WebSurv is the Web equivalent of a CATI system. In fact, it's based upon our TelSurv CATI system which in turn is based on ABstat.

Instead of interviewers calling people over the phone, WebSurv awaits its respondents on a website (yours or ours). The system can currently handle up to 128 simultaneous survey projects. Each survey is comprised of one or more frames of questions. For larger multi-frame surveys, skip patterns are possible. That means that the next set of questions the respondent sees will be custom tailored to him/her depending upon how one or more previously asked questions were answered. It's also possible to randomly rotate questions and answer choices. That means we can present questions in random order where appropriate to help remove bias resulting from their always being asked questions or being given possible answers in the same sequence as could happen if a particular brand name is always mentioned first.

Surveys can be quite small or quite large, up to 512 questions. Questions themselves can be numeric, text (open-ended), multiple-choice, or all-that-apply.

Statistics are kept on length of time for completion of each survey response as well as time the respondent has to wait on the system, the date and time the survey was started, etc.

At the back end, the automated reporting tools of ABstat are available at all times. Typically, a base set of reporting is set up so that as the survey progresses, results can be monitored. Each respondent's record is available as soon as it is completed. The monitored information is sometimes used to help with ensuring that segmentation quotas are reached.

When all responses are in, the real back-end work begins. Data cleaning is accomplished with the help of ABstat and the full reporting is accomplished. This is also done using the automated reporting capabilities of ABstat. Typically, a set of banner reports is produced for each major "profile variable". From this set of reports another automated system is available for producing charts which can be used for analysis as well as incorporation into a final written report. For Web surveys it is often desirable to make results, including charts, available over the Web. The automated capabilities of the system make it possible to update the results as often as desired and presented over the Web.

You can see all of this in action for yourself: see Sample Survey or Sample Results.

The second way is to license WebSurv to be installed on your own website. The system is based on a Windows NT platform. We have both 16- and 32-bit versions and have clients using Netscape and Microsoft Internet servers. There's a bit more involved with installation and doing your own survey setup. But if you plan to be doing quite a few surveys, this is going to be the best and cheapest way to go. (Note that even if you choose to run it from your own website, you may still want to do your first survey on our site as a "test run".)

WebSurv is a high-end market research tool: a serious tool for doing serious market research. Cost is $12,000 which includes one year of technical support and also allows for some customization. You may also want some on-site help with installation or training. Over the next few months we hope to have more stripped down versions of the product ready for mass marketing. They will be less capable and less expensive.