Marketing and Research Consulting for a Brave New World
Subscribe via RSS

Marketing research is obsessed with data quality, and so we tend to tread lightly at using non-survey data streams that we don’t understand as well, such as digital and social data.

And yet survey-based research programs like brand trackers with their precise sampling frames get really tiresome because they don’t produce much that is USEFUL.

I suggest we need to ask a new question.  Don’t just ask if the data are good, ask, “Are the data USEFUL?” (at producing better marketing outcomes). Then ruthlessly harness and combine useful but non-traditional data sources…digital, social, survey, location, weather…whatever…and make them work together as usefully as possible!

Perhaps nowhere is the distinction between usefulness and accuracy starker than with target segment profiling data.  For example, at any point in time, about 5% of vehicle drivers are “auto intenders”, that is, they are actively considering buying or leasing a car.  Now, I’m sure that 10 or more sources for auto intenders exist in the ad tech ecosystem.  Will any of the data be accurate?  Probably not…in the sense that it is unlikely that the majority of those who are tagged as auto intenders are really looking to get a car. But are the data USEFUL? YOU BET! Surveys can help you spot the best sources for auto intenders but the segment at scale delivers the payoff. If you have a segment where 20% are REALLY looking for a car, the segment is up to FOUR TIMES more useful than random ad placement by employing programmatic advertising or via matching to linear TV viewer profiling.  Targeting this segment of auto intenders will produce much higher marketing ROI…VERY USEFUL!

Increasingly, data sources can fail a classic quality test in terms of sampling theory but succeed on a usefulness test. Unless you are thinking this new way, you will have a blindspot that leads you to miss out on really valuable and useful data streams. Some examples…

Ad tech targeting data. You can’t get the incidence of a behavior or intention from ad targeting data but the best sources can produce enormous lift in advertising ROI.

Frequent shopper database. No it does not grab every shopping transaction or cover every retailer.  So are the data perfect? No.  But are they useful? ABSOLUTELY! Earlier this year, I worked with Viant and NCS to prove that targeting strategies can easily double your return on advertising! Why wouldn’t every CPG marketer do this?

Social media data.  Streams of conversations are underutilized because the sample is unknown and it is messy business turning unstructured conversations into quantitative data. However, if done right, it is useful for revealing insights you never would have seen. And it has forecasting value, as proven by the WOMMA marketing mix modeling study conducted with the cooperation of the social media intelligence firm I advise, Converseon.

Other blindspots where I find that data are underutilized include receipt scanning, smart TV data, and location data.

Challenging ourselves by asking if the data are useful, leads to new ideas for making traditional research more valuable when partnered with other data sources.

Segmentation research.  Integrate surveys with ad tech profiling data to create targetable segments that produce higher marketing ROI.

Concept testing and trial forecasting research. Integrate concept testing with media data to guide new product launch plans and improve the odds of success.  That is the goal of MoreCastR, a method I created in conjunction with Marketing Evolution.

Brand tracking. Integrate well-designed surveys with digital and social data sources to transform a boring retrospective tracker into a dynamic and predictive brand guidance system.

Asking “could certain data that are not perfect still be useful?” re-focuses us to think about the real goals…not trendable trackers with little value, or accurate facts on what is “nice to know”  but improving marketing ROI while creating growth.

Surveys can help to validate if ad tech data are likely to be useful, and ad tech data make marketing research valuable. A new approach that starts by asking, “are the data useful?”

*the idea for this blog came from a discussion I had with my long-time friend and colleague, Frank Cotignola

 

Tags: , , , , , , , , , ,

Comments

One Response to “Marketers…don’t just ask if the data are good, ask are the data useful?”

  1. Gerard Broussard

    Well said, Joel! It’s time we get pragmatic about smartly using the data that we have versus wishing for perfection.