A Content Analysis of Registration Processes on Websites: How Advertisers Gather Information to Customize Marketing Communications
Jan Ahrens
San Francisco State University
James R. Coyle
Miami University
Abstract
The proper implementation and design of registration pages is a crucial consideration in the development of websites for two reasons: First, users often navigate through them to receive future marketing communications. Second, advertisers can gain valuable contact information through registration processes that allow them to customize marketing communications. This content analysis investigates the implementation and design of such processes by comparing the registration processes established by large and small websites, as well as websites of pure play companies versus click-and-mortar companies. Differences emerge across several variables. The results have implications for advertising researchers and practitioners and suggest some registration process best practices.
Introduction
For advertisers, one of the most significant outcomes of interactivity is that it allows for advertising customization (Bozios et al. 2000). Customized or personalized advertising has the potential to be more cost efficient and effective than mass advertising, because the communications are tailored to individual users (Pavlou and Stewart 2000). The information necessary for this customization often is gathered through web forms. Thus, the proper implementation and design of web forms (e.g., registration, checkout) is a crucial consideration for the development of websites and can have a great impact on business goals (Wroblewski 2008). Efficient registration processes and the forms they consist of are particularly important for two specific reasons: (1) Users often must navigate through them to receive future marketing communications, and (2) advertisers can gain valuable contact information through registration processes that in turn allows them to customize marketing communications. Major online companies such as eBay consider increasing the number of registered users a key financial lever and thus spend a considerable amount of time ensuring that the registration process is as user-friendly as possible (Herman 2004). The factors that may be relevant to the effectiveness of a registration process are many, including its usability, location on the page, and incentives to register.
A registration process is typically an initial invitation to online users to engage in permission marketing, which implies a relationship (Gaffney and Szuc 2006; Krishnamurthy 2001). To gain the benefits they can earn through registration, consumers agree to provide personal information. These benefits might take the form of opt-in e-mail communication and discounts, for example. For advertisers, the exchange offers an opportunity to ask for important pieces of information that reveal users’ preferences, which is key to relationship management (Gaffney and Szuc 2006; Marinova, Murphy, and Massey 2002), and enable them to tailor subsequent messages (Bozios et al. 2000; Ha and James 1998; Pavlou and Stewart 2000). Moreover, a user-friendly registration process can contribute to user satisfaction with the site (Herman 2004).
Two areas of research are relevant to the study of how registration processes can be implemented and designed most effectively to enable advertisers to collect key pieces of information and customize their marketing communications. First, exchange theory (Houston and Gassenheimer 1987) suggests that certain conditions must be met before successful registration will occur. To persuade visitors to register, site managers face at least two challenges: provide adequate incentive for visitors to share personal information and convey a level of trustworthiness, such that visitors feel comfortable sharing personal information (Gaffney and Szuc 2006). These two factors together predict whether a successful exchange occurs. For this study, exchange theory predicts the kinds of strategies that may be most effective in persuading site visitors to register, the different ways registration processes can be structured, and the kinds and amount of information that is appropriate to gather at the point of registration. Second, by increasing the levels of interactivity in Internet communications, site managers capitalize on the unique potential that the medium offers (Coyle and Thorson 2001; Johnson, Bruner, and Kumar 2006). Human-computer interaction (HCI) theory suggests that information collection is an important dimension of interactivity and plays a part in the registration process as information is collected from site visitors (Ha and James 1998). In general, information collection is a crucial dimension on which successful Internet companies compete (Evans and Wurster 1999), and interactivity in turn provides advertisers with valuable opportunities for customization (Bozios et al. 2000; Pavlou and Stewart 2000). Therefore, customized advertising can allow consumers to feel more engaged (Lombard and Snyder-Duch 2001).
Despite the importance of proper registration form designs, most are flawed, typically because the organization considers a registration form a database of fields that need to be completed, rather than an invitation to join a site or community that users should be able to complete quickly and easily (Wroblewski 2008). We consider registration process best practices within the context of exchange theory and HCI theory by comparing and contrasting incentive strategies and registration process structures across e-commerce sites of different sizes and of two different business models (i.e., pure play and click-and-mortar). These differences likely affect how companies implement and design their registration processes to gather information.
Regarding the potential differences across business models, Evans and Wurster (1999) discuss information collection practices among different types of e-commerce companies and note that physical retailers looking to compete online have a natural advantage over e-retailers, because of their offline experience building, maintaining, and mining rich information databases. This experience is unique to click-and-mortar businesses and a source of differentiation among them, pure Internet, and traditional firms (Steinfield 2002). However, developing an integrated information architecture that must benefit from online and offline synergies is complex and requires long-term planning (Saeed, Grover, and Hwang 2003). Thus we investigate whether the supposed synergistic advantages that click-and-mortar businesses enjoy, compared with pure play companies, actually exist.
Regarding differences in firm size, larger companies tend to integrate the Internet more quickly into their business strategy than smaller companies (Teo and Pian 2003), likely because they have more resources than smaller companies to develop fully featured websites (Perry and Bodkin 2000). This explanation accounts for differences in the incorporation of interactive website features among companies of different sizes in the transportation industry (Ellinger, Lynch, and Hansen 2003). Similarly, an early study of e-commerce adoption by local merchants (Steinfield, Mahler, and Bauer 1999) reveals that small companies are unprepared to take full advantage of e-commerce and unlikely to see any gains from it, in part because well-established companies can better capture and use market-relevant data. Perhaps companies of different sizes implement and design registration processes differently because of the investments required to develop and manage such processes.
