficiency and profitability of banks. However, while comparing the results of the previous study with this study, one has to note that in the earlier study entire Web sites were evaluated for Web features, products, and services. This study evaluates neither the usability of entire Web sites nor the usability of the specific homepagerelated Web features, products, and services evaluated in the previous study alone. These two projects are as much independent as they are related conceptually. A body of literature pertaining to the period 1998 to 2003 [8] – [10] shows that though bank profitability is strongly correlated with Internet banking, the impact of Internet banking on bank profitability is not statistically significant. Apparently banks that adopted Internet banking already had higher profitability, accounting efficiency, and scale than other banks. However, by the end of 2005 the impact of Internet banking on bank profitability is found to be statistically significant as reported by our previous work [3] – [6] probably because rest of the community banks sensed
competition and opportunity for Internet banking and caught up with the early birds in offering Webbased services. Now, this study found that bank profitability is correlated with usability of Internet banking Web sites, but the impact of usability on bank profitability is not statistically significant. It could be the case that Web site usability in community banking sector is not mature enough yet to have significant impact on performance. However, on cross examining the data it is found that more efficient banks tend to maintain quality Web sites that have higher usability index though higher usability index is not leading to higher efficiency. This clearly indicates a potential trend where it appears that more efficient banks with scale are the first ones to adopt Internet banking and they are also the first ones to ensure the quality of their Web offerings by emphasizing on usability of their Web sites. Following the trend, one could expect in 2 to 3 years from now, less efficient banks will feel compelled by the competition to make their Web sites more usable at which point Web usability might show significant impact on bank profitability. Some of the limitations in the study might explain this unexpected negative and insignificant impact of usability index which differs from the positive and significant impact of the online banking index observed in the earlier study. The sample size of 58 used in this study, as opposed to 1271 used in the previous study, is very small for a powerful modeling technique like regression analysis that generally requires larger sample size especially for multiple regression analysis thus making the results less reliable. The asset category of $500 million to <$1billion used as a criteria to select the sample may have made a difference. The homepages are evaluated against a small set 32 homepage usability design guidelines out of a total of 113 guidelines available. Evaluating the remaining guidelines as well will provide a truer picture of the impact of the homepage usability on performance. The procedure used to develop the usability index in this study is not a sophisticated one capable of providing a measure of usability or user experience of the homepages that is capable of capturing the abstract nature of the impact of usability of homepages or Web sites as a whole on the performance of banks. Indirect measurement techniques or latency models like structural equation modeling may be better suited to develop a measure for the unobservable effects of Web usability that can then be used in regression models against performance measures. Moreover, homepage itself is a subset of the Web site. Data for some of the variables for the sample banks show skewness which might account to some extent for the unexpected results. More importantly, many of the community banks may neither have sufficient financial resources nor the critical mass of potential online customers to justify a reasonable budget for the usability of their Web banking initiatives. Pertaining to part 1 of the problem statement, the simple regression model for predicting usability index using asset
10
size as explanatory variable shows that asset size is not a good predictor of usability index. This concludes that banks with greater asset size need not necessarily build more usable Web sites. Again, the sample size, the asset size of the banks in the sample, the small subset of guidelines used to evaluate the homepages, and the procedure used for developing usability index may have influenced this result. B. Investigating the Joint Impact/Relationship of Web Features Index and Usability Index on Performance The previous study on Web features, products, and services showed that Web features and products/services do matter and they have significant impact on the bank performance. However, this study shows that usability does not matter or that usability has no significant impact on the bank performance though the limitations in this study and the previous study are differing. But existing literature pertaining to general usability studies and return on investment (ROI) studies points out that the Web features and their usability are closely related concepts and that the usability of Web sites does have significant impact on revenue generation. Though this may not imply that usability has an impact on efficiency of revenue generation which is different from simple measure of revenue generation, it is however still important to combine the two aspects of Web features and their usability and conduct a postmortem analysis to investigate the joint impact of Web features and usability on the performance of banks. First, the sample banks have been segregated into high, medium, and low categories based on their Web features/products/services scores from previous study and usability index score in this study. The natural data breaks in the sample data and even distribution of subsample sizes are the two criteria employed in segregating the sample banks. Then, with the high, medium, and low categories of usability index on xaxis and with the high, medium, and low categories of Web features/products/services categories on y axis, a cross tabulation matrix containing 9 cells is plotted using SPSS as shown in section C of Appendix. Of these 9 cells, the cell containing banks that are high on both usability and Web features (which is termed highhigh cell) and the cell containing banks that are low on both usability and Web features (which is termed lowlow cell) are of importance to test the joint impact of Web features and usability on bank performance. The highhigh cell has a sample size of 8 banks and the lowlow cell has a sample size of 7 banks. The same performance measures ROE and ROA, which are modeled in the model development section, are chosen again for analysis. The mean ROE and ROA for the banks in the highhigh cell and the lowlow cell are calculated separately after removing the outlying data points. The mean ROE and ROA for the entire sample of 58 banks are also calculated after removing the outlying data points. The pooled variance ttest for testing the statistical significance of the difference in
