WO2001024057A2

WO2001024057A2 - Automated internet research tool including proxy server

Info

Publication number: WO2001024057A2
Application number: PCT/US2000/026739
Authority: WO
Inventors: Douglas King Van Duyne; James Anthony Landay; Mark Newman
Original assignee: Netraker Corporation
Priority date: 1999-09-28
Filing date: 2000-09-28
Publication date: 2001-04-05
Also published as: WO2001024057A3; AU7731300A

Abstract

A method and apparatus for automatically gathering research data for a website. The method includes identifying question in a question set that are a measure of a research metric to be gathered for the website and embedding one or more instructions in a first question in the question set. The invention includes instructions to record participant response data using a proxy server disposed between the participant and the website. The method includes initiating a research test including providing a first question in the question set to the participant and gathering response data as each question in the question set is answered by the participant.

Description

AUTOMATED RESEARCH TOOL INCLUDING PROXY SERVER

This application is a continuation- in-part (and claims the benefit of priority under 35 USC

5 120) of U.S. application serial no. 09/407,802, filed September 28, 1999. The disclosure of the prior application is considered part of (and is incorporated by reference in) the disclosure of this application.

The present invention relates generally to computer systems and more particularly to methods and apparatus for gathering research data for a website. 0 The explosion of the Internet has provided new business opportunities and markets for a broad spectrum of industries. The Internet provides unprecedented opportunities for businesses to reach customers by providing a website that describes a business 's products or services. The appeal of the website often can influence purchase decisions by a consumer. The appeal is a measure of both the subject matter, the method of presentation and other intangible factors. If a 5 website includes content that is appealing to a consumer, the consumer is more likely to purchase the services or products advertised. Advertising has long been used in other media forms in order to attract purchasers. The web is no different. The success of a product or service on-line is generally measured in sales. However, reasons for closing a sale vary. Some sales will complete irrespective of the environment. Other customers may have to be convinced that the 0 product or service is both valuable and necessary. The website can be a powerful tool in the persuasion process.

In addition, a website may be used to convey information, separate and apart from a sales environment. Success for an informational website may be measured in terms of the number of users that come to the website and the amount of time they stay. How long a user stays at a

25 website can be influenced by the appeal of the website. Again, the appeal is a measure of both the content and the method of presentation.

In order to measure the appeal of a website, conventional web page builders have numerous options. In some cases, a web page may include a request for feedback and provide users with an ability to send compliments and, more often, complaints back to the website sponsor. While these means are helpful, the information received may be biased or not representative of the general appeal of the website and therefore of little help to the website sponsor. Alternatively, a user may be asked to participate in a survey. Often, the surveys are long and unfocused. Just as with the feedback information, the information received can be biased or unreliable.

SUMMARY In one aspect the invention provides a method for automatically gathering research data for a website and includes steps for identifying questions in a question set that are a measure of a research metric to be gathered for the website and embedding one or more instructions in a first question in the question set. The instruction includes instructions to record participant response data using a proxy server disposed between the participant and the website. The method includes initiating a research test including providing a first question in the question set to the participant and gathering response data as each question in the question set is answered by the participant. Aspects of the invention can include one or more of the following features. The research data can be selected from usability, product and organizational research. The gathering step can be performed without a permission of a website's sponsor.

The method can further include forming questions in a plurality of question sets that are a measure, respectively, of a plurality of research metrics to be gathered for the website and partitioning each question set into a plurality of subsets of questions. The method can further include only presenting a subset of questions for each question set to each participant including gathering responses from each participant to the questions presented for each research metric and combining the participant's responses to produce a plurality of indexes, one for each of the research metrics. Only a subset of the questions for one or more question sets can be presented to each participant including gathering responses from each participant to the questions presented for each research metric and combining the participant's responses to produce a plurality of indexes, one for each of the research metrics.

