Gathering customer feedback via the Internet:
`instruments and prospects
`
`For many years companies
`have collected feedback from
`customers through means
`such as comment cards and
`toll-free telephone numbers.
`The feedback data can be
`used by companies to track
`quality, locate quality prob-
`lems, and identify sugges-
`tions for improvement. Gath-
`ering feedback from cus-
`tomers has become a recent
`but prevalent phenomenon on
`the Internet. Many companies
`designate an e-mail address
`for submitting comments and
`questions. Companies with
`information on the World
`Wide Web frequently include
`a feedback form that cus-
`tomers can complete on
`screen and send at the click
`of a mouse. This article con-
`siders current practice and
`the potential for customer
`feedback collection over the
`Internet. The nature of Web-
`based feedback forms is
`compared to corresponding
`features of conventional
`(paper) comment cards.
`Explanations for differences
`are supposed, and future
`prospects for Web-based
`feedback are discussed.
`
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`© MCB University Press
`[ISSN 0263-5577]
`
`Scott E. Sampson
`Department of Business Management, Marriott School of Management,
`Brigham Young University, Provo, Utah, USA
`
`Introduction
`A widely accepted objective in quality man-
`agement is being customer driven. A common
`way companies work towards this objective is
`to provide systems to gather feedback from
`customers. Along comes the Inter net, with its
`great potential to transfor m various business
`processes. One might wonder how Inter net
`technologies might transfor m the means of
`gathering (and using) customer feedback.
`This article presents an exploratory study
`of cur rent practice of soliciting customer
`feedback over the Inter net and discusses
`prospects for the future. The exploratory
`nature of the study is imposed because of the
`time frames of Inter net technologies.
`Although this and other articles pertaining to
`the Inter net are written as forward-looking
`treatises, within a few years of publication
`they will probably be regarded as historical
`documents. Not only is the technology chang-
`ing rapidly, but the rate of change is increas-
`ing dramatically. As a civilization we are on
`the verge of a revolution in the way customers
`and companies communicate with one
`another. This revolution will influence how
`companies gather and use feedback from
`customers – the topic of this article.
`The remainder of this section introduces
`ideas behind customer feedback, the Inter net,
`and implementations of customer feedback on
`the Inter net. The next major section looks at
`prior articles pertaining to Inter net-based
`feedback. An exploratory study which com-
`pares Inter net and conventional customer
`feedback instruments is then described.
`Results of the comparison are outlined and
`observations are made. The penultimate sec-
`tion considers prospects for customer feed-
`back systems in light of Inter net technologies,
`including potential problems and limitations.
`The final section offers a summary and con-
`clusions.
`Customer feedback
`It is common for companies to gather feed-
`back from customers. The feedback can take
`many for ms, including on-site customer com-
`plaints, calls to toll-free customer-response
`phone numbers, and customer comment
`cards. In each of these for ms, the feedback
`
`infor mation is either unsolicited, or passively
`solicited (Sampson, 1996). Passive solicitation
`represents an appeal to customers in general
`without focusing on any specific customer.
`In comparison, active solicitation is an
`appeal to specific customers, as with market
`research. The sample frame is usually
`selected with care to avoid sample frame
`bias. Further, active effort is taken to encour-
`age response so as to avoid non-response
`bias.
`With passive solicitation, the company has
`little or no control over sample frame and
`non-response bias, since the respondents are
`completely self-selected. Never theless, certain
`advantages exist with passive solicitation of
`feedback. The cost of gathering feedback is
`low. A passive appeal to each and every cus-
`tomer might represent no more cost than the
`staffing and maintenance of a toll-free tele-
`phone line and a sign at the service location
`or a notice on the product. Active solicitation
`is accomplished at moderate cost (e.g. mail
`surveys) to high cost (e.g. personal
`interviews) (Churchill, 1995, p. 377). It would
`probably be prohibitive to actively survey
`every single customer of a company.
`Another advantage of passive solicitation of
`feedback is in the use of the data. Since the
`data is inherently biased, it is not as useful as
`market research is in estimating general
`consensus of a target market. However, the
`nature of the bias can be exploited. One might
`assume that customers with exceptionally
`positive or negative views of the company are
`more likely to respond than the customer
`population in general. This would result in an
`extreme-response bias that would be more
`likely to identify cur rent quality problems
`than a controlled survey of equal sample size
`(Sampson, 1996). Therefore, passive data col-
`lection is particularly useful in monitoring
`and controlling quality in the day-to-day oper-
`ations of the business, and in identifying ideas
`for quality improvement.
