No one in our industry gets fired for saying they are data-driven. But maybe they should?
There is plenty of excellent data and analysis in fmcg and retail – and plenty of poor data, poorly analysed. If you are being driven by poor data, analysed poorly, then being data-driven is dangerous.
So, knowing you are working with good data and analysis is critical.
Sorting good data from bad
There are some great data sources. We’ve probably all got examples of Kantar Worldpanel marginally misreading a category or retailer’s growth rate. But step back and it’s a brilliantly managed, high-quality data source that transforms our understanding of what shoppers are doing.
You could say the same about Kantar’s Usage panel: excellent qualitative researchers conducting insightful in-depth interviews and focus groups, and surveys designed and analysed by people with the proper skills and training.
But for every good case, there’s a bad case. The analysis of Kantar or survey data seems to tell us that (guess what?) almost every new product or shopper activation drives incremental sales. The surveys run cheaply through a do-it-yourself panel provider, inappropriately structured and full of leading questions, designed to give the answers that suit the survey commissioner.
These are common problems that cause expensive mistakes and discredit the insight industry. Worst case, they can make good people wonder if all this insight stuff is really worth the paper it’s written on.
Doing due diligence
So what can you do to ensure you’re working with quality data, well analysed?
First, test for expertise. Don’t be shy. If you’re presented with data, ask the detail of what it is, how it is collected and how it is analysed. Don’t be satisfied until you really understand it. With a few good questions, it will be obvious whether the presenter is fully on top of it. If they aren’t, ask for someone who is.
Second, don’t be blinded by science. I’ve rarely met someone who properly understands gains/loss and switching from Kantar, or how incrementality is calculated. Don’t accept something you can’t understand yourself. You might think others understand it, but they probably don’t.
Third, have an expert in your camp. You wouldn’t walk into a legal discussion without a lawyer or make a big financial decision without an accountant. So, don’t be data-led unless you have a trusted advisor who understands where the data comes from and how it should be analysed and interpreted. There’s a reason successful companies invest in highly qualified insight people.
Is it good to be data-led? Depends on the data. Depends on the quality of the analysis.
Jeremy Garlick is a partner at Insight Traction
No comments yet