The method can further include, at the participant's web browser, extracting the embedded instruction from the first question received and generating a request to load a page from the website including the embedded instruction, transmitting the request from the participant's web browser to the proxy server and enabling a recorder at the proxy server to record participant response data. Included with the embedded instruction can be an IP address for the proxy server and a code designating the website. The method can further include determining the IP address for the website from the code and forwarding the request to the website.

The method can further include receiving back from the website a page responsive to the request and forwarding the page to the participant's web browser. Prior to forwarding the page to the participant's website, a code can be embedded in the page to measure a metric. The code can include an engine for recording participant responses generated at the participant's web browser. The participant responses can be selected from the group of audio, video, gestural, facial, mouse, pen, joystick and keystroke responses. The code can include an engine for capturing a screenshot of a user interface presented to the participant.

The method can further include embedding one or more instructions in a last question in the question set where the instruction causes a recording device to stop recording participant response data using the proxy server and to initiate the download of the gathered response data to a centralized research site for analysis.

The method can further include incentivizing a candidate to become a participant including identifying a reward to be given to a participant upon completion of participation in the research test. The method can further include characterizing research participants into one or more types and customizing the reward for each participant type.

The method can further include storing historical results for the research metric and a representation of the website or portion of the website to which the metric pertains, determining that the website has changed, gathering new responses from participants to the questions presented for the research metric, combining the participant's responses to produce an index of the research metric for the changed website and comparing the indexes for the website against the changed website to measure a change of the research metric between the two. The method can further include recording at the proxy server each page request received from a participant. The method can further include recording at the proxy server each page received from the website prior to forwarding the page to the participant's web browser.

The method can further include recording participant response data at the proxy server and, after the research test has been completed by the participant, downloading the participant response data to a centralized research site for analysis. The method can further include compiling answers to each question at the participant's computer and, after the research test has been completed by the participant, downloading the participant response data to a centralized research site for analysis. The method can further include compiling answers to each question at the participant's computer, forwarding the participant responses to the proxy server for recording and, after the research test has been completed by the participant, downloading the participant response data to a centralized research site for analysis. The participant response data can be analyzed including playing back the participant response data in real time. The response data and a visualization of the website can be played back concurrently at a centralized research site to analyze of the response data. The participant response data can be played back at a faster rate than a recording rate for the participant response data. The faster rate can be twice the recoding rate.

The proxy server can be located at a Internet service provider accessed by the participant for linking to the website. The proxy server can be located at a hosting server that couples the website to the Internet. The participant can be coupled to the web site using a PDA and the proxy server can be located at a PDA proxy that is accessed by the participant for linking to the website. The participant can be coupled to the web site using a PDA that includes a PDA proxy for linking the PDA to the Internet and the proxy server can be coupled between the PDA proxy and the Internet. The participant can be coupled to the web site using a wireless telephone and the proxy server can be located at a WAP proxy that is accessed by the participant for linking to the website. The participant can be coupled to the web site using a wireless telephone that includes a WAP proxy for linking the wireless telephone to the Internet and the proxy server can be coupled between the WAP proxy and the Internet. Aspects of the invention may include one or more of the following advantages. Research can be automatically gathered from a website. The research can include website marketing, usability, product and organizational research. The research can be performed with or without the permission of the website sponsor. User responses and interaction including audio and video can be recorded and analyzed to allow for easy resolution of usability problems.

Other advantages and features will be apparent from the following description and claims.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic block diagram of a web site research tool.

FIG. 2 is a flow diagram of a process for gathering research. FIG. 3 is a flow diagram of a process, from the perspective of the participant, for gathering research.

FIG. 4 is a schematic block diagram for an Internet service provider based web site research tool.

FIG. 5 is a schematic block diagram of a hosting server based website research tool.