`This article focuses on passive solicitations
`for feedback, and explores the implications of
`such data collection over the Inter net. The
`next sub-section provides a brief description
`of the Inter net and its provisions for customer
`feedback.
`
`[ 71 ]
`
`Qualtrics, LLC
`Exhibit 1007
`CBM of U.S. Patent No. 8,041,805
`Page 001
`
`

`
`Scott E. Sampson
`Gathering customer feedback
`via the Internet: instruments
`and prospects
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`
`[ 72 ]
`
`The Internet
`It is assumed that many readers are both
`familiar and experienced with using the Inter-
`net, which has been called a “network of [com-
`puter] networks” (Kehoe, 1994). Yet, some
`readers (even academics) are unversed in the
`technology, and many others are unaware of
`its origins and history. Thus, this section is
`presented. The following is a brief summary,
`followed by a description of Inter net faculties
`which are relevant to customer feedback.
`The Inter net began in 1969 as a US Depart-
`ment of Defense project called ARPANET
`(Stuart, 1994). By the end of that year, the first
`four “host” computers were attached to the
`network. (Other computers can access the
`Inter net by connecting, or dialing-in, to a host
`which is already connected to the Inter net.)
`Since its inception, the Inter net has included
`provisions for person-to-person electronic
`communication, i.e. e-mail (Sproull and
`Kiesler, 1986). This opens the way for
`customer feedback. A market researcher
`might use e-mail as a means of contacting
`individuals to actively solicit their feedback
`about a company or a product. Prior to 1993,
`such surveys would be largely limited to sci-
`entists and military personnel, who repre-
`sented most of the people with e-mail
`addresses. The mid-1990s saw explosive
`growth in Inter net access among the general
`population of developed nations. For example,
`from January 1995 to January 1996 the num-
`ber of hosts attached to the Inter net increased
`from 4.9 million to 9.5 million (Infoworld,
`1996). There is a large amount of speculation
`as to the scope of the Inter net at present – as
`the Inter net becomes larger and larger, it is
`increasingly difficult to deter mine what hosts
`are attached to other hosts. As the Inter net’s
`reach continues to expand, market
`researchers will have more opportunity to
`actively solicit feedback from various cus-
`tomer groups at lower cost than traditional
`mail and phone surveys.
`However, e-mail in and of itself does not lend
`itself to passive solicitation of feedback. A
`customer cannot go out and grab an e-mail
`message that has not been specifically sent to
`him or her (Sproull and Kiesler, 1986). Compa-
`nies may put customer-response e-mail
`addresses on product labelling, however, such
`passive solicitation are through the product,
`not through electronic means.
`The breakthrough for electronic passive
`solicitation of feedback came with the intro-
`duction of the World Wide Web (or “Web”) in
`1993-94. (The Web was prototyped at the Euro-
`pean Laboratory for Particle Physics – CERN
`– in 1990-91). (Ber ners-Lee, et al. 1994). The
`Web represents an electronic communication
`medium which is user (i.e. customer) initiated
`
`– a fundamental requirement of passive solici-
`tation of feedback. Documents (Web pages)
`are “posted” on the Web (i.e. made available
`for downloading). An uncountable number of
`organizations and individuals have Web sites
`which contain their Web pages. Among other
`things, companies can post general solicita-
`tions for feedback. Customers and others who
`access a company’s Web pages may elect to
`respond to these passive solicitations, as they
`might if they were presented with comment
`cards or toll-free telephone numbers.
`Feedback in HTML
`In addition to requests for response, the Web
`contains provisions which facilitate the feed-
`back process. These provisions are compo-
`nents of the language of Web: HTML, or
`HyperText Markup Language. Typically,
`HTML documents contain for matted text and
`hyperlinks (or “links”) which guide the user
`to other Web pages or other resources. HTML
`also includes provisions for for ms and
`“m ailto” links which can be used to gather
`feedback (Hoffman et al., 1995)
`A m ailto (as in “mail a message to some-
`one”) link is a code in an HTML document
`which, when selected by the user, opens an e-
`mail window on the user’s screen so that a
`message may be sent to a pre-specified e-mail
`address at the company. The e-mail window is
`simply a box for composing any textual mes-
`sage. The message is unstructured, thus is an
`Inter net version of a toll-free telephone num-
`ber to a customer service employee, but with-
`out personal interaction.
