Incentives: The Right Ones Can Work!

While many of our online studies source sample from online panels, we also do a substantial number of studies using client provided opt-in customer lists. Not that long ago, customers willingly completed a survey, purely based on their relationship with the client company. But these days, customers have come to expect an incentive – either in the form of a sweepstakes or as an individual gift. Today I want to talk about individual incentives.

Incentives serve as a nice perk for respondents and increase the likelihood that they will take a survey. However, it’s crucial to consider various ramifications that may come into play. Many clients are tempted to give an incentive that appears innocent, and to be a cost saver for them, but that can actually introduce a major bias. For example, if a restaurant chain wants to give a “free appetizer” incentive – what about the customer that is so disgruntled that they would never revisit your restaurant. The free appetizer would not appeal to them, and you would never get their valuable feedback. The same thing goes for the CPG firm that wants to give one of their products as an incentive. These types of incentives are very effective for an advisory panel, but not for your typical survey.

Another thing to keep in mind when it comes to incentives is to offer something broad and neutral, which can accommodate anyone. Let’s say you were to offer a gift certificate to Bed Bath and Beyond. This prize would most likely appeal to a certain demographic, aka soccer moms, as opposed to young male teens. Thus, you have biased your sample (unless your customer base is only comprised of soccer moms). Something like an Amazon gift certificate would appeal to a broader audience. We have tried other incentives, such as Gap, LL Bean, or Starbucks gift cards. Each of these can have a biasing impact. And of course cash is very neutral, but we have found gift cards much easier to administer.

So the next time your study calls for an incentive, think about any potential bias it could introduce.

"I Want to do a Segmentation Study..."

It is fairly common for me to get a call from a client, or a potential client, saying “I want to do a segmentation study”. Naturally I get excited at the prospect of a new study, and segmentation studies are one of my favorites. But these days, I have learned I need to ask what they mean by a segmentation study.

To a traditional researcher, segmentation means you start with a clean slate. You ask a lot of behavior and attitudinal questions appropriate for your category, along with extensive psychographic, lifestyle and demographic questions. Then you use a multitude of multivariate techniques such as cluster analysis (hierarchical and partition), tree analysis (such as CHAID, CART), K-means and Latent Class segmentation to segment the data. A typical approach may result in 3-4 segmentation solutions. Selecting the final solution requires a mix of art and science. The result is unique segments; with complex, distinctive profiles and personalities. The marketer quickly gets to “know” these personalities. They learn who represents their best opportunity, and the messages that will best resonate with them.

Now quite often, “I want to do a segmentation study” does not mean what I described above. In these instances, marketers typically have pre-identified groups that they want to understand in more detail. It may be as simple as males vs. females, teens vs. adults, brand loyalists vs. brand dabblers, or techies vs. technically challenged. The predefined segmenting is usually on some behavior or identifiable trait, and not based on attitudes or psychographics. Basically, we are just “slicing and dicing” the data by these groups, and doing profiling analysis. To limit confusion with traditional segmentation studies, we typically refer to this type of a study as a Market Profile Study, or Market Analysis.

Both of the above mentioned approaches have their purpose, and can be extremely valuable to a marketer. For now, when I get that call about a segmentation study, I will still be asking “what type of segmentation do you want to do?”

Wow - a first post!

With this post we join 175,000 other blogs started today on a journey into the future of communication. We intend to use this medium as a forum for insightful conversation about marketing research, and the world through the eyes of a researcher.

I'm excited about our experiment, and I hope you'll join us.