DETAILED DESCRIPTION Referring now to FIG. 1, a system 100 to conduct research on a website is shown. The system includes a proxy server 102, a client web server 104, a web browser 106, and a test server 108. Web browser 106 is associated with a user's personal computer and allows access to a client web server site associated with the client web server 104. Associated with the client web server 104 is an Internet protocol (IP) address. The IP address for the client web server 104 can be provided to or is otherwise known by proxy server 102. An IP address for the proxy server 102 is provided or otherwise made known to the web browser 106 (user) and can be associated with the client web server IP address. In one implementation, a user generates requests that are sent to proxy server 102 (to be forwarded to client web server 104) that include the IP address of the proxy server and a code for designating a particular client web server 104. In this implementation, proxy server 102 can service plural client web servers 104, each designated by an associated code.

To access the client's web server site, web browser 106 requests access to a page associated with the client's web site, located at client web server 104. In one implementation, the 5 web browser generates a request/(activates a link) to the client web server 104 through proxy server 102. More particularly, the request includes a pointer to the client web server site through the proxy server IP address. Proxy server 102 forwards or communicates the page request to the client web server 104. Client web server 104 responds to the request providing, for example, a page of information, back to proxy server 102 for transmission back to the individual end user. 0 At the proxy server 102, code may be inserted into the material returned from the client web server producing a modified page. The modified page is then provided to the web browser 106.

In one implementation, the modified page includes code for conducting research. The code, when inserted into the returned page, allows for research on a client's site without requiring the client site to insert any special code in their own web page. Codes for conducting research 15 are described in greater detail in co-pending application entitled "Automated Research Tool", filed September 28, 1999 and assigned United States Patent Application serial number, 09/407,802, the contents of which are expressly incorporated herein by reference.

In one implementation, proxy server 102 includes a screening mechanism 110 for screening particular requests that are received from a web browser 106. Screening mechanism 20 110 can include a screening list defining a criteria to screen requests. If a particular request satisfies the screening criteria, then proxy server 102 can invoke an associated action (depending on the criteria). In one implementation, proxy server 102 inserts codes into requests to be forwarded to the client web server 104. In another implementation, proxy server 102 uses the criteria to select which returned pages are to be modified. 25 In one implementation, either a special code or no code can be inserted in pages returned from the client web server for ultimate transfer to the web browser. The code can be in the form of a recorder or other means for recording information associated with the actions of the user or web browser 106. In one implementation, the code gathers and provides data to test server 108. The data provided to test server 108 can be of the form of keystrokes, URLs associated with the pages that the web browser links to or other data.

In one implementation, proxy server 102 transforms each page request received prior to passing it on to an associated client web server 104. The transformation may include the insertion of other information or request data to the client web server 104. In one implementation, information can be appended to the request and returned to the web browser as part of the returned page. For example, a cookie can be appended to the request that is forwarded onto to the client web server. The cookie can in turn be attached to the returned page from the client web server 104 and forwarded in normal course through the proxy server 102 to the web browser 106.

Referring now to FIG.s land 2, a method for executing research on a web site is shown. The method begins with test initiation (202). In one implementation, test server 108 initiates a test by providing the IP address for a site to be reviewed to the web browser (user). Alternatively, the user may access the client server web site through a link on a web page or by other means. In each case, access to the client web server web site after this initiation is through proxy server 102. The research gathering process begins when a first test question is sent to the user (204). In one implementation (the stand alone mode), the test questions can be provided by or through proxy server 102. In this configuration, the proxy server can initiate a research project upon receipt of a page request for the client web server' web site. As described above, not all requests may prompt a research project. A screening mechanism can be employed to only select a portion of the requests that are forwarded through the proxy server. Test questions can be stored locally at proxy server 102 or alternatively can be stored remotely. Alternatively, test server 108 can provide test questions either directly or indirectly (through proxy sever 102) to the user of web browser 106. The test questions can be of the form of interrogatories concerning a particular web site or alternatively, can be an instruction for the user to perform a particular task at a particular web site (i.e., the client web server site). In one implementation, the question includes an embedded instruction to the proxy sever 102, the client web server 104 or both. In one implementation, the test question includes an embedded instruction directed to proxy server 102 to record information relating to the question. In one implementation, the information recorded is input received from the web browser 106. Client input including key strokes and URL's can be recorded as received by the proxy server 102. In another implementation, actual web pages returned from the client web server are recorded. In another implementation, other information including user parameters are recorded by proxy server 102. The user parameters can include screen resolution, screen size and browser window size.

