Editor's note: Teri Kaslow  is executive director of decision science at Deluxe Corporation, a Shoreview, Minn., business development company. This article appeared in the August 23, 2010, edition of Quirk's e-newsletter.

In addition to low-cost online and e-mail survey tools that allow companies to design and conduct their own survey research, niche social networking sites are growing in popularity among marketers for their ability to target specific demographic, firmographic and interest groups. Examples of niche social media sites include Slashdot (tech nerds), RealEstateVoices (real estate gurus) and ShowHype (movie and entertainment fans).

Enterprises that are looking to survey a specific household demographic, business firmographic or interest group can now look to niche social media sites to gain attitude and behavioral insights about their target market. And with the size of these niche sites growing daily, businesses may feel more comfortable deploying surveys as the sample size increases.

So what are some of the opportunities and challenges for marketers when conducting surveys on niche social networking sites? Specifically, what should market researchers keep in mind when conducting a niche social networking site survey?

As precise as possible

First, consider the style of questions when developing an online niche survey. The overarching principle when designing questions is to be as precise as possible. In other words, make sure that these questions tie in to the focused objectives that the survey is expected to answer. Remember that not all clients are aware of your products or services, so create simple questions and avoid the use of any jargon. You may have to provide product descriptions or additional details in order for this audience to understand your business. For those clients who are aware of your products or services, be sure not to ask questions that you already have gathered information on. For instance, these clients may be registered members of your Web site, which means that you would likely have access to specific household demographic information including gender, age, number of persons in the household, etc. Similarly, if you are in the business-to-business space, you may already have retained information on the industry of the registrant, number of years in business, number of employees, annual sales volume and additional firmographics. If demographic or firmographic data gaps exist, be aware that third-party data providers are able to append additional information to help supplement your stored data internally as well as the survey responses.

Offer a variety

Secondly, consider the form or type of questions when designing your survey. It's no secret that building a survey questionnaire can be difficult. Marketers often wonder if they should include more open-ended questions than closed-ended questions or vice versa. The best practice in this instance is to offer a variety of questions, of which the majority should be closed-ended. It's important to minimize the number of open-ended questions because of the raw grunt work involved to analyze the responses. Sifting through verbatims is too time-consuming for the analyst, but may provide useful qualitative insights. Text mining software can help by highlighting interesting findings such as the most popular buzzwords in the responses. Unfortunately, many market researchers are keenly aware of the fact that many respondents tend to inflate their answers and stretch the truth. This is in sharp contrast to surveys conducted via telephone or face-to-face. To avoid this, consider changing up the order of questions and inserting additional questions to double-check their answers.

Golden questions

Similarly, it's important to think in terms of golden questions. Golden questions are strategic questions that will help your company uncover important information about a customer.  What is the purpose of golden questions with your customers or prospects? These questions help you better qualify the respondents for your product usage; determine their readiness to buy; help you understand which need or pain point they are experiencing; and can help you categorize them into actionable segments for marketing initiatives. In this age of social media, consider asking questions about social media usage. This is another great way to understand the attitudes and needs of your customer.

No longer than five minutes

In terms of survey length, it's important to minimize the number of questions asked. The general rule of thumb is that the survey should be no longer than five minutes because a consumer's attention span is extremely short. Before launching a survey, test it out with a few customers or prospects from your targeted audience or with co-workers. In doing so, you will have a better idea if they are able to complete the survey within your set of time constraints, as well as if they found any necessary edits to make the questions and answers more clearly stated.

Statistical design

It's also important to think in terms of sampling design - will you offer the survey to each person who visits your site or will you only offer it to one out of 10 visitors? Or will you offer it to the visitor who has opened the highest number of pages? Remember, think of the best statistical design to achieve the most relevant results on the back end, as you will want to obtain enough responses to have statistical significance. This means that you may have to have your survey in-market for a longer duration until you reach an amount that correctly reflects your targeted audience.

Value their input

Let them know how much you value their input and how their responses can improve your Web site or product. In the case of a customer satisfaction survey, a customer will likely understand how his or her responses will help improve your service or product. Be sure to ask questions that reveal insight into the entire purchase process, from start to finish. You should also consider asking standard benchmark questions, including how often they shop at your competitors. As a result, you'll have a better idea of where their dollars are going. Customer satisfaction questions work best if given in a scale format (e.g., from 1 to 5) as it is easy to analyze the response data between questions and within different audience groups. Make sure that the scale format is clearly defined. 

Modify and refine

Unlike direct mail and other conventional surveys, the beauty of online surveys is their adaptability. It's simple and easy to modify and refine questions as you go. For instance, after the survey has been up and running for a day or two, you can check in on the status of the survey to determine if - and how many - respondents have taken the survey. If responses are lower than expected, consider revising the questions to make sure they are not as complex, or cut back on the number of survey questions if you are seeing a high abandonment rate. Moreover, an initial read of the responses may uncover answers being completely different from what you hypothesized, which may require you to reword or reformat a question to be more clear or to add further questions to do a deeper dive into the topic. Even after refinement of the survey, you may still experience lower-than-expected response rates. Additional options would be to consider sending out reminders to the audience to complete the survey within a specific time duration or to provide an incentive (e.g., monetary offer, free download, product coupon) as respondents will more likely complete the survey if they receive a benefit. Another common practice today online is to have the survey go viral by suggesting customers share the survey with their online network.

Reap the benefits

Analyzing survey data is tricky but rewarding as you uncover insights to your objective goal. Be sure to let the survey run its course to ensure the the correct sample size, but if you do the proper planning upfront with survey design and sampling, you will likely reap the benefits on the back end.