two means is employed to check if the averages of the performance measures of banks in highhigh cell are significantly different from the averages of respective performance measures of banks in lowlow cell. To make the tests more robust, the ttests are conducted 3way: (1) between means of performance measures of banks in highhigh cell and lowlow cell, (2) between means of performance measures of banks in highhigh cell and overall sample of 58, and (3) between means of performance measures of banks in lowlow cell and overall sample of 58. The null hypothesis in all these tests states that the means of the performance measures of the populations of highhigh and lowlow banks are equal failing to reject which we cannot prove that banks in highhigh cell perform significantly better than banks in lowlow cell which in turn means that banks in highhigh cell are no more efficient than banks in lowlow cell. The average ROE for overall sample, highhigh cell, and lowlow cell are 14.35, 14.68, and 13.42 respectively which means that the performance of banks in highhigh cell is better than the performance of overall sample of 58 banks which in turn is better than the performance of banks in low low cell. This is in line with theoretical expectation of higher efficiency of banks with better Web banking features and also usability. However, all the 3way ttests failed to reject the null hypothesis showing that the difference in the means of performance of these 3 groups in not significant. The average ROA for overall sample, highhigh cell, and lowlow cell are 1.30, 1.23, and 1.46 respectively which means that the performance of banks in lowlow cell is better than the performance of overall sample of 58 banks which in turn is better than the performance of banks in highhigh cell. But, this is not in line with the theoretical expectation of higher efficiency for banks with better Web banking features and also usability. The potential explanation for this unexpected result is that banks in highhigh cell may have significantly invested in assets for IT applications and these assets are tiedup and cannot produce any immediate returns. It is a fact that investments in IT applications will not give returns in the short term. In such a scenario, this unexpected result is bound to happen with any asset based performance measure such as ROA. However, again all the 3way ttests failed to reject the null hypothesis showing that the difference in the means of performance of these 3 groups in not significant. This clearly shows that the ROA of banks in highhigh cell is not significantly lower than ROA of banks in lowlow cell which is an expected result. The overall conclusion is that assetbased performance measures appear to be insignificantly lower for banks with high Web features and high usability compared to banks with low Web features and low usability while nonasset based performance measures appear to be higher for banks with high Web features and high usability compared to banks with low Web features and low usability probably because of higher investments in IT assets by banks that have high Web
11
features and high usability. We know that all these banks had Web sites from 2003 though we do not precisely know how long they had Web sites before 2003. Probably the ROA is lower for banks in highhigh cell because they have not recovered their investments in IT applications as yet and they may require more time to realize the performance efficiency benefits of OB applications. This trend of the results based on ROE and ROA, and the rejection of the null hypotheses in all the 6 ttests indicates that the banks, especially those in high high cell, may be in a twilight stage in realizing the benefits of Web banking and the usability of Web banking. This prediction is based on the fact that ROE, a measure closely related to assetbased performance measures, though not a direct assetbased performance measure in itself, is showing a higher performance for banks in highhigh cell compared to those in lowlow cell while ROA, a direct assetbased performance measure, is showing an insignificant lower performance for banks in highhigh cell compared to those in lowlow cell. The implications of these two findings have to be read in conjunction rather than in an insulated manner. However, it has to be noted that the insignificant results and the less precise conclusions are because of the small sample of banks in the highhigh and lowlow cells of the cross tabulation matrix and the overall sample size. Future research endeavors must make sure of sufficient sample size. C. Summary This research analyzed the impact of usability of community bank Web site homepages on the performance of community banks that are implementing electronic commerce with Internet banking. A usability index is developed as a measure of compliance to the homepage usability design guidelines to proxy the usability of homepages and this index is included in econometric models as an explanatory variable to examine whether it explains differences in community bank performance. Based on the sample, the results show that (1) the asset size of banks has no impact on the usability of the homepages of Web sites, and (2) the impact of usability of bank Web site homepages on bank performance is insignificant as a result of which no meaningful and reliable conclusions can be drawn. It could be the case that Web site usability in community banking sector is not mature enough yet to have significant impact on performance. However, it is also found that more efficient banks tend to maintain quality Web sites that have higher usability index though higher usability is not leading to higher efficiency. An important point to note is that usability is measured relative to certain users and certain tasks. Usability measurement therefore starts with the definition of a representative set of test tasks, relative to which the different usability attributes can be measured. Certain tradeoffs are involved in designing Web sites for usability simultaneously for both novice and expert users and hence usability is in a way a contradiction in itself [11].