`Of even greater feedback potential are the
`for m provisions of HTML. An HTML docu-
`ment structured as a for m may contain text
`fields, check boxes, and/ or drop-down lists of
`selections. An example of code for an HTML
`for m is shown in Figure 1, and a depiction of
`the resulting for m is shown in Figure 2. After
`the user has entered infor mation to such an
`on-screen for m, a “submit” button can be
`selected to automatically send the infor ma-
`tion to the company’s computer. The company
`receives the infor mation in a structured for-
`mat which allows various options for han-
`dling the data – a topic which will be revisited
`later.
`
`Prior research
`Much research pertaining to active solicita-
`tion of customer feedback has been published.
`However, research regarding passive solicita-
`tion of customer feedback is scarce, to say the
`least. Since research pertaining to the Web is
`still in its genesis, it is particularly hard to
`find articles that address the feedback
`
`Qualtrics, LLC
`Exhibit 1007
`CBM of U.S. Patent No. 8,041,805
`Page 002
`
`

`
`Scott E. Sampson
`Gathering customer feedback
`via the Internet: instruments
`and prospects
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`
`potential of the Web. Most of the writing about
`such applications of the Web is in popular
`press and trade publications.
`Mur phy, For rest and Wotring (1996) con-
`sider customer feedback one of four Web-site
`communications functions for businesses.
`They do not go beyond just saying that feed-
`back will provide ideas and ways for improve-
`ment.
`Hoffman et al. (1995) identify the potential
`for engaging customers in communication
`with the fir m in order to receive infor mation
`from customers about their needs. They indi-
`cate that “e-mail buttons” (i.e. m ailto links)
`and for ms can be used for such feedback, but
`do not discuss the implementation of feedback
`systems.
`
`Figure 1
`HTML code for a sample form
`<TITLE>sample Feedback Page</TITLE>
`<FORM ACTION=”/cgi-bin/formdata” method=“POST”>
`<CENTER><H2>YOur feedback is important to us!</H2></CENTER>
`Which of our services have you used? (check all that apply)<BR>
`<INPUT NAME=“regular” TYPE=“CHECKBOX>regular
`<INPUT NAME=“custom” TYPE=“CHECKBOX>custom
`<INPUT NAME=“express” TYPE=“CHECKBOX>express<BR>
`Overall, how would you rate our company’s services?<SELECT NAME=“rating”>
`<OPTION>excellent<OPTION>adequate<OPTION>inadequate</select><BR>
`How might we improve our services?<BR>
`<TEXTAREA NAME=“howimprove” cols=35 rows=4></TEXTAREA><BR>
`What is your e-mail address? <INPUT TYPE=“text” NAME=“email” SIZE=“30”
`MAXLENGTH=“80”><BR><CENTER>
`<INPUT TYPE=“SUBMIT” VALUE=“Click here to submit feedback”></CENTER>
`</FORM>
`
`Figure 2
`How that form would typically display
`
`Emerick (1995) discusses Inter net feedback
`and gives some suggestions such as to keep
`for ms simple and ask open-ended questions
`when appropriate. Emerick also talks about
`using the feedback to capture e-mail
`addresses for future contact (which one may
`suppose is a less altruistic motive than quality
`improvement).
`The most descriptive report of electronic
`feedback data use is a case study by Marelli
`(1995). Marelli describes a beverage company
`which collects customer feedback via their
`Web site and uses the data a number of ways.
`They discuss automatic acknowledgment of
`feedback, personal response, monthly evalua-
`tion of feedback by a team, and communica-
`tion with repeat customers. Such ideas will be
`visited in a later section of this article.
`In an article about the role of the Web in
`marketing communication, Berthon
`et al.(1996) cite customer feedback as the sixth
`and final stage of the buying and selling
`process. However, they make no observations
`about implementation or effectiveness, stat-
`ing “… we re-emphasize the fact that the Web
`is still in its infancy, which means that no
`identifiable attempts have so far appeared in
`scholarly jour nals that methodically clarify
`its anticipated role and perfor mance”
`(Berthon, et al. 1996, p. 46). This being the case,
`it is not difficult to consider the present
`report to be foundational.
`
`Instrument comparison study
`Given the lack of prior research in this area, a
`descriptive exploratory study is war ranted.
