Maximizing respondent cooperation

Editor’s note: Keith Price is senior vice president at Greenfield Online, a Wilton, Conn., research firm.

Today’s marketing researcher has a great variety of methodologies to choose from to obtain market intelligence, and clearly online research has many advantages. By 2006, reports Inside Research ’s Larry Gold, more than 33 percent of overall research spend will be done online.

Why has online research quickly gained such acceptance over other methods for collecting attitudes and opinions? In addition to timing and cost efficiencies, there is the remarkable ability to collect information about a wide variety of topics, from household purchases to health issues and ailments from low-incidence, geographically-diverse respondents. Second, respondents can answer surveys thoughtfully during leisure time, when they are uninterrupted, which can yield robust and well thought-out responses. By coupling the ability to target virtually any respondent group and enabling them to respond at any time of the day, no other methodology rivals the Internet.

Researchers are migrating studies online at a rapid pace for many reasons, in addition to the aforementioned benefits. The vast majority of consumers are online and the industry is no longer skeptical about the validity of Internet-based work. Currently more than 60 percent of the populations in both the U.S.  and Canada have access to the Internet.

We have found that consumers want to “speak out and be heard” and appreciate the convenient, and non-intrusive nature of online surveying. We know what works and doesn’t work in designing surveys for the online environment and encourage our clients to design surveys with the respondent in mind. The following is a list of Top 10 tips that we recommend you adhere to when conducting an online interview.

The efforts we - and you - can put in on the development of any online research project invariably pay off with improved speed and efficiencies, better cooperation rates and better research results overall.

1. Ask yourself: would you enjoy taking this survey?

Remember, the Internet is a self-administered environment. Unlike a telephone survey, online respondents are required to read survey instructions and answer questions themselves. To avoid confusion and frustration, make sure that questions are logical and consistent. Can technology applications be included that enhance the overall experience? Have you over-engineered the study design?

Here are some ways of enhancing questionnaire design:

  • Place screening and quota management questions up front. No one likes to answer 25 questions only to find out they don’t qualify for a study.
  • Use drop-down lists and numeric open-ends sparingly - they take longer to answer.
  • To minimize respondent fatigue, place at least three or four questions on a page.
  • The maximum number of product or service attributes shouldn’t exceed 20.
  • Always have clear directions on how to answer the question, i.e., “Select one,” “Select all that apply.”
  • Provide an escape - offer a “don’t know” or “prefer not to answer” option.

2. Remember: the respondent is working for you!

Respondents have busy lives. Before they even receive an invitation to take your survey, they’re responding to e-mails, phone calls and handling a myriad of daily chores. Taking surveys should be enjoyable. Make things easy for them with clear, explicit instructions and useful, comprehensible error messages.

3. Design attribute ratings for visual ease of use.

Attribute ratings scales should appear once per page, usually along the top or left-hand margin of the survey. If the respondent must scroll down to answer all questions, keep the scale visible on screen.

4. Beware of long surveys.

A lengthy survey is a daunting task for anyone - no matter how interested they are in the subject matter. The average respondent begins to tire after 10-15 minutes. The longer the respondent participates, the more likely they are to terminate before completion. Consider which questions are the most critical and look for ways to enhance the survey with gaming exercises or visuals. Provide rest stops in the text and allow for transitions from one question set to another.

5. Pay attention to specific methods and audiences.

Many targeted audiences require extra attention, whether you are working with a difficult sample, or need a long-term commitment from the respondent. For example:

- Teen studies: Be mindful of the need for exciting, appealing graphics. Ask relevant questions - queries about income or life insurance don’t belong.

- B2B studies: These are busy people. Stick to the relevant topic and offer creative incentives for study cooperation.

- Product placement studies: Respondents like these types of studies. When mailing, always remember to include all components and provide clear instructions on product usage.

6. Be honest about sensitive information.

Provide respondents with a factual privacy policy, informing them as to how the information they provide will be used. When asking respondents to provide sensitive information, tell them why you need it. This increases the likelihood that they will share this type of information.

Finally, don’t ask irrelevant questions. Ask only what you need to know.

7. Incentives: pick the appropriate incentive for the right audience.

Greenfield Online conducts thousands of studies each year, and we have developed metrics that help us to measure the effectiveness of incentives. Realize that lower-incidence groups and lengthy surveys require more generous incentives - especially if quick turnaround is required.

8. Avoid dropout triggers.

Open-ended or sensitive questions asked at the beginning of the survey can be jarring to the participant and increase the likelihood of dropout. These types of questions are better placed towards the end of the survey. Ultimately, consider providing the respondent with an option not to answer.

To further avoid dropouts, ask only relevant questions of the target group. Respondents can tell within five to 10 questions if the survey applies to them. If the first few questions are vague or irrelevant you may increase the dropout rate.

Finally, be straightforward about survey length and provide a progress meter; they’re both appreciated. If you tell a respondent that they will be participating for five minutes and your survey runs eight, you increase respondent frustration and encourage dropout.

9. Mix methodologies to reach respondents.

E-mail is a fast and cost-effective means of inviting survey participants, however, sometimes it makes sense to use other data collection methods, such as telephone or interactive voice response along with e-mailed invitations, increasing the number of touch points and ensuring survey participation.

10. Use Flash technology to enhance certain exercises.

Roughly 10 years ago, when the Internet was in its infancy, Web sites were nothing more than static pages with limited functionality. Interactivity has grown to the point where now most Web surfers possess the processing power and bandwidth to take full advantage of full-motion, high-speed interactivity. Purchasing can be done with a click of a button, and commonly-viewed sites can be customized to meet individual needs and preferences.

Online surveys have charted a similar course, from “electronic paper” to early interactive surveys to the point where online research is now ready to embark on a fully interactive journey. Researchers are taking full advantage of this new era, with engaging new technology tools and enhancements designed with one goal in mind: high-quality, actionable data.

Consumer insights that were once considered cost-prohibitive or nearly impossible to obtain offline, such as testing commercials or simulating an in-store shopping experience, are now being executed via the Internet using cost-effective Flash and HTML technology. What’s more, these advanced technologies can be used for standard surveys.

Consider new technologies like card sort. Card sort can improve concept, text and graphic ranking. Respondents are asked to place cards in order of preference, in a more robust fashion than a traditional ranked list. Card sort has been especially effective on ranking attributes lists or images. The technique elicits careful responses to lengthy attribute lists and provides enhanced “think-through” and respondent enjoyment.