In one implementation, the question includes an embedded instruction to insert code into web pages returned by client web server 104. As described above, the modified pages can be forwarded to the user. In one implementation, the returned pages from the client web server 104 are modified to include code to detect keystrokes and other actions of the user executed when answering the question or performing the designated task. The code can include instructions for recording the user input and forward the user input to the proxy server 102 or directly to the test server 108 as appropriate. In one implementation, the code can include instructions to record audio, video (including facial, gestural, mouse, pen or joystick responses) or other user input. In another implementation, the code can include instructions to measure the speed at which pages or other web based activities are performed at the user's computer. By measuring the performance directly at the user's computer, information can be gathered about particular performance problems that can be of a regional nature, time sensitive or otherwise only locally detectable. Test server 108 can be of the form of a centralized research site that includes tools for constructing, gathering and analyzing research. Alternatively, test server 108 may only be used to gather research, leaving the construction and analysis operations to a third party. The gathering of research includes the execution of a research project. In one implementation, each research project includes a questionnaire. The questionnaire defines a set of questions or tasks that are to be performed by the participants to the research project. In one implementation, the questions are of the form of index questions and benchmark questions. Index questions gather information specific to a particular site and produce a measurable result, typically of the form of a numeric rating. The term "index," as used herein, refers to the rating assigned by the participants to a particular index category. The rating is measured through the responses received from the participants of the research project. Examples of categories of index questions are usability, site satisfaction, retention, brand value, service, content and community indexes. Usability is a measure of a web site's ease of use as judged by the site's visitors. Site satisfaction refers to the overall satisfaction of the website as judged by the site's visitors.

Retention is a measure of the likelihood that a site visitor will perform a predefined action. Example of retention categories include likelihood the visitor will return to the website, use the product, tell friends or associates or purchase a product.

The brand value index category is a measure of the appeal of the website to a visitor. Distinguished from the satisfaction index, the brand value index is more of an assessment as to the promises made, or the marketing and images that are used in the website in marketing the product or service.

The service index is a measure of the satisfaction that the visitor finds as to whether or not the products or services that are being offered will accomplish the stated purpose and whether or not there is a belief that the offerings are in fact valuable.

The content index is a measure of the satisfaction of the visitor to the content presented on the site, that is, the information provided about the product or service.

The community index is a measure of the community impact for a website. Fostering a sense of information and idea sharing as well as the opportunity to develop mutual learning opportunities among a site's visitors and between visitors has been one of the hallmarks of currently successful sites. The community index attempts to measure the impact of the website from this unique perspective.

Benchmark questions gather information about the participant. Benchmarks can be used as screening factors to determine which candidates should be selected to be participants in a particular research project. For example, if the customer is seeking to get feedback for a particular site from people that live in the south, that are over the age of 60, then the responses to the benchmark questions can be used to screen in (or out) participants. In one implementation, the categories of benchmark questions include audience identifier, lifestyle, demographics and computing environment. Audience identifier questions are used to identify some detail of the audience, other than demographics, that is tied to the likes and dislikes of a particular participant.

Lifestyle questions can be used to determine the background or lifestyle from the perspective of what is important to the particular participant. Demographics questions can be used to screen for gender, age, country, language, education and other vital visitor statistics.

Computing environment questions can be used to screen for a participant's experience associated with a computer or with a particular computing environment as well as the participant's hardware, software or other computing experience. No matter the instructions received, the user is directed to go to a web site at the client web server 104 and the user generates, and the proxy server receives, a request for the site (206). The web page requested by the web browser is returned to the web browser through the proxy server 102 as described above (208). After a user completes a task, or other wise answers a question, responses are recorded (either locally at the web browser or at the proxy server 102) for forwarding to the test server 108 (210).