`This will help us understand what potential
`research would be interesting and useful to
`pursue in the future. Even though research in
`Web-based feedback is scarce, passive solicita-
`tions of feedback are common on the Web.
`This study is an attempt to characterize cur-
`rent implementations of Web-based customer
`feedback mechanisms. The content and for-
`mat of customer feedback instruments are
`analysed with Web-based instruments being
`compared with conventional instruments, i.e.
`customer comment cards.
`Including an analysis of conventional
`instruments will serve two pur poses. First, it
`will help devise a taxonomy of instrument
`content components. The content of conven-
`tional instruments is expected to be more
`diverse than the content of Web-based instru-
`ments, since conventional instruments have
`been in existence much longer. Web-based
`instruments were not assumed a priori to be
`completely homogenous, but more homoge-
`neous than conventional instruments. The
`diversity of conventional instruments will
`[ 73 ]
`
`Qualtrics, LLC
`Exhibit 1007
`CBM of U.S. Patent No. 8,041,805
`Page 003
`
`

`
`Scott E. Sampson
`Gathering customer feedback
`via the Internet: instruments
`and prospects
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`
`lead to a taxonomy of instrument elements
`that is more extensive. Details of the taxon-
`omy will be published elsewhere, but key
`elements will be presented herein (such as the
`categorization of questions in Figure 3 and
`Table III, and the general types of textual
`feedback appeals of Table II).
`The second pur pose for analysing conven-
`tional instruments is to provide a basis for
`comparison. Such an exploratory comparison
`will give clues as to how Web-based and con-
`ventional gathering of customer feedback
`differs, and inferences will be made about
`how the use of feedback from the two methods
`may differ.
`Data collection
`Two features of HTML were described as
`mechanisms for feedback: m ailto links and
`for ms. A m ailto link is analogous to toll-free
`telephone numbers, being virtually free of
`any structure. As such, m ailto links (and toll-
`free numbers) will not be included in this
`analysis. Neither will mere descriptions of
`customer-response e-mail addresses. Instead,
`the analysis will focus exclusively on HTML
`for ms, and cor respondingly, customer com-
`ment cards.
`The author created a database of 71 HTML
`feedback for ms and 176 comment cards. The
`for ms and comment cards were collected in
`an unscientific manner: The for ms were dis-
`covered by “browsing” the Web or searching
`for postings with words like “feedback” and
`“comments” on them. The comments cards
`were likewise obtained by asking various
`businesses and organizations for them.
`Table I summarizes sources of the instru-
`ments (by US Standard Industrial Code cate-
`gorizations). Observe that most of the HTML
`for ms and all of the comment cards in the
`database are from the service sector.
`
`Table I
`Industries of feedback instruments
`Comment cards (# = 176)
`HTML forms (# = 71)
`81 Retail trade
`17 Data processing
`49 Restaurants and lodging
`15 Miscellaneous service
`27 Miscellaneous service
`9 Restaurants and lodging
`7 Air transportation
`7 Electrical and electronics
`3 Photography
`6 Printing and publishing
`2 Printing and publishing
`4 Computers and office equipment
`2 Banking
`4 Telecommunications services and equipment
`1 Broadcasting/cable TV
`2 Wholesale trade
`1 Recreation and entertainment
`2 Retail trade
`1 Data processing
`1 Recreation and entertainment
`1 Insurance
`1 Paper and wood products
`1 Health care
`1 Chemicals
`1 Automotive
`1 Miscellaneous manufacturing
`
`[ 74 ]
`
`Nevertheless, the present study is exploratory
`and will not control for differences in com-
`pany type. A separate study is underway
`involving a more systematic collection of
`instruments in a single industry.
`The initial database contained nine other
`HTML customer-feedback pages which were
`not included in the analysis: Seven had no
`fields but only m ailto links, and two were
`marketing infor mation request for ms with a
`comment field. Eliminating those nine entries
`assured focus on for ms which contain fields
`with the primary pur pose being customer
`feedback. Likewise, 28 out of 204 conventional
`instruments were omitted from analysis for
`similar reasons: The remaining 176 conven-
`tional instruments were comment cards with
`a distinct customer feedback pur pose.
`The primary reason so many fewer HTML
`for ms were entered into the database than
`comment cards is that after collecting about
`half of the for ms it became obvious that Web-
`based instruments were in fact quite homoge-
`nous. This will be seen in the next section.