D. Future Direction Of Research As community banks continue to implement EC applications, the usability of these applications cannot be ignored if community banks were to withstand the aggressive online competition from competitors in the market which is only bound to increase in the future as almost all banks have gone online and remaining ones are also following suit to catch up the online market. This leads to many future usability research endeavors in the banking sector beyond this project. To better study the effect of the usability of the homepages on financial performance of the banks, it would be interesting to evaluate the homepages with all the 113 homepage usability design guidelines and 40 homepage usability design statistics proposed in [13] based on extensive user testing. However, homepage is only a part of Web site and a heuristic evaluation of the entire Web sites for usability will be more reliable to assess the true and comprehensive impact of the usability of Web activities of banks on their financial performance. REFERENCES
[1] H.Beck, “Banking is essential, banks are not. The future of financial intermediation in the age of the Internet,” Netnomics, vol. 3 (1), pp. 722, 2001. [2] B. Foster, “Do or die time for community banks?” The RMA Journal, pp. 15–19, April 2001. [3] R. N. Acharya, A. Kagan, and S. R. Lingam, “Online banking and profit efficiency: Implications for community banking,” Journal of Money, Credit, and Banking, submitted for publication. [4] A. Kagan, R. N. Acharya, and S. R. Lingam, “Electronic business applications within community banks: A modeling approach,” presented at the Annual Meeting of the Decision Sciences Institute, San Francisco, California, 2005. [5] A. Kagan, R. N. Acharya, L. S. Rao, and V. Kodepaka, “Does Internet banking affect the performance of community banks?” presented at the Annual Meeting of the American Agricultural Economics Association, Providence, RI, 2005. [6] S. R. Lingam, “Performance aspects of community banks implementing electronic commerce with Internet banking,” M.S. thesis, Department of Agribusiness and Resource Management, Arizona State University, Tempe, 2005. [7] K. Furst, W. W. Lang, and D. E. Nolle, “Internet banking: Developments and prospects,” Economic and Policy Analysis Working Paper, Office of the Comptroller of the Currency, pp. 158, September 2000. [8] K. L. Egland, K. Furst, D. E. Nolle, and D. Robertson, “Banking over the Internet,” Quarterly Journal, Office of the Comptroller of the Currency, vol. 17, no. 4, pp. 2530, December 1998. [9] K. Furst, W. W. Lang, and D. E. Nolle, “Internet banking: Developments and prospects,” Program on Information Resources Policy (Jointly sponsored by Harvard University and Center for Information Policy Research), pp. 153, April 2002. [10] A. N. Berger, “The economic effects of technological progress: Evidence from the banking industry,” Journal of Money, Credit and Banking, vol. 35, no. 2, pp. 141176, 2003. [11] J. Nielsen, Usability Engineering. San Diego, California: Academic Press, 1993. [12] J. Nielsen, R. Molich, C. Snyder, and S. Farrell, ECommerce User Experience. Fremont, California: Nielsen Norman Group, 2001. [13] J. Nielsen and M. Tahir, Homepage Usability: 50 Websites Deconstructed. Indianapolis, Indiana: New Riders, 2002. [14] J. Nielsen and H. Loranger, Prioritizing Web Usability. Indianapolis, Indiana: New Riders, 2006. [15] J. Nielsen, Designing Web Usability: The Practice of Simplicity. Indianapolis, Indiana: New Riders, 2000.
