Editor's note: Aleksandra Polovina is market intelligence manager at Kelley Blue Book, an Irvine, Calif., automotive research company. This article appeared in the July 26, 2010, edition of Quirk's e-newsletter.

Online market research is becoming more and more prevalent as Internet-facing companies query Web site visitors about what they want and need. The two most commonly used methods of researching users of Web sites - online panels and site intercepts - differ in two major ways:

1. Individuals in panels have signed up for the privilege of taking surveys and are fully expecting to receive them.  This is not the case for Web site respondents, who are visiting the Web site to accomplish a task.

2. Panelists are typically compensated for survey participation (i.e., monetary incentives, points that allow them opportunities to win/earn monetary incentives, gifts, etc.). This compensation - to some degree - motivates them to finish a survey they have started regardless of how excessively long that survey might be. This is not so for the Web site respondent, as their sole consolation is often the additional text in the survey invitation that promises to improve future Web site experience based on survey responses.

So how can a company gather this all-too-crucial, in-the-moment feedback from customers and Web site visitors without prior consent or an incentive? Try implementing the following five strategies to help gain more insightful information from online visitors without damaging their experience on the site.

Keep it brief. Research companies can get away with 20-30-minute online questionnaires because panelists are prepared and rewarded. However, if you were to launch a 20-30-minute survey on your Web site among visitors who are trying to accomplish a task, well, you don't need me to tell you what that would do to their user experience. Web site surveys typically should not exceed five to seven minutes for optimal response rates.

Make it sound fun. For panelists already confirmed to be interested in the survey topic at hand, a direct invitation stating the objectives of the study will work well. On a Web site where visitors are trying to complete a task, generating interest is more of a challenge. Thus, invitations that foreshadow a survey that is fun and engaging will work best. Additionally, make a point to communicate the importance of visitors' opinions and how they will stimulate change to the Web site.

Terminate the terminating. It makes sense for market research companies to terminate individuals who do not qualify to take their survey. However, on a Web site, your visitors are taking the time out of their day to help you. Visitors want to be heard and disqualifying them from being able to provide valuable feedback may send the message that their opinions are not important. Respondents may lose their positive view of the company, avoid using the Web site or be discouraged from participating in future market research - regardless of the source. Therefore, for Web site research it's best to let all respondents fill out the survey and filter to the target group on the back end.

Give them a reason. Survey incentives can get expensive and there are a few risks with providing them. If you provide too little of an incentive it may lower your response rate. Provide too large of an incentive and you may influence the responses you receive as individuals feel compelled to provide the information you would like to hear simply to receive the incentive. Remember, the best way to get optimal survey results is to tailor incentives to the wants of the respondents. Human beings in general are interested in self-discovery, therefore invitations that promise to tell them something about themselves will typically increase response rates. For example, if your Web site is in the automotive industry, an invitation that promises to suggest a vehicle that will best meet their needs may be more enticing than a chance to win a gift card. Additionally, sharing preliminary results of the survey with respondents upon completion or comparing their input to other respondents' results anonymously may be motivation enough.

Time your invitation well. Survey invitations are often served at the beginning of a visitor's experience with a Web site (i.e., on the home page, secondary landing page or tertiary landing page). Serving invitations this early in a visitor's task completion cycle can be very disruptive. It's best to serve an invitation upon exit from the Web site or on pages that are known to be at the end of the visitor's task completion cycle. Responses to the survey will be of higher quality if visitors are not inconvenienced by invitations.

Helpful suggestions

These tips are just helpful suggestions to improve a Web site's research strategy. However, not all Web sites are created equal, so be selective in implementing the techniques that suit your project and your corporate style best.