`For each instrument, 170 different charac-
`teristics were recorded in the database. Exam-
`ples of characteristics include numbers and
`types of questions, response for mat (e.g. open-
`ended or rating scales), and types of textual
`appeals for feedback. The following highlights
`that data.
`
`Results and observations
`In this section, instrument characteristics are
`summarized in four general areas: feedback
`method and location, methods for encourag-
`ing feedback, types of questioning, and appar-
`ent uses of feedback data.
`Feedback method and location
`The means of submitting feedback for all
`HTML for ms was the same. The customer
`calls up the HTML for m and enters feedback
`infor mation on the computer screen. When
`the customer selects the for m’s “submit”
`button (which may be labelled something else
`such as “send comments”), the completed field
`infor mation is retur ned to the company’s host
`computer. The data are submitted as an ASCII
`string with field identifiers in what is known
`as Common Gateway Interface (CGI). A pro-
`gram (or script) residing on the company’s
`host computer can react to the submission.
`Comment cards, on the other hand, can be
`submitted in a number of ways. Of course, any
`comment card could be handed to an
`employee of the company. Of the cards, 21 per
`cent (37 of 176) actually included text stating
`that the customer can leave the completed
`card with a company employee if they care to.
`
`Qualtrics, LLC
`Exhibit 1007
`CBM of U.S. Patent No. 8,041,805
`Page 004
`
`

`
`Scott E. Sampson
`Gathering customer feedback
`via the Internet: instruments
`and prospects
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`
`The majority of the cards (77 per cent or 136
`cards) included an address for retur n mail,
`and 87 per cent of these (118 cards) included
`pre-paid postage (generally business reply
`mail).
`Thus we see a first major distinction
`between HTML for ms and comment cards:
`HTML for ms must be submitted from a com-
`puter which is connected to the Inter net;
`comment cards can be submitted at a com-
`pany location or, in most cases, at any mail-
`box. (An exception was that 8 per cent of the
`comment cards and 1 per cent of the HTML
`for ms listed a customer response phone num-
`ber, allowing submission of feedback from any
`phone.) Further, comment cards can be com-
`pleted at any time, whereas HTML for ms are
`typically completed only when they are dis-
`played on the customer’s screen. The method
`of submitting feedback may have significant
`ramifications for response rates: One might
`suppose that response rates would cor relate
`with the opportunities for completing and
`submitting the feedback instrument.
`Response rates for comment cards are gen-
`erally not high. In a couple of customer feed-
`back studies, response rates were 3.5 per cent
`(auto service industry) and 8.6 per cent (hotel
`industry) (Sampson, 1996; Sampson and
`Weiss, 1993). Such response rates may seem
`low, but since they are based on a passive
`sampling of entire customer populations, the
`absolute number of responses is quite high.
`Nevertheless, if response rates are too low the
`potential for accurately monitoring quality
`and identifying improvement opportunities is
`diminished.
`It would be difficult to tabulate response
`rates with passively solicited Web-based feed-
`back, since companies generally do not know
`what portion of the customer population sees
`the solicitation. (Companies can easily tabu-
`late the number of times the feedback for m is
`downloaded, but it would be difficult if not
`impossible to know what per cent of those
`downloads are by customers.) However, the
`results of a CommerceNet/ Nelson Inter net
`Demographics Survey (1995) suggest that
`response rates to passive solicitations over the
`Web might be quite high. A questionnaire that
`was placed on a Web site for four weeks
`resulted in more than 32,000 responses. (Dur-
`ing the same four weeks, telephone surveyors
`made 280,000 calls which yielded approxi-
`mately 4,200 completed interviews.) Again,
`the surveyors would not know how many
`unique individuals accessed the Web survey
`for m or saw the solicitation for response, but
`32,000 responses does suggest that it is not too
`difficult to get Web users to submit feedback.
`Perhaps an explanation for potentially high
`response rates with Web-based feedback
`
`solicitations is the ease of response (typing or
`clicking boxes with a mouse), and ease of
`submission (clicking the “submit” button).
`Comment cards involve a more complex
`process, including locating a pen or pencil,
`writing the feedback infor mation, and locat-
`ing an employee or mailbox. (We assume that
`there is little hope for cards which require
`finding a stamp.) If submitting feedback is
`perceived as requiring too much effor t, cus-
`tomers are likely to complain with their feet
`(Hirschman’s (1970) “exit” category) instead
`of with their comments (the “voice” category).