Finally, we note that content analysis methodology has been applied infrequently to the study of Internet marketing issues, despite being singled out as a particularly important tool for understanding website-related issues, such as how sites evolve over time (Parasuraman and Zinkhan 2002). Content analyses conducted in the field tend to use websites as the unit of analysis, focusing on communication strategies in general (Perry and Bodkin 2000) and cross-cultural communication strategies in particular (Kim, Coyle, and Gould 2009; Okazaki and Rivas 2002; Wurtz 2005). Ha and James’s (1998) content analysis also contains some aspects of registration processes, as we discuss in greater depth in the literature review section.
Overall, the purpose of our study is to conduct a content analysis to investigate the current state of commercial website registration processes, in terms of their implementation and design. Furthermore, we compare how the registration processes of large and small sites and the sites of pure play companies and click-and-mortar companies may vary.
Literature Review
Registration as a Form of Interactive Exchange
In their investigation of levels of interactivity in company-sponsored websites, Ha and James (1998) consider the prevalence of information collection and reciprocal communication functionality in the first half of 1996. This early study included two instruments for the operationalization of information collection: registration mechanisms and counters displaying the number of visitors to a site. At the time, 110 websites conducted business on the Internet, and Ha and James included all of them in their analysis. Only 14% included a registration procedure, and for them, registration prior to access to the sites was twice as frequent as registration during site visits. They consider this trend unsurprising, “because after admission to a site, consumers have little incentive to reveal their identity or disclose other information” (Ha and James 1998, p. 468). This early picture shows that the first step in facilitating the exchange of information online between marketer and consumer (i.e., registration procedures) rarely existed, and the typical implementation of registration procedures (i.e., before a visitor could enter sites) was primitive. The researchers also call for a longitudinal study of the dimensions of interactivity, including information gathering through registration, in company-sponsored websites. We grasp this opportunity to examine, with a longitudinal perspective, whether registration process have evolved to become important parts of a company’s information-gathering practices. Specifically, we address the following research question:
RQ1: How prevalent are registration processes in company-sponsored websites?
In determining the conditions necessary for an exchange, Blalock and Wilken (1979) identify five, three of which are relevant to registration processes:
- Humans basically are goal-seeking animals.
- Humans are able to anticipate the consequences of their actions.
- Humans direct their behaviors toward their preferred anticipated consequences.
In general, users who encounter a website that conforms to their expectations demonstrate increased intentions to return to the site (Coyle and Gould 2002; Dellaert and Kahn 1999). Advertisers then need to understand the goals and expectations that site visitors have, regarding what they can achieve and what they will receive for registering (Gaffney and Szuc 2006). Visitors who are satisfied with the incentive should be more likely to complete the registration process. For example, in an experiment comparing two incentive strategies (Gamberini et al. 2007), participants either were required to complete a form that asked for contact and demographic information before they gained access to site content (reward strategy) or were granted access to site content before completing the form (reciprocity strategy). The reciprocity strategy was more effective than the reward strategy in persuading participants to complete the form. Although research in the 1990s speculated that company websites rarely included registration processes because companies did not give site visitors an incentive to share their information (Ha and James 1998), advertisers today should have learned the value of incentives for persuading site visitors to register. Therefore, we investigate whether the inclusion of incentives, which may include a consumer’s expectation of the exchange that will take place during registration, represents a best practice in modern registration process designs and implementation approaches. Accordingly, we examine the following research question:
RQ2: How prevalent are incentives offered to persuade site visitors to register?
Of course, regardless of an incentive strategy, the extent of information consumers are willing to share is an important consideration (Pavlou and Stewart 2000). The kinds and number of questions that make up the registration process can have a significant influence on whether site visitors choose to complete the process. A process with too many or irrelevant questions may discourage site visitors from completing their goals (Gaffney and Szuc 2006). In general, advertisers should study each potential question in a form to ensure the information requested is necessary, cannot be retrieved some other way, or would be more sensible to gather at some other time or place (Wroblewski 2008). Considering the importance of both the quantity and type of information gathered, we develop two further research questions:
RQ3: What kinds of information are gathered through registration processes?
RQ4: How much information is gathered through registration processes?
Another challenge for advertisers is to identify when consumers are most likely to wish to be engaged (Pavlou and Stewart 2000). Designers of online processes must identify opportune times to present registration opportunities to visitors, in ways that contribute to their experience and are not overly disruptive (Bailey, Konstan, and Carlis 2001). By identifying the most strategic places to ask visitors to register, designers can develop processes that significantly reduce annoyance and frustration, are considered respectful, and require less mental effort (Adamczyk and Bailey 2004). Best practices for questionnaire designs for survey research also address the need to organize questions into topic areas and thus facilitate processing by the research participant (Malhotra 2006). Similarly, registration forms should be organized into content groups, whether as sections of a single webpage or across multiple webpages, each of which consists of a set of questions pertaining to a particular topic (Wroblewski 2008). When site visitors begin to register, the organization should be consistent with their schema related to how they expect to encounter sets of questions. This congruence between a consumer’s existing schema for a company’s website and the structure of the site offers a positive predictor of perceived ease of navigation, attitude toward the brand, and the quality of brand decisions made on the site (Bellman and Rossiter 2004). Similarly, a consumer’s expectation of the structure of a website registration process can influence that consumer’s willingness to complete the process. Thus, it is important to understand how online retailers structure their registration processes, and we address the following research question:
RQ5: How are registration processes structured?