A check is made to determine if all questions/ tasks have been answered/performed (212). If all questions/tasks have been completed, the user can collect a reward or otherwise terminate the research project. If more questions remain, then the next question is retrieved and transmitted to the user (214). In one implementation, the last question received can include an embedded code to tum-off the recorder or otherwise end the monitoring process performed by proxy sever 102. After the test has terminated, the recording entity (either the web browser or the proxy sever 102) gathers any recorded information and forwards the same to the test server 108 for evaluation (216). Thereafter the process ends.

Test server 108 can playback the questions, responses and the information gathered by the various recorders to evaluate the client web server's site. The playback can include screen shots viewed by the user shown in real time. Alternatively, the playback can be accelerated (e.g., 2x or 4x normal rate) so that results can be gathered quickly. The results can be aggregated, visualized graphically and submitted to a third party for automatic analysis. The results can be analyzed graphically by comparing the retrieved data with an "expert" trace to determine where a user had difficulty in completing a task. Alternatively, results from users can be compared to determine deviation among the users.

Participant Process Referring now to Fig. 3, a process 300 from the perspective of the participant is shown for gathering research. The process begins with the receipt of an invitation (302). The invitation can be sent directly to the candidate, or can be displayed on a website (where candidates are selected from website visitors). The candidate links to the client web server via the proxy server (304). In one implementation, the link is provided as part of the invitation and can be of the form of a URL.

Thereafter, the participant completes the research questions and tasks (if any) (306). In one implementation, a participant is given a subset of a questions to answer in an index set. Alternatively, the participant can be given a subset of the questions in each index set to answer. After the participant has completed the question set(s), the participant completes a benchmark question set(s) (308). In one implementation, a participant is provided a subset of questions from each benchmark type. The test data and any recorded information is uploaded to the proxy server or directly to the test server (310). Thereafter, the participant can collect an identified reward (312) and the process ends.

Alternative Implementations

In one implementation, the results received at test server 108 or proxy sever 102 can be interpreted in real time to determine a next question or instruction to generate/send to the user. In one implementation, the proxy server can be located at an Internet service provider (ISP) for a user. This configuration is shown in Fig. 4. A user's computing device 400 is coupled to the Internet 404 through server 402. Server 402 can be of the form of a proxy server that includes all of the attributes of proxy server 102 described above. Server 402 can include a screening mechanism for screening particular requests from users directed to a client web server's 406 web site, or alternatively can be used to perform research for all requests to a given web site. The research gathering can be performed without the knowledge or permission of the client web sever 406. As described above, server 402 can provide research questions to the user or alternatively can be linked to a test server which in turn can provide the research materials. In another implementation, the test server can interact directly with the user. Code included in page requests from the user can be used to initiate recording operations at server 402. The user's computing device can be of the form of a personal digital assistant (PDA)

(e.g., Palm Computing Pilot or the like) and include a PDA browser. The PDA browser may be configured to link to a PDA proxy for access to the Internet. The PDA proxy may translate requests received from the PDA and forward them on to the Internet. Similarly, the PDA proxy may format responses received from the Internet for linking back to the PDA. The linking can be wireless or wired through a host computer. The PDA proxy may include functionality that is similar to proxy sever 102. That is, the PDA proxy may include screening mechanisms and devices for conducting research for PDA users that access a particular client web sever web site. Alternatively, the PDA proxy may in turn be coupled through a proxy server (e.g., an ISP as discussed above with regard to FIG. 4) to the Internet. The user ' s computing device 400 can be of the form of a wireless telephone and include a small screen version of a browser. The browser may be configured to link to a wireless access protocol (WAP) proxy for access to the Internet. The WAP proxy is a specialized proxy that supports HDML for small screen devices. The WAP proxy may translate requests received from the wireless telephone and forward them on to the Internet. Similarly, the WAP proxy may format responses received from the Internet for linking back to the wireless telephone. The WAP proxy may include functionality that is similar to proxy sever 102. That is, the WAP proxy may include screening mechanisms and devices for conducting research for wireless telephone user's that access a particular client web sever web site. Alternatively, the WAP proxy may in turn be coupled through a proxy server (e.g., an ISP as discussed above with regard to FIG. 4) to the Internet.

