Photo Credit: AMC Networks
Brand trackers usually serve the role of ensuring the brand remains healthy. But what if the thermometer is broken and the tracker itself is not healthy?
Here are four ways to ensure your tracker is worth the money...
Design your tracker to trend with the business
Sadly, trackers built on soft brand measures (awareness, preference, love) often find key measures stop trending with the business and eventually the program loses credibility. The proven way to ensure that your measures tie back to market performance is to use a constant sum question (e.g. “allocate 10 chips based on your likelihood to buy each brand…”) to measure each respondent’s individual probabilities of purchase towards each brand in their consideration set. By forcing a respondent to make choices (more chips for one brand leave less for the others) it avoids cultural bias, respondent fatigue and most importantly, can be calibrated back to actual share using a Beta probability distribution adjustment process (contact me for details or follow my blog for a forthcoming description). Then you will have a system where the key metric is solid and the rest of the metrics give valuable insight into the change in performance across quarters.
Make sure the tracker has media planning actionability
If you follow the breakthrough work via the MMA regarding the Movable Middle, (those with a 20-80% probability of buying your brand) you now know that this segment is where profitable response to your advertising comes from. The Beta distribution approach allows you to identify who is in the Movable Middle from the brand tracker, then you can start to profile their media habits, their brand perceptions, and so on. In addition, you might be able to build a lookalike audience for programmatic targeting. This gives guidance on how to reach the Movable Middle and what to say. THAT is useful!
Design the tracker to detect and predict change
One way is to build digital metrics into the tracking system, and not rely solely on survey responses which often have their feet in concrete. Digital metrics such as changes in levels of search activity or social conversation are reflective of not only brand preference but category shopping interest and “shocks”. Especially in seasonal categories, behaviors move a lot so incorporating digital behaviors makes the tracker more predictive and useful.
Don’t fall into the “I Can’t Believe It’s Not Butter” trap
When a brand is known for a given attribute (or it’s built into the name, like the above brand of margarine), standard derived importance regression methods might paradoxically report that the attribute that defines the essence of the brand is NOT a driver! Statistically, this comes about because there is little variance across respondents in ratings on that attribute (like “buttery taste”) but lots of variation in the probability of buying…so little partial correlation. There are ways around this fallacy (contact me for ideas…)
I have been brought in to consult with numerous marketers and research providers when their trackers start breaking bad. Even better would be to design them the right way from scratch following these simple principles.