An equally important aspect of the value proposition inherent to the registration process is whether site visitors experience sufficient trust in the company (Ha and James 1998). Trust is a key influence of customer purchase decisions on e-commerce sites (Cho 2006; Garbarino and Maxwell 2009). Whether trust gets established at registration pages may be crucial. For example, one survey indicates that the top reason respondents do not fill out registration forms is because the sites do not provide information about how the data will be used (Georgia Tech 1998). To establish trust, sites might explain why the information is needed and how it will be used (Gaffney and Szuc 2006). This argument suggests that including this explanation as part of an online registration process should increase the likelihood that consumers share information and prompts the following research question:
RQ6: How frequently do registration processes include statements about why the information being asked for is needed and how it will be used?
Our last question focuses on what companies do after site visitors register. More than 10 years ago, Ha and James (1998) identified information collection and reciprocal communication as important dimensions of interactivity. In more recent research on facets of interactivity, Johnson, Bruner, and Kumar (2006, p. 40) have described reciprocity as a situation in which consumers have “the opportunity to participate jointly and convers[e] with firms as opposed to hearing a monologue from them.” Thus, the way companies respond to registered site visitors is another vital component of interactivity. Because the depth and detail of information a business can collect from and provide to visitors online can predict its online success (Evans and Wurster 1999), we attempt to learn whether companies go beyond simply requesting consumer information and subsequently respond to consumers according to the information they provide, with the following research question:
RQ7: How frequently do companies respond to site visitors who register?
We use research questions, rather than hypotheses, because only one previous study (Ha and James 1998) has considered the characteristics of registration processes, and this study was conducted when e-commerce was in its infancy. To explore all of the research questions more deeply, we have established comparisons across companies of different sizes and types. Although there are many reasons differences may emerge in these comparisons (e.g., marketing budget sizes, business strategy considerations), there is insufficient research in this area to support hypotheses development.
Method
Sampling Frame and Sample
The sample includes 200 companies’ websites. To capture companies of different sizes, the sample features the top 100 retail websites from Internet Retailer’s 2008 “Top 500 Guide.” The 2007 web-based sales for these companies ranged from $140 million to $14.8 billion. For smaller companies, we selected the 2008 top 25 fastest growing companies in four categories (i.e., consumer products, computers and electronics, financial services, and retail), according to Inc.com (www.inc.com/inc5000/2008/index.html). These four categories closely mirror the categories represented by the larger companies. Only 2 of the 100 companies from the Inc.com sample generated 2008 revenues greater than $140 million, which suggests that this sample includes significantly smaller companies than the “Top 500 Guide” sample. For the comparisons between large and small companies, we moved these exceptions to the large-company sample.
For the comparisons between types of companies, we classified the sample into two groups: click-and-mortar companies (with both an online and an offline presence) and pure play companies (operating exclusively online). This distinction is especially interesting in relation to registration processes for two reasons. First, click-and-mortar companies can use registration processes to their advantage in cross-channel communications, in that they use the information gathered during registration to generate communications that can drive consumers to different channels. Second, for pure play companies, the information gathered during registration may be the primary way to learn how to communicate with current and potential customers. The sample contains 137 (69.2%) pure play companies and 61 (30.8%) click-and-mortar companies.
Content Variables
We coded registration processes according to several dimensions. First, we noted whether an incentive was offered to site visitors to register. Second, we determined whether the following types of information were captured in the registration process: name, e-mail address, mailing address, birth date, and gender. Third, we counted the number of steps to complete the registration process. A step was defined as an action that submitted information, such as clicking a button labeled “submit” or “continue” that would complete the registration process or continue it by linking to a subsequent step. Fourth, we determined whether information provided during the registration process clarified what the sponsoring company would do with the gathered data. Fifth, we counted the total number of e-mails sent subsequent to the registration over a 15-17-week period. To gather all this information, we evaluated the registration response, using an e-mail address created specifically to register at the site. In Table 1, we provide each research question and the coding variables used in the analyses to evaluate it.
Table 1
Summary of Research Questions and Associated Coding Variables
| RQ# | Research Question | Coding Variable(s) |
| RQ1 | How prevalent are registration processes in company-sponsored websites? | For all 200 companies, the presence of one or more registration opportunities on the home page. |
| RQ2 | How prevalent are incentives offered to persuade site visitors to register? | For companies with at least one registration opportunity, the presence of any incentive to register (e.g., free newsletters). |
| RQ3 | What kinds of information are gathered through registration processes? | For companies with at least one registration opportunity, each of the following variables’ presence or absence was coded: name, physical mailing address, age, and gender. |
| RQ4 | How much information is gathered through registration processes? | For companies with at least one registration opportunity, three categories revealed:
1) Companies that do not capture contact information (name, mailing address) or demographic information (age, gender). 2) Companies that capture either contact or demographic information but not both. 3) Companies that capture contact and demographic information |
| RQ5 | How are registration processes structured? | For companies with at least one registration opportunity, the number of steps required to complete the registration process. A step was an action in which information was submitted, such as clicking a button labeled “submit” or “continue” that then completed the registration process or continued it by linking to a subsequent step. |
| RQ6 | How frequently do registration processes include statements about why the information being asked for is needed and how it will be used? | For companies with at least one registration opportunity, the presence of one or more statement regarding what the sponsoring company would do with the data it gathers during the registration process. |
| RQ7 | How frequently do companies respond to site visitors who register? | For companies with at least one registration opportunity, the total number of e-mails sent in the 15-17-week period subsequent to the registration. |
Coder Reliability
All variables were recorded by two trained coders, namely, one of the coauthors and a second coder recruited and extensively trained by this coauthor. Intercoder reliability was calculated using Scott’s pi (Potter and Levine-Donnerstein 1999), and the reliability coefficients for all variables fell within the expected range of .80 to 1.00, above the minimal agreement level (Riffe, Lacy, and Fico 1998). Specifically, they reached 100% agreement regarding whether an incentive was offered to register, whether a visitor’s name was captured during registration, the number of steps required to complete registration, and the absence or presence of a statement of how the company would use the gathered data. Scott’s pi was equal to .80 for decisions about whether a visitor’s street address was captured during registration, .80 for decisions concerning whether gender was captured, and .85 for whether a visitor’s birth date was captured. Differences were reviewed and discussed by both coders to resolve disagreements. Coding took place over a three-week period between May 15 and June 6, 2009.