12
[16] J. Nielsen. (1998, April 5). Nielsen’s Law of Internet bandwidth. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/980405.html [17] J. Nielsen. (2004, September 13). The need for Web design standards. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/20040913.html [18] J. Nielsen. (2000, August 6). Why Doc Searls doesn’t sell any books. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/20000806.html [19] J. Nielsen. (2000, February 6). Novice vs. expert users. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/20000206.html [20] A. Kagan, G. Post, and C. M. Noel, “Selling the Web: Web features used by retailers,” The Journal of Applied Business Research, vol. 16, no. 1, pp. 1526, 2000. [21] R. DeYoung, W. C. Hunter, and G. F. Udell, “Whither the community bank?” Journal of Financial Services Research, vol. 25:2/3, pp. 8184, 2004. [22] R. DeYoung, W. C. Hunter, and G. F. Udell, “The past, present, and probable future for community banks,” Journal of Financial Services Research, vol. 25:2/3, pp. 85133, 2004. [23] L. I. Nakamura, “Small Borrowers and the Survival of the Small Bank: Is Mouse Bank Mighty or Mickey?” Business Review, Federal Reserve Bank of Philadelphia, pp. 315, November/December 1994. [24] J. Nielsen. (2006, January 23). Ten best intranets of 2006. Alertbox: Current Issues in Web Usability. [Online]. Available: http://www.useit.com/alertbox/intranet_design.html [25] R. DeYoung, and I. Hasan, “The performance of de novo commercial banks: A profit efficiency approach,” Journal of Banking & Finance, vol. 22, pp. 565587, 1998. [26] K. J. Stiroh, “Do community banks benefit from diversification?” Journal of Financial Services Research, vol. 25:2/3, pp. 135160, 2004.
[27] V. Flanders and D. Peters, Son of Web Pages That Suck: Learn Good Design by Looking at Bad Design. Alameda, California: Sybex, 2002. [28] J. F. Hair, Jr., R. E. Anderson, R. L. Tatham, and W. C. Black, Multivariate Data Analysis, 5th ed. Upper Saddle River, New Jersey: Prentice Hall, 1998. [29] A. N. Berger and L. J. Mester, “Inside the black box: What explains differences in the efficiencies of financial institutions?” Journal of Banking & Finance, vol. 21, pp. 895947, 1997. [30] A. Akhigbe and J. E. McNulty, “The profit efficiency of small US commercial banks,” Journal of Banking & Finance, vol. 27, pp. 307325, 2003. [31] T. J. Yeager, “The demise of community banks? Local economic shocks are not to blame,” Journal of Banking & Finance, vol. xxx, pp. xxxxxx, 2003. [32] Gary C. Zimmerman, “California’s community banks in the 1990s,” Weekly Letter, Federal Reserve Bank of San Fancisco, no. 9604, January 26, 1996a. [33] Gary C. Zimmerman, “Factors influencing community bank performance in California,” Economic Review, Federal Reserve Bank of San Fancisco, no. 1, pp. 2642, 1996b. [34] A. P. Meyer and T. J. Yeager, “Are small rural banks vulnerable to local economic downturns?” Review, Federal Reserve Bank of ST. Louis, pp. 2538, March/April 2001. [35] A. N. Berger and G. F. Udell, “Relationship lending and lines of credit in small firm finance,” The Journal of Business, vol. 68,3, pp. 351381, July 1995. [36] P. E. Strahan and J. P. Weston, “Small business lending and the changing structure of the banking industry,” Journal of Banking & Finance, vol. 22, pp. 821845, 1998. [37] A. N. Berger, A. Saunders, J. M. Scalise, and G. F. Udell, “The effects of bank mergers and acquisitions on small business lending,” Journal of Financial Economics, vol. 50, pp. 187229, 1998.