`This can be a serious problem – one study
`estimates the general ratio of exit to voiced
`complains (TARP, 1979, 1986) at 25 to 1.
`Anonymity may also positively effect
`response rates. Customers may have com-
`plaints or negative ratings and may not want
`to be identified. The 40 comment cards that
`must be presented at the company location
`(i.e. the 23 per cent with no addresses) limit
`the potential for anonymity. A total of 86 per
`cent of the comment cards ask for the cus-
`tomer to identify himself or herself (34 per
`cent of these indicate it is “optional”,
`although completing any field is always
`optional).
`Submitting HTML for ms can always be
`done anonymously. Clicking the “submit”
`button does identify the computer host to
`which the customer is attached, but does not
`identify the specific customer. Further, the
`company cannot respond to the customer at a
`future date unless the customer has submitted
`identification infor mation. Of the HTML
`for ms, 86 per cent asked for some type of iden-
`tification, the most common of which is for
`the customer’s e-mail address (asked for on 94
`per cent of the HTML for ms which asked
`customer identification questions). A total of
`13 per cent of the for ms asking for customer
`identification indicate that identification is
`optional.
`Another difference between electronic and
`conventional feedback is the temporal
`response frame. Comment cards with pre-
`printed addresses have the advantage of
`potentially being completed during, subse-
`quent to, or well after service is received.
`Again, comment cards without retur n
`addresses are at a disadvantage, since the
`feedback must be submitted when the cus-
`tomer is present at the company location
`(either at that visit or another visit to the
`service location). HTML for ms may also be at
`a disadvantage since the customer can only
`complete them when they are accessing the
`Web. (Of course, HTML for ms could be
`printed, completed, and mailed, but at a sig-
`nificant sacrifice of convenience. Besides,
`only a small per centage of the HTML for ms in
`[ 75 ]
`
`Qualtrics, LLC
`Exhibit 1007
`CBM of U.S. Patent No. 8,041,805
`Page 005
`
`

`
`Scott E. Sampson
`Gathering customer feedback
`via the Internet: instruments
`and prospects
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`
`the database listed a mailing address.) The
`on-line submission requirement is less of a
`disadvantage if the company’s line of business
`is providing Web infor mation, or if the com-
`pany desires feedback about their Web site
`(which a few for ms specifically said they
`desired). Otherwise, the response opportunity
`is limited to when the customer happens to be
`accessing the Web, if they ever do. Despite the
`exponential growth in the number of individ-
`uals with Web access (Inforworld, 1996),
`presently most people do not have access. The
`previously cited CommerceNet/ Nielsen (1995)
`study revealed that as of August 1995, only 17
`per cent of persons aged 16 and above in the
`USA and Canada have access to the Inter net.
`Encouraging feedback
`As was discussed in the prior sub-section,
`assuring an adequate response rate is an
`important issue. With the comment cards, this
`is accomplished in a number of ways. First,
`many of the cards included a textual appeal
`for feedback. With 32 per cent of the cards, the
`appeal is described as coming from a key
`person at the company (e.g. president or other
`executive). Also, more than half of the cards
`with a retur n address had the addressee listed
`as a top executive of the company. Of the
`cards, 37 per cent describe a major objective of
`the company, such as providing excellent
`service, and 31 per cent of these said they
`wanted to know if they were succeeding at
`that objective. A total of 21 per cent of the
`cards asked “How are we doing?” in various
`words, and 6 per cent said they wanted to
`know how the customer feels about the com-
`pany.
`On the other hand, none of the HTML for ms
`included notes from, or references to, top
`executives or key people at the companies,
`refer red to a company objective, nor said
`“How are we doing?”. However, the for ms
`used other types of appeals.
`Many of the HTML for ms emphasize the
`importance of feedback to the company.
`Table II describes types of other appeals listed
`on HTML for ms and comment cards. As can
`
`Table II
`Textual appeals for feedback
`
`Type of textual appeal
`We “value” or “appreciate” customer comments
`Your feedback is “important”
`We are “interested” in your feedback
`“We want to know” or “let us know” (your comments)
`We “welcome” your feedback or we “invite” your feedback
`“Please …” (give feedback)
`“Thank you” (for your feedback)
`
`Per cent of
`Web forms
`17
`1
`0
`18
`17
`48
`30
`
`Per cent of
`cards
`20
`7
`2
`3
`9
`56
`41
`
`[ 76 ]
`
`be seen, the occur rence of these types of
`appeals is not substantially different between
`for ms and cards.