Results
Regarding the prevalence of registration processes in company-sponsored websites, we find that of the 200 companies in the sample, 131 (65.5%) have registration processes. The chi-square analysis indicates that prevalence and company size are not independent; significant differences exist across company size (c2 = 39.76, df = 1, p < .001). The chi-square analysis also indicates that prevalence and business model are not independent, such that significant differences in prevalence exist across business models (c2 = 7.9, df = 1, p < .01). We provide the frequencies, percentages, and chi-square analyses related to RQ1 in Table 2.
Table 2. Registration Prevalence Overall, by Company Size, and by Business Model
| Company Descriptors | Including Registration Process | No Registration Process | ||
| Percentage | Frequency | Percentage | Frequency | |
| Overall | 65.5 | 131 | 34.5 | 69 |
| Large | 86.3 | 88 | 13.7 | 14 |
| Small | 43.9 | 43 | 56.1 | 55 |
| Pure play | 59.9 | 82 | 40.1 | 55 |
| Click-and-mortar | 80.3 | 49 | 19.7 | 12 |
Notes: For the size-based measures, c2 = 39.76, df = 1, p < .001. For the firm type measures, c2 = 7.9, df = 1, p < .01.
For RQ2, regarding how frequently incentives are offered to site visitors, we find that of the 131 companies that included registration processes on their sites, 104 (79.4%) offer some type of incentive to persuade site visitors to register. The chi-square analysis indicates that prevalence of incentives and company size are not independent, with significant differences in their prevalence appearing across company sizes (c2 = 5.0, df = 1, p < .05). However, we find no differences in the prevalence of incentives across business models (c2 = .88, df = 1, p = .35), as we detail in Table 3.
Table 3. Incentive Prevalence Overall, by Company Size, and by Business Model
| Company Descriptors | Including Incentive | No Incentive | ||
| Percentage | Frequency | Percentage | Frequency | |
| Overall | 79.4 | 104 | 20.6 | 27 |
| Large | 73.9 | 65 | 26.1 | 23 |
| Small | 90.7 | 39 | 9.3 | 4 |
| Pure play | 76.8 | 63 | 23.2 | 19 |
| Click-and-mortar | 83.7 | 41 | 16.3 | 8 |
Notes: For the size-based measures, c2 = 5.0, df = 1, p < .05. For the firm type measures, c2 = .88, df = 1, p = .35.
In RQ3, we asked what kinds of information are gathered during registration processes. To answer this question, we organized the information into two buckets: the contact (name and/or mailing address) and the demographics (age and/or gender) buckets. We did not include e-mail addresses, because they were part of every registration process and thus do not offer a point of differentiation. Of the 131 companies that included registration processes, 66 (50.4%) captured registrants’ name and/or mailing address. The chi-square analysis indicates no differences in the frequency of capturing name and/or street mailing address information across company size (c2 = .25, df = 1, p = .62) or business models (c2 = .37, df = 1, p = .54). Furthermore, 17 (13%) of these sites captured gender and/or age data, but the chi-square analysis reveals no differences in the frequency of this information capture across company sizes (c2 = .2.04, df = 1, p = .15). We find that the frequency of capturing age and/or gender information and the business model are not independent, and significant differences in this data capture exist across business models (c2 = 6.22, df = 1, p < .05). We summarize these frequencies, percentages, and chi-square analyses relevant to RQ3 in Table 4.
Table 4. Type of Information Capture Overall, by Company Size, and by Business Model
| Company Descriptors | Capture Name and/or Street Address | Capture Neither | ||
| Percentage | Frequency | Percentage | Frequency | |
| Overall | 50.4 | 66 | 49.6 | 65 |
| Large | 48.9 | 43 | 51.1 | 45 |
| Small | 53.5 | 23 | 46.5 | 20 |
| Pure play | 52.4 | 43 | 47.6 | 39 |
| Click-and-mortar | 46.9 | 23 | 53.1 | 26 |
| Capture Age and/or Gender | Capture Neither | |||
| Overall | 13.0 | 17 | 87.0 | 114 |
| Large | 15.9 | 14 | 84.1 | 74 |
| Small | 7.0 | 3 | 3.0 | 40 |
| Pure play | 7.3 | 6 | 92.7 | 76 |
| Click-and-mortar | 22.4 | 11 | 77.6 | 38 |
Notes: For the name/address data, for the size-based measures, c2 = .25, df = 1, p = .62; and for the firm type measures, c2 = .37, df = 1, p = .33. For the age/gender data, for the size-based measures, c2 = 2.05, df = 1, p = .15; and for the firm type measures, c2 = 6.21, df = 1, p < .05.
