The delights and dangers of ‘data-driven’ anything
Most brand strategy boils down to three/four axes:
Category truth
Consumer truth
Company truth
Agencies like Droga5 also use Cultural truth (i’m a fan)
Which raises two interesting questions:
What is truth?
Can you have more than one?
My take: truth is highly contextual and yes multiple things can be true at once. Quantum much? Yes, but also, consider Nassim Taleb:
“Statistics is a knife that cuts on both sides.”
By which he means you can use numbers to tell many types of narratives. Have a big enough data set and you can prove that people both love and hate a product, depending on your agenda. Or that having lizards as a pet is both a good and a bad idea.
The question then is, which version of the truth to you adopt?
You adopt the one that gives you the largest chance of competitive advantage. Which means data-driven approaches are mostly a performance. Data is a Rorschach test, we see what we want to see. The same report can yield wildly different interpretations.
The bad news: post-truth, oh noes. The good news: we get to choose which version feels most compelling. We obsess so much with being right, we forget the point is to broadly be less wrong.
So your data doesn't always need to undoubtedly confirm what you're saying, as long as it doesn't deny it either. From there, you can shape the story how you see fit. Narrative cuts through harder than pure numbers ever could. Always has. Always will do. Sorry.
