1. Ask the right questions (defining your objectives) ...

What do you want to know?

Just as too many cooks spoil the broth, a "fishing expedition" survey with too many (or fuzzy) objectives will spoil your efforts.

What do you want to know? Do you want to: improve your visitors’ experience at your website? Increase the size of their orders? Discover unmet needs? Increase the frequency of their visits? Follow the KYSS motto: Keep Your Survey Simple. Trying to satisfy too many objectives in your question mix will confuse and annoy your audience and skyrocket the non-completion rate.

2. ... Of the right people (they’re not limited to your customers and prospects)

About and from whom do you want information? Of course, you want to keep profitable, loyal customers coming back. This are the first and foremost group you want to relate to and obtain information from.

What about dissatisfied customers? A separate "complaint" survey form for disgruntled, but still current, customers can do a lot for your bottom line. Contrary to what so many business people fear, most complaints are a strong indication of interest in what you are offering and will usually give you important information on how to beat your competitors. Ask the usual demographic and product questions, or their ID number if you’ve got their basic information at hand. Follow these with "what went wrong" and "how can we help" -- with an indication of when they should expect a response from you. Make sure you follow through, of course. Raised expectations that are ignored are worse than no expectations at all!

There are several important groups beyond your customers and prospects that you ignore at your peril. One group is former customers. Profitable customers who have defected are a very valuable source of strategic business information. While these defectors may not be particularly inclined to visit your website to fill in a survey, they may respond to a brief, polite e-mail survey with an expression of regret, but respecting their decision, and asking just a couple of open-ended questions something like: "Would you tell us the most important factor(s) in your choosing another vendor?" and "Do you have any specific suggestions where we should change or improve?" Of course, you might also want to keep the door open by asking: "May we keep you informed of any changes we make that might be of interest to you?"

A lot of your visitors are surfers who will probably never purchase anything from you, so why bother even considering them? Because some of your non-customer visitors are important influencers. Youngsters, for example, may not pay their own way into Disneyland, but they sure have influence on mom's and dad's decisions on vacations. Likewise, corporate engineers may not be buying your electronics components directly, but they’re the ones writing the specs for their purchasing departments to fulfill. If you don’t ask them questions, you won’t get their answers.

3. ... In the right sequence (encouraging cooperation)

Just as you need to attract respondents to your survey with your introduction and incentive offer, you need to draw them into the survey by starting with the easy questions first: name, affiliation, other demographics. Beware of making this section too lengthy or personal, however, or you’ll lose people before they get to your good stuff.

4. ... In the right way (single choice, "multiple choice," open-ended)

This will strike a lot of big-organization researchers as heresy, but after the demographic information, I like to alternate single- and multiple-choice questions with open-ended ones. The reason is simple. It’s easy for people to mindlessly click on boxes without a whole lot of thought. Most will. While the information you gather from this data will have some generic value to you, you can greatly increase the value of some questions by asking "why do you feel this way?" or some such. An example: "Select the best color printer manufacturer" of a list you provide. Follow it with "Why do you feel this company produces the best printer?" Those few people who have something of importance to say about it ... will!

Just as with traditional -- paper or telephone -- surveys, a lot of small businesses avoid asking open-ended questions because of two perceived difficulties: lack of response, and analysis (simple spreadsheets simply won’t work). As implied above, a few thoughtful responses are better than none at all. And open-ended questions usually can be sorted into broad categories and sub-categories, assigned identifier key words or numbers, and then subjected to routine analysis.

5. ... In the right number and format (avoid those long down-load times, "when-will-it-ever-end" blues)

Have you ever clicked on an online form and waited while it downloaded? You KNEW it was going to take a long time to fill out, and unless you were greatly motivated to start, you probably just clicked the "back" button or got away from the site entirely.

Even worse is the single-question-per-page with the "submit" button just below your screen’s view. Eight to ten questions into the survey, and you have NO idea how many there are to go. Yes, these are great for surveys that branch off into different question sets based on each submitted answer, but awful for completion rates.

Don’t make your respondents feel like Rodney Dangerfield. Give them some respect!

6. ... At the right time (e.g., product opinion when it’s "top of mind")

You might have a special survey URL for on-line buyers and product registrations. Your products and services are top-of-mind when someone’s just made a purchase or registered their retailer purchase at your site. Capture their interested state of mind and presence at your site by inviting them to click on your survey URL.

7. ... To get the right results (is statistical validity still valid in this era of "market size of one"?)

Contrary once more to conventional wisdom, I suggest that EVERY response to your survey is important in this 1 to 1 marketing era. Yes, you need to keep a heads-up on trends. But going for numbers (volume or numerical data) alone simply isn’t taking advantage of the technology available to you. By this I mean eliciting comments through open-ended questions -- questions that your competitors are probably not asking. Only the most mouse-dedicated visitors to your survey URL will resist telling you if something important to them is on their mind. Try it. All you’ve got to lose is some great ideas.