To address RQ4, we delineated three levels of information captured during registration: no information from either the contact (name and/or mailing address) or demographics (gender and/or age) buckets; information in either the contact or demographics bucket but not both; and information in both the contact and the demographics buckets. Of the 131 companies that included registration processes on their sites, 59 (45%) capture information in neither bucket, 61 (46.6%) capture information in one or the other but not both, and 11 (8.4%) capture information in both buckets. Of the 61 companies in the second category, we determine that 55 captured just contact information, and 6 captured just demographic information. The chi-square analysis indicates no difference in the level of information quantity across company sizes (c2 = .19, df = 2, p = .91). The level of information quantity and business model are not independent though, such that significant differences exist across business models (c2 = 7.55, df = 2, p < .05), as we show in Table 5.
Table 5. Level of Information Quantity Overall, by Company Size, and by Business Model
| Company Descriptors | Captures No Information Process | Captures Either Contact or Demographic Information | Captures Both Contact and Demographic Information | |||
| Percentage | Frequency | Percentage | Frequency | Percentage | Frequency | |
| Overall | 45.0 | 59 | 46.6 | 61 | 8.4 | 11 |
| Large | 44.3 | 39 | 46.6 | 41 | 9.1 | 8 |
| Small | 46.5 | 20 | 46.5 | 20 | 7.0 | 3 |
| Pure play | 43.9 | 36 | 52.4 | 43 | 3.7 | 3 |
| Click-and-mortar | 46.9 | 23 | 36.7 | 18 | 16.3 | 8 |
Notes: For the size-based measures, c2 = .19, df = 2, p = .91. For the firm type measures, c2 = 20.35, df = 2, p < .001.
Regarding how the registration processes are structured, 39 (29.8%) companies offer one-step registration, 65 (49.6%) have a two-step registration process, and 27 (20.6%) use a registration process that requires more than two steps. The chi-square analysis indicates that registration process structure and company size are not independent and that significant differences in the registration process structure exist across company sizes (c2 = 20.35, df = 2, p < .001). Furthermore, the registration process structure and business model are not independent; significant differences in registration process structure exist across business models too (c2= 7.01, df = 2, p < .05). We provide the frequencies, percentages, and chi-square analyses related to RQ5 in Table 6.
Table 6. Registration Process Structure Overall, by Company Size, and by Business Model
| Company Descriptors | One-Step Registration Process | Two-Step Registration Process | More than Two Steps in Registration Process | |||
| Percentage | Frequency | Percentage | Frequency | Percentage | Frequency | |
| Overall | 29.8 | 39 | 49.6 | 65 | 20.6 | 27 |
| Large | 19.3 | 17 | 51.1 | 45 | 29.5 | 26 |
| Small | 51.2 | 22 | 46.5 | 20 | 2.3 | 1 |
| Pure play | 37.8 | 31 | 42.7 | 35 | 19.5 | 16 |
| Click-and-mortar | 16.3 | 8 | 61.2 | 30 | 22.4 | 11 |
Notes: For the size-based measures, c2 = .19, df = 2, p = .91. For the firm type measures, c2 = 7.01, df = 2, p < .05
In RQ6, we asked how frequently registration processes include statements about why the information being asked for was needed and how it would be used. Of the 131 companies with registration processes, 93 (71%) do not include such a statement. The frequency and company size are not independent; significant differences in frequency exist across company sizes (c2 = 5.04, df = 1, p < .05). However, we find no difference across business models (c2 = 2.27, df = 1, p = .13). These details for RQ6 appear in Table 7.
Table 7. Privacy Statement Inclusion Overall, by Company Size, and by Business Model
| Company Descriptors | Including Statement | No Statement | ||
| Percentage | Frequency | Percentage | Frequency | |
| Overall | 29 | 38 | 71.0 | 93 |
| Large | 35.2 | 31 | 64.8 | 57 |
| Small | 16.3 | 7 | 83.7 | 36 |
| Pure play | 24.4 | 20 | 75.6 | 62 |
| Click-and-mortar | 36.7 | 18 | 63.6 | 31 |
Notes: For the size-based measures, c2 = 5.04, df = 1, p < .05. For the firm type measures, c2 = 2.27, df = 1, p = .13.
Finally, we investigated how frequently companies respond to site visitors who register. The mean number of e-mails sent in the 15-17-week period after registration by the 131 companies that used registration processes was 1.00 per week, though 40 (30.5%) did not send any e-mail during that period. To determine whether the frequency of responses varied across company size, we conducted an independent samples t-test and found a significant difference in the average number of weekly e-mails sent by large companies (M = 1.34) versus small companies (M = .30) to registrants (t(129) = 17.12, p < .001). In addition, we find a significant difference in the average number of weekly e-mails sent by click-and-mortar companies (M = 1.53) compared with pure play companies (M = .68; t(129) = 6.41, p < .05). We provide the means, standard deviations, and t-test analyses related to RQ7 in Table 8.
Table 8. Weekly E-Mail Frequency (Overall Mean = 1.00)
| Company Descriptors | Mean | Standard Deviation | t-Value |
| Large | 1.34 | 1.57 | 4.19*** |
| Small | .43 | .30 | |
| Pure play | 1.53 | 1.84 | 6.41* |
| Click-and-mortar | .68 | .99 |
* p < .05.
*** p < .01.
Discussion
Too often, registration processes on websites have tended to be slow, difficult to use, and overly intrusive (Gaffney and Szuc 2006). The redesign of a poorly developed web form can increase registration completion rates by 10-40%, thus increasing the number of site members and sales (Wroblewski 2008). Despite their importance though, very little empirical research investigates registration process development and deployment. With this study, we provide a theoretical underpinning for best practices identified by designers and practitioners, as well as new best practices that are not yet being discussed. In doing so, we shed further light on why these best practices have emerged.