In one implementation, the proxy server can be located at a hosting service (host) for a client web server. This configuration is shown in Fig. 5. A user's computing device 500 is coupled to the Internet 502. Client web server 506 is coupled to the Internet 502 through a hosting server 504. Server 504 can be of the form of a proxy server that includes all of the attributes of proxy server 102 described above. Server 504 can include a screening mechanism for screening particular requests from users directed to a client web server's 506 web site, or alternatively can be used to perform research for all requests to a given web site. The research gathering can be performed without the knowledge or permission of the client web sever 506. As described above, server 504 can provide research questions to the user or alternatively can be linked to a test server which in turn can provide the research materials. In another implementation, the test server can interact directly with the user. Code included in page requests from the user can be used to initiate recording operations at server 504.

In one implementation, a user is selected to be a participant in a research project by random selection from among the user's that visit a particular web site. In this type of implementation, a selected participant is included in a single research project for the given site. Alternatively, a user can be a participant in a global research project that monitor's the users feedback for more than one web site. In this implementation, a user's web browser can be configured to always route requests to the Internet through proxy server 102. In one implementation, whenever the user accesses a web site from a preselected group of web sites, then research is gathered as described above. Alternatively, research can be gathered each time the user accesses the Internet. In one implementation, the research can be of the form of diary events. The diary research can span a long period of time and include plural diary entries, one for each diary event (access) of the Internet. Test server 108 can include an invitation engine for selecting participants to be included in a research project. The invitation engine can be used to locate participants. Research participants may be provided by a user (e.g., the website owner provides the participants), selected from a list of participants (an approved list) or candidates can be randomly invited to join in the research. A web site can include an invitation means. In one implementation, the invitation means generates invitations to be sent to participants using a pop-up palette. Various forms of invitations are stored on the pop-up palette. The palette can be automatically initiated at various times to invite visitors to join in a research project. This feature can provide a truly random sampling of the users for inclusion into a research project. Alternatively, research participants may be chosen from a list of research participant candidate. A pop-up palette controller can measure the success rate associated with responses that are received from a particular pop-up invitation. The success or failure of any individual pop-up invitation can be tracked and an algorithm can be used to self-tune the pop-up palette so that the default pop-up invitation selected for a particular site based on the highest response rate. Test server 108 can include various tools to construct research projects. Tools for constructing research projects are described in greater detail in copending application entitled "Automated Research Tool".

In one implementation, a recorder is downloaded to a participant prior to commencing research questioning. On the user side, the system provides for multilevel screening and recording of the interaction of the users. The recorder can be configured to record keystrokes, screen progression, audio and video at the participant's site. The recorder can automatically bookmarks recordings. Recording can be initiated upon receipt of a command from the run-time module 104. The recorder also can stop recording and upload, back to the run-time module 104, data retrieved on command (from the run-time module). In one implementation, test server 108 includes an intelligence center to share customer intelligence and team management information. An example of an intelligence center is discussed in greater detail in copending application entitled "Automatic Research Tool".

In one implementation, each balloter is provided a subset of one or more index question sets as well as one set or more of benchmark questions. Each balloter generates a number rating for a website page in one or more of the index sets. A single number can be used to summarize the usability, retention, etc., for the particular index. In one implementation, the benchmark is a measure of the deviation from an Internet standard and provides a percentage in terms of the deviation for the particular user group. Correlations can be drawn by a correlation engine in the test server 108 between the benchmark data and the index data in order to provide the customers with more specific information. In addition, the correlator engine can be used to provide a longitudinal measure of customer ratings. By a longitudinal customer ratings measurement, we refer to the system's ability to gather information associated with the website over a period of time. Customer ratings can then be measured in a particular group over a particular time. The results can be gathered and built into a database and the database can be updated to provide a strategically sound result.

