A Case for Directional Sampling

Posted by Mike Irwin on Thursday, April 01, 2010 at 3:27 PM  |  More From This Author »

“Is that a reliable sample?”
It’s a common question we get in developing market research. Many clients mistakenly confuse reliability with statistical validity. If your data is statistically valid, it’s likely (if done properly) that it is also highly reliable. But what if the situation truly requires nothing more than a qualified gut check? These projects still demand reliable research data and can be accomplished when we work towards “directional sampling.”

In a typical sampling, we’d require approximately 385 respondents in order to collect enough independent, objective responses that not only reflect a sense of reality but also deliver high-quality generalizations about what we’re researching. And while there are situations where your project might necessitate a sample size of two-to-three times that magic number, there are also instances where your project could thrive with the level of confidence and reliability that you get with 100 completed responses. And in these cases, directional sampling can deliver the insight and intelligence needed.

Here are a few examples where directional sampling can provide reliable research data efficiently and effectively:

• Reaction to New Creative: If we’re rolling out a new look for a campaign, especially when that look may be a new direction for a client, directional sampling can give us a sense of the target audience’s reaction. Does the ad make a meaningful impression on the target audience? What is the nature of that meaning? What are the key take-aways of the new creative direction? Using directional sampling to gain insight into these questions can help brand managers understand how creative will resonate when it rolls out to the market.
• Key Players in the Market: Maybe we want to understand the general landscape of the market and some perceptions of a set of known providers. If you use directional sampling to gain insight from, let’s say, a hundred local restaurant patrons, you’ll have a good level of confidence in the findings for the local restaurant with the highest quality Steak De Burgo or the coolest atmosphere, or even who has the best customer service.
• Evaluating Sentiment for Polling Issues: Let’s say that a community is about to bring together a vote for a local referendum. To get a general idea of the breakdown of the upcoming vote and, in general, which way the vote may swing, directional sampling could be used to get the pulse of the community.

In each of these examples, Flynn Wright would sample an adequate number of respondents to be able to provide reliable information to make critical strategic decisions. It’s a powerful research tool and provides a unique opportunity to get a macro view of key research questions.