Perhaps surprisingly, only 65.5% of companies in the sample have registration processes in place. Clearly the industry as a whole must evolve further to implement the predictive targeting capabilities touted by optimization providers (Omniture 2008). Still, compared with 13 years ago, when the percentage of companies with a registration process stood at 14% (Ha and James 1998), it seems clear that companies are beginning to understand the benefits they can accrue by gathering information through registration. Three-quarters of the companies with registration processes provide an incentive to site visitors to register, and companies send an average of one e-mail per week to site registrants, which further confirms this claim. However progress on other fronts appears to be slower. Very few registration processes capture age or gender during the registration process, and more than 70% of the companies do not provide any statement regarding why they are gathering the information or what it will be used for.
We also considered how best practice deployment varied between large and small companies and between pure play companies and click-and-mortar companies. In the first comparison, we find that twice the percentage of large companies compared with small companies include registration processes. However, large companies are less likely to include an incentive for site visitors to register. We also note differences regarding the structure of the registration processes: Whereas more than 50% of the registration processes of small companies take place in one step, fewer than 20% of those of large companies are similarly brief. Instead, large companies are much more likely than small companies to stretch the registration process across multiple steps. Also, twice the percentage of large companies, relative to small companies, disclose why they need the gathered information and how it will be used. Finally, large companies tend to send, on average, five times as much follow-up e-mail to website visitors after their registration than small companies.
The differences across business models show that click-and-mortar companies are more likely than pure play companies to include registration processes on their sites and three times more likely to capture age and/or gender information during registration. Click-and-mortar companies are also more likely than pure play companies to capture both demographics and contact information during registration. Regarding the structure of registration processes, twice the percentage of pure play companies, relative to click-and-mortar companies, permit registration to occur in one step. Click-and-mortar companies tend to send, on average, twice as much follow-up e-mail to website visitors subsequent to their registration than do pure play companies.
Conclusions and Implications
The most effective online advertising serves the right message to the right customer at the right time in the right context (Kazienko and Adamski 2004). Profiling visitors with site registration enables advertisers to capture some information needed to deliver such advertising. Thus, the purpose of this study has been to fill a void in extant literature by (1) addressing how advertisers currently use registration processes to gather information that they can then use to customize advertising; (2) bringing together two streams of literature, exchange theory and interactivity theory, to provide support for existing and new best practices regarding registration processes; and (3) examining whether registration processes vary across two key company variables, company size and business model.
Offering incentives to gain prospects’ contact information is a proven offline method that has moved online successfully. Consider a department store such as Macy’s, which offers 15% off all purchases during a day if visitors sign up for a store credit card. This registration allows Macy’s to identify the person and convert him or her from a prospect (who spends cash or uses a third-party credit card in the store) to a customer. Macy’s then adds this customer’s name and mailing address to its database to receive regular mail and e-mail. Of the companies offering an incentive on our study, the most common practice online is to offer e-mail communications or free newsletters. Many prospects are likely aware that such follow-up information may include promotionally oriented and targeted discounts. Small companies outpace large companies in the decision to offer such an incentive; despite their limited resources, which often put them at a disadvantage compared with larger companies (Teo and Pian 2003), smaller companies appear to consider registration opportunities a cost-effective early step in integrating the Internet into their business strategy. They use registration incentives to facilitate the exchange of information.
Regarding the type of information captured during registration, we note several reasons the companies in our sample seem to prioritize the capture of contact information, such as name and mailing address, over demographic information such as gender and age. First, they can personalize communications, which may increase perceptions of familiarity and lead to stronger relationships with customers (Kasanoff, Peppers, and Rogers 2001). Second, by capturing the mailing address, they can communicate through multiple channels. According to a V12Group/Winterberry Group (2006) report, multichannel customers spend approximately 30% more annually than single-channel customers. Third, companies may conclude that requests for demographic information require additional time from consumers, which may make the difference between a completed and an abandoned registration form. The greater likelihood that click-and-mortar companies, compared with pure play companies, gather both contact and demographic information may reflect their greater reliance on multichannel marketing.
How much information should a company try to capture? Ideally companies want to know a tremendous amount about each site visitor: name, e-mail, address, telephone number, Facebook page, gender, age, income, marital status, home ownership, presence of children, shopping behavior, and more. Realistically though, asking for too much information will likely increase registration abandonment rates. We have observed both simple one-step registration processes on a home page (e.g., Apple Store, Ralph Lauren) and complex registrations requiring as many as four steps (e.g., ShopNBC, eBags). The former maximize their registrations and reduce the risk of registration abandonment, but they gather little more than e-mail addresses. The latter companies risk the ire of site visitors with their multistep registrations, but they gather a great deal of information that they can use to enhance their relationship marketing and personalize their advertising.
In this era of consumer concern about privacy and security online, we were surprised to observe that 71% of companies do not offer any statement about how they or third parties would use the contact information. The absence of such information may make wary site visitors less likely to complete the registration. A related observation reveals that larger companies tended to provide privacy information more frequently than smaller companies, consistent with the finding that firm size relates positively to disclosure of information related to social responsibility (Patten 2002). Smaller companies may receive less public pressure to make such disclosures, whereas larger companies likely have shareholders concerned about the types of programs developed by the company (Cowen, Ferreri, and Parker 1987). For similar reasons, larger companies may be more inclined than smaller companies to state explicitly how they will use registration process information.
We had anticipated that companies would contact registrants by e-mail to thank them for registering and begin communicating with them through customized product and service advertising. Yet only 69.5% of all companies did so. To provide the registration opportunity and then do nothing with the registrant’s e-mail address is puzzling for two reasons. First, the registrant expects that the company will send e-mail, and if that does not happen, the potential customer may not return to the website or lose trust in the company. Second, the company has missed the opportunity to begin a relationship with an interested party who has opted-in to its communications. Less surprising perhaps is that click-and-mortar companies are more likely to reply to new registrants than are pure play companies. With their experience in coordinating information gathered at their website and physical stores, they likely better understand the importance of responding immediately to new site members.
