Different Kinds of Surveys

My visit to the ESOMAR Congress in Amsterdam provoked plenty of thought about how we can accommodate different kinds of respondents by offering different survey experiences. Historically, we have been a “one size fits all” industry. Survey design templates are pretty much the same regardless of who is being surveyed, the topic, the study goals.

I can’t beat us up too badly for having one size fits all. After all, Office Depot has the same website for business buyers and consumers. The desktop of Windows Home Edition looks pretty much the same as Windows Enterprise Edition minus a few bells and whistles (and Mac OS X looks exactly the same, as there is only one edition). Companies produce one size because it’s cost efficient, which certainly benefits them but also benefits us as consumers. So it goes with research – if we have too many survey SKUs, as it were, prices will rise to the sponsoring client and incentives will go down to the panel members.

However, recent experimentation with game-ish surveys opens up a path to – and conversation about – a different model for surveys. Most of those pursuing game-ish surveys seem to recognize that this is not the next size that fits all, rather another tool in our kit. I have to believe that having more options for how surveys can be designed can only benefit us in the long run.

Perhaps we can find a middle ground that offers a bit of customization and sensitivity to context: more than one basic design but fewer than a lot. It strikes me that there may really be no more than 3 basic forms of survey:

  • Transactional: in this kind of survey, we assume a certain level of familiarity with the research process on the part of the respondent (e.g. they are panel members, or they’ve completed a satisfaction survey before). Questions are likely to be more centered on facts, behaviors, reactions. The study goals are confirmatory rather than exploratory. This kind of study characterizes a lot of what’s done by high tech companies with IT professionals, or by pharmas with prescribing physicians, or by fast-moving consumer goods with customers.
  • Conversational: in this kind of survey, we may not be able to assume that the respondent is as familiar with the research process. Questions are more likely to be experiential or emotional, and verbatims may assume great importance. Study goals are exploratory. This kind of study characterizes a lot of CEM, marketing or product development and brand work across industries.
  • Diversionary or embedded: in this kind of study, the traditional Q&A approach is de-emphasized by creating or embedding surveys goals in a task that engages people on a different level. Data may be collected through the process as well as through embedded questions. As this is new domain for MR, I don’t think we quite know what the appropriate uses of this approach are. It seems to me that it should work equally well with consumers who are familiar or unfamiliar with research since they will not view it strictly as a research event. If they do, however, view it as research, the picture changes. I can imagine quite a lot of experienced panelists having little tolerance for even a very well-designed game-ish survey. They want to get in, give their responses in a familiar format and collect their incentive, thank you very much.

Limiting us to 3 survey SKUs is a useful thought exercise; I can’t imagine any organization being able to support the design standards and programming requirements of more than 3. No doubt this is one among many possible taxonomies, even with a cap of 3 – send me some more!

Leave a Reply

Your email address will not be published.