APPENDIX SPSS output tables for the hypothesized econometric models. A. Output for part 1 of the problem statement
Model Summary Model 1 R R Square a .170 .029 Adjusted R Square .012 Std. Error of the Estimate 8.91589687
a. Predictors: (Constant), Total Assets
b ANOVA
Model 1
Regression Residual Total
Sum of Squares 133.130 4451.620 4584.750
df 1 56 57
Mean Square 133.130 79.493
F 1.675
Sig. a .201
a. Predictors: (Constant), Total Assets b. Dependent Variable: Usability Index
13
a Coefficients
Model 1
Unstandardized Coefficients B Std. Error (Constant) 58.499 5.997 Total Assets 1.109E05 .000
Standardized Coefficients Beta .170
t 9.756 1.294
Sig. .000 .201
a. Dependent Variable: Usability Index
B. Output for part 1 of the problem statement ROE Model:
b Model Summary
Model 1
R R Square a .890 .792
Adjusted R Square .743
Std. Error of the Estimate **********
a. Predictors: (Constant), UINDX, BLON, IMGN, CLON, AGR, FASET, LAST, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROE
b ANOVA
Model 1
Regression Residual Total
Sum of Squares 2489.825 652.378 3142.203
df 11 46 57
Mean Square 226.348 14.182
F 15.960
Sig. a .000
a. Predictors: (Constant), UINDX, BLON, IMGN, CLON, AGR, FASET, LAST, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROE
14
a Coefficients
Model 1
Unstandardized Coefficients B Std. Error (Constant) 107.778 30.200 FASET 1.131 .873 BLON 4.506E02 .049 CLON .017 .085 NIEX .204 .043 NEREV .257 .055 LAST 1.395 .316 AEMP .183 .549 IMGN 3.156 .786 AGR .172 .071 EGR .128 .068 UINDX .074 .063
Standardized Coefficients Beta .127 .071 .018 .505 .488 .344 .040 .337 .472 .376 .090
t 3.569 1.296 .928 .204 4.698 4.661 4.408 .333 4.014 2.420 1.882 1.187
Sig. .001 .201 .358 .839 .000 .000 .000 .741 .000 .020 .066 .241
Collinearity Statistics Tolerance VIF .471 .767 .570 .391 .412 .741 .319 .639 .119 .113 .787 2.125 1.304 1.756 2.555 2.428 1.350 3.138 1.566 8.412 8.832 1.270
a. Dependent Variable: ROE
ROA Model:
b Model Summary
Model 1
R R Square a .933 .870
Adjusted R Square .838
Std. Error of the Estimate **********
a. Predictors: (Constant), EQCAP, IMGN, CLON, FASET, AGR, UINDX, BLON, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROA
b ANOVA
Model 1
Regression Residual Total
Sum of Squares 40.036 6.003 46.039
df 11 46 57
Mean Square 3.640 .131
F 27.890
Sig. a .000
a. Predictors: (Constant), EQCAP, IMGN, CLON, FASET, AGR, UINDX, BLON, NEREV, NIEX, AEMP, EGR b. Dependent Variable: ROA
15
a Coefficients
Model 1
Unstandardized Coefficients B Std. Error (Constant) .590 .952 FASET .145 .084 BLON 7.779E03 .005 CLON 8.561E03 .008 NIEX 2.453E02 .004 NEREV .025 .005 AEMP 4.626E02 .053 IMGN .413 .075 AGR .021 .007 EGR 1.692E02 .007 UINDX .009 .006 EQCAP 6.318E02 .030
Standardized Coefficients Beta .135 .102 .074 .501 .394 .083 .365 .477 .410 .090 .129
t .620 1.735 1.670 1.044 5.882 4.745 .878 5.482 3.087 2.593 1.493 2.082
Sig. .538 .089 .102 .302 .000 .000 .384 .000 .003 .013 .142 .043
Collinearity Statistics Tolerance VIF .471 .767 .570 .391 .412 .319 .639 .119 .113 .787 .741 2.125 1.304 1.756 2.555 2.428 3.138 1.566 8.412 8.832 1.270 1.350
a. Dependent Variable: ROA
C. Cross Tabulation Chart Comparing Usability Index and Web Features (Banking) Index:
Crosstabulation of Usability Index and Banking Index Usability Index 1.00 2.00 7 4 50.0% 28.6% 10 6 37.0% 22.2% 2 7 11.8% 41.2% 19 17 32.8% 29.3%
3.00 3 21.4% 11 40.7% 8 47.1% 22 37.9%
Banking Index
1.00 2.00 3.00
Total
Count % within Recoded Bindx Count % within Recoded Bindx Count % within Recoded Bindx Count % within Recoded Bindx
Total 14 100.0% 27 100.0% 17 100.0% 58 100.0%
Note: Codes 1.00, 2.00, and 3.00 for Usability Index (Uindx) and Banking Index (Bindx) correspond to low, medium, and high values respectively for the concerned index.
16
doc_819536545.pdf