Limitations and Areas for Further Research
In this study, we have examined how companies’ registration processes vary across firm sizes and business models. Further content analyses could compare other important company variables. For example, Fader and Hardie (2009) distinguish between contractual and noncontractual companies. Contractual companies typically rely on a subscription-based revenue model and may value their membership information differently than noncontractual companies, as well as make different use of their registration processes in terms of the type and quantity of information gathered, the salience of the registration opportunity, and so on. In addition, companies in different industries may adopt different registration process practices, considering the various retail formats that exist across industries (Palmer 1997).
Although our literature review offers some best practices suggested by industry professionals, and we propose some new ones, additional survey research should focus on whether the registration process expectations of consumers match the registration process experiences defined by these best practices. Experiments could then determine whether best practices are predictors of intentions to register, time spent at site, intentions to return a site, and so forth.
Finally, eye-tracking methods may shed light on how users visually process information on Web forms (Wroblewski 2008). For example, the path of eye fixations could help determine the importance of including registration process elements such as graphics and privacy statements. These data also could allow researchers to explore the tolerance consumers have for how much information is gathered during the registration process.
References
Adamczyk, Piotr D. and Brian P. Bailey (2004), “If Not Now, When? The Effects of Interruption at Different Moments Within Task Execution,” in Proceedings of the SIGHI Conference on Human Factors in Computing Systems, Vienna, Austria, April, 271-78.
Bailey, Brian P., Joseph A. Konstan, and John V. Carlis (2001), “The Effects of Interruptions on Task Performance, Annoyance, and Anxiety in the User Interface,” INTERACT ’01, Tokyo, Japan, 593-601.
Bellman, Steven and John R. Rossiter (2004), “The Website Schema,” Journal of Interactive Advertising, 4 (Spring), http://www.jiad.org/article48 (accessed September 21, 2010).
Blalock, Hubert M., Jr. and Paul H. Wilken (1979), Intergroup Processes. New York: The Free Press.
Bozios, Theodoros, Georgios Lekakos, Victoria Skoularidou, and Kostas Chorianopoulos (2000), “Advanced Techniques for Personalized Advertising in a Digital TV Environment: the iMEDIA System,” inProceedings of the eBusiness and eWork Conference, Venice, Italy, October, 107-13.
Cho, Jinsook (2006), “The Mechanism of Trust and Distrust Formation and Their Relational Outcomes,” Journal of Retailing, 82 (1), 25-35.
Cowen, Scott S., Linda B. Ferreri, and Lee D. Parker (1987), “The Impact of Corporate Characteristics On Social Responsibility Disclosure: A Typology and Frequency-Based Analysis,” Accounting, Organizations and Society, 12 (2), 111-22.
Coyle, James R. and Stephen J. Gould (2002), “How Consumers Generate Clickstreams through Web Sites: An Empirical Investigation of Hypertext, Schema, and Mapping Theoretical Explanations,” Journal of Interactive Advertising, 2 (Spring), http://www.jiad.org/article23 (accessed September 21, 2010).
— and Esther Thorson (2001), “The Effects of Progressive Levels of Interactivity and Vividness in Web Marketing Sites,” Journal of Advertising, 30 (3), 65-77.
Dellaert, Benedict G.C. and Barbara E. Kahn (1999), “How Tolerable Is Delay? Consumers’ Evaluations of Internet Web Sites After Waiting,” Journal of Interactive Marketing, 13 (Winter), 41-54.
Ellinger, Alexander E., Daniel F. Lynch, and John D. Hansen (2003), “Firm Size, Web Site Content, and Financial Performance in the Transportation Industry,” Industrial Marketing Management, 32 (3), 177-85.
Evans, Philip and Thomas S. Wurster (1999), “Getting Real about Virtual Commerce,” Harvard Business Review, 77 (6), 84-94.
Fader, Peter S. and Bruce G.S. Hardie (2009), “Probability Models for Customer-Base Analysis,” Journal of Interactive Marketing, 23 (1), 61-69.
Gaffney, Gerry and Daniel Szuc (2006), The Usability Kit. VIC, Australia: SitePoint.
Gamberini, Luciano, Giovanni Petrucci, Andrea Spoto, and Anna Spagnolli (2007), “Embedded Persuasive Strategies to Obtain Visitors, Data: Comparing Reward and Reciprocity in an Amateur, Knowledge-Based Website,” in Persuasive Technology, Second International Conference on Persuasive Technology. New York: Springer, 187-98.
Garbarino, Ellen and Sarah Maxwell (2009), “Consumer Response to Norm-Breaking Pricing Events in E-Commerce,” Journal of Business Research, 63 (9/10), 1066-72.
Georgia Tech Graphics, Visualization & Usability Center (1998), “GVU’s 10th WWW User Survey,” http://www.cc.gatech.edu/gvu/user_surveys/survey-1998-10/ (accessed September 21, 2010).
Ha, Louisa and James E. Lincoln (1998), “Interactivity Reexamined: A Baseline Analysis of Early Business Web Sites,” Journal of Broadcasting and Electronic Media, 42 (4), 456-73.
Herman, J. (2004), “A Process for Creating the Business Case for User Experience Projects,” in Proceedings of the SIGHI Conference on Human Factors in Computing Systems, Vienna, Austria, April, 1413-16.
