From the July 2001 issue of Entrepreneur

Would you like to monitor the effectiveness of your Web site? Do you want or need to know whether customers are going to like the new product you're thinking of developing?

One of the best ways to get answers to these questions is to do a survey. But don't bother with traditional polling via phone or mail if you want results. You can ask everything you need of your customers through your Web site. "The biggest reason companies survey online is because it's fast and efficient," says Chris Anne Wheeler of marketing research company ActivMedia. "You can automatically calculate results, and if you [run into] problems, you can fix the study and relaunch quite painlessly."

Furthermore, online surveys are inexpensive. According to Tim Lee, co-founder of WebCMO.com, a site for Web marketers, and editor of the Journal of Marketing Research, most online surveys qualify as do-it-yourself projects. That means you don't need a professional firm to conduct market research surveys for you. "After all," he says, "you understand your market better than anyone."

To conduct a successful online survey, be sure to keep the following tips in mind:

1. Do the survey yourself. Just because you forgo hiring an expensive online research firm doesn't mean the results will be any less effective.

2. Determine the survey's goal. Focus on gathering the information most important to you, such as navigation satisfaction, product and service needs or price expectations.

3. Keep the survey short and only ask questions that pertain to the topic at hand. If the question won't yield insight into the topic, eliminate it.

4. Select your survey channels. Will you collect information from visitors right on your Web site, or will you send e-mails asking for information? Both options are effective.

5. Test the survey before launching it fully, and evaluate the results you receive. You'll then be able to modify any questions that aren't returning the expected results or that people skipped because they were worded poorly.

6. For quantitative answers, don't overlap the survey categories. (For instance, 1 to 99, 100 to 199 and so on). Also, keep the number of items within each category the same to yield more statistically valid results.

7. Always give the person the option to check "don't know."

8. Keep it simple and avoid complex questions that require too much thought.

9. Ask for personal data last. By doing so, your respondent will be more at ease, and you'll more likely obtain that critical information.

10. Offer an incentive related to respondents' interests. That way, they're more likely to complete the survey.

11. When selecting incentives, keep them unrelated to the survey questions. For example, if you're trying to measure your customers' attitudes about price, offering a discount as an incentive will attract people who seek out low prices-which will return biased results.

12. Keep your target audience in mind. If you want to understand your current customers or site visitors, conduct a survey on your site or use your mailing list. If you want to understand online consumers in general, buy banner space on other sites or use outside mailing lists to get respondents.

13. Remember simplicity. Though sophisticated marketing research requires advanced analysis, less complex techniques (cross-tabulation, for example) can also provide valuable information.

14. Do market research for the right reasons. Conduct a market research survey because you need key information to make your marketing decisions. That way, rather than just having a stack of tables with numbers, you'll have research results you can use to improve your business.

Giving in to the temptation of getting help from an online research company could cost you $15,000 to $100,000. Doing an online survey yourself keeps that money where you need it-in your business.