How to Entice Visitors to Take Web Survey

[hairycoo]

My sites are informational in nature, I don't really sell anything. I'm looking to survey my visitors in order to get a better understanding of:

a. what they REALLY want/need
b. whether they found it on my site or not
c. some demographic data (for a media kit)

There are 2 types of visitors:

1. new
2. returning

I know that new visitors are not as keen to answer surveys without a significant incentive than say more loyal visitors are but they do form the bulk of my traffic. At the moment the issues I'm facing are:

- how short should the surveys be in order to attract more responses from new visitors? 5 questions? more? less?

- should I try to gain some permission first by getting them to subscribe to the newsletter and then ask only these people to take the survey?

- is offering an incentive absolutely necessary or do people respond well to "help us make this website more useful for you by taking our quick survey"?

- if incentives are necessary, which ones work best for info sites? white papers? special reports? On one of my destination websites I can offer high quality, really gorgeous pics that they can print at home. Would this work? Other ideas?

- what about the wording used to entice visitors? What works best?

I also have problems figuring out what questions to ask in order to reach my goal. I have some knowledge of marketing research but not enough so I could really use some tips/advice.

Thanks

[rogerd]

You might make use of a psychological finding from some years ago. People who do you a small favor are more likely to do you a larger favor; psychologists speculate that there's a subconscious mental proces that goes something like, "I did this person a little favor, therefore I must like him, so I'll do this other thing." They original study had people ask a stranger for complex directions. In some cases, they just asked. In other cases, they first asked what time it was. Nearly everyone responded to the time question. Surprisingly, rather than being irritated by another demand, the subjects in second group were more likely to give the additional assistance.

So... if you want to try this strategy, get people to do smoe little thing first - maybe a single radio-button question that is pre-checked for the most likely answer... Then, hit them with a few more questions.

Whatever you do, keep the survey short. If there's a user benefit, make that clear (e.g., respondents will see the survey results). Test different approaches, because what works on one site may not work on another.

[hairycoo]

roger, that's a good idea! I read about that.. I think in Robert Cialdini's book "Influence"... I think I'll have another look through it.