Houston, Franklin S. and Jule B Gassenheimer (1987), “Marketing and Exchange,” Journal of Marketing, 51 (October), 3-18.
Johnson, Grace J., Gordon C. Bruner II, and Anand Kumar (2006), “Interactivity and its Facets Revisited: Theory and Empirical Test,” Journal of Advertising, 35 (4), 35-52.
Kasanoff, Bruce (2001), Making It Personal: How to Profit From Personalization Without Invading Privacy. New York: Perseus.
Kazienko, Przemyslaw and Michal Adamski (2004), “Personalized Web Advertising Method,” Lecture Notes in Computer Science, 3137, 146-55.
Kim, Heeman, James R. Coyle, and Stephen J. Gould (2009), “Collectivist and Individualist Influences on Website Design in South Korea and the U.S.: A Cross-Cultural Content Analysis,” Journal of Computer-Mediated Communication, 14 (3), 581-601.
Krishnamurthy, Sandeep (2001), “A Comprehensive Analysis of Permission Marketing,” Journal of Computer-Mediated Communication, 6 (2), http://jcmc.indiana.edu/vol6/issue2/krishnamurthy.html (accessed September 21, 2010).
Lombard, Matthew and Jennifer Snyder-Duch (2001). “Interactive Advertising and Presence: A Framework,” Journal of Interactive Advertising, 1 (2), http://jiad.org/article13 (accessed September 21, 2010).
Malhotra, Naresh K. (2006), “Questionnaire Design and Scale Development,” in The Handbook of Marketing Research: Uses, Misuses, and Future Advances, R. Grover and M. Vriens, eds. Thousand Oaks, CA: Sage Publications, 83-94.
Marinova, Ana, Jamie Murphy, and Brian L. Massey (2002), “Permission E-Mail Marketing as a Means of Targeted Promotion,” Cornell Hotel and Restaurant Administration Quarterly, 43 (1), 61-69.
Okazaki, Shintaro and Javier A. Rivas (2002), “A Content Analysis of Multinationals’ Web Communication Strategies: Cross-Cultural Research Framework and Pre-Testing,” Internet Research: Electronic Networking Applications and Policy, 12 (5), 380-90.
Omniture (2008), “Overcoming Random Relevance: Using 1:1 Behavioral Targeting to Serve the Right Content,” http://www.omniture.com/offer/365 (accessed September 21, 2010).
Palmer, Jonathan W. (1997), “Electronic Commerce in Retailing: Differences across Retail Formats,” Information Society, 13 (1), 75-91.
Parasuraman, A. and George M. Zinkhan (2002), “Marketing To and Serving Customers Through the Internet: An Overview and Research Agenda,” Journal of the Academy of Marketing Science, 30 (4), 286-95.
Patten, Dennis M. (2002), “Give or Take on the Internet: An Examination of the Disclosure Practices of Insurance Firm Web Innovators,” Journal of Business Ethics, 36 (3), 247-59.
Pavlou, Paul A. and David W. Stewart (2000), “Measuring the Effects and Effectiveness of Interactive Advertising: A Research Agenda,” Journal of Interactive Advertising, 1 (1), http://jiad.org/article6, (accessed September 21, 2010).
Perry, Monica and Charles Bodkin (2000), “Content Analysis of Fortune 100 Company Websites,” Corporate Communications: An International Journal, 5 (2), 87-96.
Potter, W. J. and Deborah Levine-Donnerstein (1999), “Rethinking Validity and Reliability in Content Analysis,” Journal of Applied Communication Research, 27 (3), 258-84.
Riffe, Daniel, Stephen Lacy, and Frederick G. Fico (1998), Analyzing Media Messages. Mahwah, NJ: Lawrence Erlbaum Associates.
Saeed, Khawaja A., Varun Grover, and Yujong Hwang (2003), “Creating Synergy with a Clicks and Mortar Approach,” Communications of the ACM, 46 (12), 206-12.
Steinfield, Charles (2002), “Understanding Click and Mortar E-Commerce Approaches: A Conceptual Framework and Research Agenda,” Journal of Interactive Advertising, 2 (2), http://jiad.org/article19, (accessed September 21, 2010).
—, Alwin Mahler, and Johannes Bauer (1999), “Electronic Commerce and the Local Merchant,” Electronic Markets, 9 (1 & 2), 51-57.
Teo, Thompson S. H. and Yujun Pian (2004), “A Model for Web Adoption,” Information & Management, 41 (4), 457-68.
V12 Group/Winterberry Group (2006), “Tracking the Trends: A Comparison of Above-the-Line & Below-the-Line Expenditure Trends,” http://myadpsyche.files.wordpress.com/2010/04/report-on-btl-marketing1.pdf (accessed February 21, 2011).
Wroblewski, Luke (2008), Web Form Design: Filling in the Blanks. Brooklyn, NY: Rosenfeld.
Wurtz, E. (2005), “A Cross-Cultural Analysis of Websites from High-Context Cultures and Low-Context Cultures,” Journal of Computer-Mediated Communication, 11 (1), 274-99.
About the Authors
Jan Ahrens (DBA, Golden Gate University) is an Adjunct Professor of Marketing at both San Francisco State University and Golden Gate University. Her research interests include Internet marketing, electronic referrals, and online consumer behavior. E-mail: [email protected].
James R. Coyle (Ph.D., University of Missouri-Columbia) holds a joint appointment in the Marketing Department of the Farmer School of Business and Armstrong Interactive Media Studies at Miami University. His research interests include how consumers process and share interactive and rich media content in commercial websites. E-mail: [email protected].
Both authors contributed equally to this research.