`Another appeal for feedback which did not
`occur very often on either cards or HTML
`for ms are incentives such as entry into a con-
`test with a prize. Of the cards 3 per cent
`included such an incentive, and only one per
`cent of the for ms did. This is interesting, since
`a prior study suggests that such an incentive
`can have a dramatic effect on response rates
`(Sampson and Weiss, 1993).
`An additional type of appeal for feedback is
`a promise that someone from the company
`will respond to the feedback (assuming the
`customer identifies himself or herself). A total
`of 11 per cent of the HTML for ms include such
`a promise as do 13 per cent of the cards.
`Questioning
`A major difference between the HTML com-
`ment for ms and the conventional comment
`cards is the extent of questioning on each.
`Comment cards had an average of 18.46 ques-
`tions on each card, whereas HTML for ms
`were much simpler, with an average of only
`4.65 questions per card.
`Another major difference is the presence of
`the various types of questions. As mentioned
`previously, one type of questioning allows the
`customer to identify himself or herself. Six
`other general types of questions were found
`on comment cards:
`1 Dem ographics – questions about the cus-
`tomer (such as age and income).
`2 Buyer behavior – questions about how the
`customer came to select that company (such
`as “How did you hear about us?”).
`3 Incident reports – questions that ask the
`customer to report or describe incidents
`that occur red while the customer was
`receiving service from the company (such
`as: “When and where were you served?” or
`“Did the waiter offer you dessert?”).
`4 Evaluations – questions in which the cus-
`tomer is asked to evaluate the goodness
`(appropriateness, quality, excellence, etc.) of
`general or specific aspects of the service.
`5 Com m ents (or suggestions) – open-ended
`questions for general comments or other
`feedback.
`6 Mark eting – questions that try to sell the
`customer something more (such as “Which
`of our products could we send you more
`infor mation about?” or “Can we put you on
`our mailing list?”) recall that cards with
`marketing as the primary focus were omit-
`ted from this analysis.
`Comment cards contain a variety of each type
`of question. This is depicted in Figure 3.
`Observe that identification, incident reports,
`
`Qualtrics, LLC
`Exhibit 1007
`CBM of U.S. Patent No. 8,041,805
`Page 006
`
`

`
`Scott E. Sampson
`Gathering customer feedback
`via the Internet: instruments
`and prospects
`Industrial Management &
`Data Systems
`98/ 2 [1998] 71–82
`
`evaluation, and comment questions were
`present on most of the cards. HTML for ms, on
`the other hand, mostly contained two types of
`questions: identification and comment. The
`identification question on HTML for ms was
`generally asking for the customer’s e-mail
`address (on 94 per cent of the for ms with iden-
`tification questions) or his/ her name (on 88
`per cent of those for ms).
`
`Figure 3
`
`cust. identification
`
`demographics
`
`buyer behavior
`
`incident report
`
`evaluation (rating)
`
`general comments
`
`marketing
`
`0%
`20%
`40%
`60%
`portion with that question type
`Key
`comment
`
`HTML forms
`
`80%
`
`100%
`
`Interestingly, if we compare the for ms that
`contain a particular type of question to the
`cards that contained that same type of ques-
`tion, we observe that the number of questions
`of that type per card/ for m is somewhat more
`similar between cards and for ms. This is
`shown in the fourth and last columns of
`Table III.
`For example, we see that of cards or HTML
`for ms that ask for customer identification,
`about three such questions are presented.
`Also, cards or HTML for ms asking for general
`comments tend to ask about one and a half
`such questions. One thing that differentiates
`HTML for ms in regard to comment questions
`is the presence of a question labelled
`“subject”. One quarter of the HTML for ms
`asked for a one-line subject for the comments,
`which seems to be an idea coming from the
`“subject” line of standard e-mail messages.
`Such a subject line might be useful in direct-
`ing (e-mail forwarding) comments to specific
`individuals and in maintaining a database of
`feedback received.
`The most significant disparity in the com-
`parison of Table III is in evaluation questions.
`Evaluation questions are dominant in com-
`ment cards, with an average of more than 12
`questions per card. Few HTML for ms ask for
`evaluations, and those that do