Various research tools can be invoked by the team members using the intelligence center. Examples of tools include bug tracker, competition tester, site tester, performance measurer, tasker, action tracker, search services morale meter, organization tool, product tool and service tool.

Bug tracker can be used to ensure that a website is free from costly and embarrassing bugs. The tracker can be invoked for one time operation or invoked on a regular basis to screen a customer's website. Competition tester is a tool that can be used to invoke research on a competitors website.

A competitor's strengths and weaknesses can be measured and leveraged by the customer using the index data derived for a competitors site. Changes in the competitor's site can be monitored and correlation data derived reflecting changes in the usability of the competitor's site.

Site tester is a tool designed to automatically test a customer's site for common errors. In one implementation, tests are run for resolving domain names (DNS entries), missing pages, missing graphics, poor response and other common errors.

Performance measurer provides diagnostic information for problems associated with performance for a particular site. The tool provides performance measurement diagnostics as well as site screen capture and review features. Tasker is a tool for performing task based research. Tasker allows the user the ability to set up tasks for a website including the identification of scenarios of entry and task research to see if different scenarios are being satisfied.

Action tracker is used to measure workflow for revisions or other tasks associated with the collaboration services. The tracker can be used to track organizational issues and requests for problems for a particular customer and provides managed workflow for an organization. Search services provide a comprehensive search option capability including search operators and dictionaries forΔ customer's website. The search service tool can be inserted as a script in a customer's website. Once inserted, the service runs on a remote server, but provides search results to the website's visitors. The search service delivers search results and queries users as to their satisfaction with the results. The feedback can be used to tune the search engine response. The website operator can use result statistics and performance ratings (feedback) to improve the user's website experience.

Morale meter is used to measure the morale for a team. The morale meter provides a mechanism for team members to report their morale using one or more graphical indicators, so that managers and teammates can assess the morale of the team and take action as appropriate to boost team morale.

Organizational tool provides tools for polling employees of an organization and delivering aggregate team member assessments of company or project progress and organizational efficiencies.

Product tool provides a mechanism for engaging customers in an ongoing dialogue about company products. The tool provides product designers and management better insight into their customers needs and desires and includes feedback on products and areas for improvement.

Service tool provides features similar to the product tool, but is directed to company services.

Research may be gathered about a customer's own website or alternatively, may be gathered about another website. Customer rating research can be run against a competitor's website. Customer rating information can be retrieved for the competitor's site by doing the same kinds of screening both based on index as well as benchmarks. In the second method, the particular site is identified and participants can be selected from the visitors to the customer's website or by other means. For example, a participant may be located at a customer's website and then, there afterwards, may be asked questions about a second or different site. A pop-up menu may be used at the site location to query a user as to whether or not a user may desire to be a participant in the study. The present invention has been described in terms of specific embodiments, which are illustrative of the invention and not to be construed as limiting. Other embodiments are within the scope of the following claims.

Claims

WHAT IS CLAIMED:

1. A method for automatically gathering research data for a website comprising: identifying questions in a question set that are a measure of a research metric to be gathered for the website; embedding one or more instructions in a first question in the question set, the instruction to record participant response data using a proxy server disposed between the participant and the website; initiating a research test including providing a first question in the question set to the participant; and gathering response data as each question in the question set is answered by the participant.

2. The research method of claim 1 wherein the research data can be selected from usability, product and organizational research.

3. The research method of claim 1 wherein the gathering step is performed without a permission of a website's sponsor

4. The research method of claim 1 further including forming questions in a plurality of question sets that are a measure, respectively, of a plurality of research metrics to be gathered for the website; and partitioning each question set into a plurality of subsets of questions.

5. The research method of claim 4 further including only presenting a subset of questions for each question set to each participant including gathering responses from each participant to the questions presented for each research metric; and combining the participant's responses to produce a plurality of indexes, one for each of the research metrics.

6. The research method of claim 4 further including only presenting a subset of questions for one or more question sets to each participant including gathering responses from each participant to the questions presented for each research metric; and 5 combining the participant's responses to produce a plurality of indexes, one for each of the research metrics.

7. The research method of claim 1 further including at the participant's web browser, extracting the embedded instruction from the first 0 question received and generating a request to load a page from the website including the embedded instruction; transmitting the request from the participant's web browser to the proxy server; and enabling a recorder at the proxy server to record participant response data.

5 8. The research method of claim 7 further including including with the embedded instruction an IP address for the proxy server and a code designating the website; determining the IP address for the website from the code; and forwarding the request to the website. 20

9. The research method of claim 8 further including receiving back from the website a page responsive to the request; and forwarding the page to the participant's web browser.

25 10. The research method of claim 9 further including prior to forwarding the page to the participant's website, embedding in the page a code to measure a metric.

11. The research method of claim 10 wherein the code includes an engine for recording participant responses generated at the participant's web browser.

12. The research method of claim 11 wherein the participant responses are selected from the group of audio, video, gestural, facial, mouse, pen, joystick and keystroke responses.

13. The research method of claim 10 wherein the code includes an engine for capturing a screenshot of a user interface presented to the participant.

14. The research method of claim 1 further including embedding one or more instructions in a last question in the question set, the instruction to stop recording participant response data using the proxy server and to initiate the download of the gathered response data to a centralized research site for analysis.

15. The research method of claim 1 further including incentivizing a candidate to become a participant including identifying a reward to be given to a participant upon completion of participation in the research test.

16. The research method of claim 15 further including characterizing research participants into one or more types and customizing the reward for each participant type.

17. The research method of claim 1 further including storing historical results for the research metric and a representation of the website or portion of the website to which the metric pertains; determining that the website has changed; gathering new responses from participants to the questions presented for the research metric; combining the participant's responses to produce an index of the research metric for the changed website; and comparing the indexes for the website against the changed website to measure a change of the research metric between the two.

18. The research method of claim 1 further including recording at the proxy server each page request received from a participant.

5

19. The research method of claim 1 further including recording at the proxy server each page received from the website prior to forwarding the page to the participant's web browser.

10 20. The research method of claim 1 further including recording participant response data at the proxy server; and after the research test has been completed by the participant, downloading the participant response data to a centralized research site for analysis.

15 21. The research method of claim 1 further including compiling answers to each question at the participant's computer; and after the research test has been completed by the participant, downloading the participant response data to a centralized research site for analysis.

20 22. The research method of claim 1 further including compiling answers to each question at the participant's computer; and forwarding the participant responses to the proxy server for recording; and after the research test has been completed by the participant, downloading the participant response data to a centralized research site for analysis. 25

23. The research method of claims 14, 21 or 22 further including analyzing the participant response data including playing back the participant response data in real time.

24. The research method of claims 14, 21 or 22 further including concurrently playing back the response data and a visualization of the website at a centralized research site to analyze of the response data.

5 25. The research method of claim 24 further including playing back the participant response data at a faster rate than a recording rate for the participant response data.

26. The research method of claim 25 where the faster rate is twice the recoding rate. 0

27. The research method of claim 1 where the proxy server is located at a Internet service provider accessed by the participant for linking to the website.

28. The research method of claim 1 where the proxy server is located at a hosting server that 15 couples the website to the Internet.

29. The research method of claim 1 where the participant is coupled to the web site using a PDA and the proxy server is located at a PDA proxy that is accessed by the participant for linking to the website.

20

30. The research method of claim 1 where the participant is coupled to the web site using a PDA that includes a PDA proxy for linking the PDA to the Internet, and where the proxy server is coupled between the PDA proxy and the Internet.

25 31. The research method of claim 1 where the participant is coupled to the web site using a wireless telephone and the proxy server is located at a WAP proxy that is accessed by the participant for linking to the website.

32. The research method of claim 1 where the participant is coupled to the web site using a wireless telephone that includes a WAP proxy for linking the wireless telephone to the Internet, and where the proxy server is coupled between the WAP proxy and the Internet.