Don’t Let Data Strangle the Insight
One of the best consumer insights I’ve ever heard belongs to Pepto Bismol.
The delightful observation that we feel like victims to our headaches but culprits to our stomachaches.
For a long time, this led to creative work that was designed to make us laugh — to give ourselves a break & be part of the “trouble” that got ourselves into. Irrelevant work, perfect from a weird, bright pink medicine brand.
That insight (and therefore, that work) would be strangled today. Killed by the availability, abundance and religious zeal to (quantitative) data.
When you have all the data in the world, you are lulled into believing that the insight is always there — it’s just a matter of unburying it.
Make a chart. Run a correlation. Throw machine learning at it. Design a model. Make a word cloud from social listening. Recut the audience. Use 1st party data instead of 3rd party data. Do the right pivot table and the data will reveal the insight you’ve been looking for.
This happens exactly zero times. Ever.
(If it has for you, please tell me how. Seriously, I want to be wrong).
The sheer quantity and sophistication of data available today is an illusion. One that has blinded us to a simple reality that data captures existing attitudes and behavior. It is, by definition, a lagging indicator and therefore, cannot tell you how to create a future that doesn’t yet exist.
Can it help optimize, yes.
Squeeze out incremental growth, most definitely.
Offer interesting hints, for sure.
But actually light the way to business-changing creativity?
In my humble opinion, no.
I am not a data deny-er. Or a luddite. Or shaking my fist at the Excel.
I love data. It makes me feel scientific. It makes me sound smarter than I am.
I’ve just come to see how dangerously seductive it is.
How easily it can lead to solving ever-increasingly clearer, more statistically significant but smaller, less important questions.
I’ve come to realize that squinting harder at data does not make it more useful. The kinds of answers I’m looking for simply aren’t in there.
Don Draper is dead. Google Analytics killed him.
Brands also underestimate their own role in creating attitudes and behaviors that do not yet exist (and therefore, cannot be captured in any data set). Consumers can be taught something new if, for whatever reason, it makes sense to them and makes them feel something.
A brand convinced humans to believe an abundant, colorless rock represented eternal love and was worth 3 months’ salary.
A brand convinced humans that something that used to be free should be bottled and cost more than gasoline per liter.
A brand convinced humans that air, when injected into a shoe’s sole, increased the shoe value to $200.
If I were the planner clever enough to find and articulate the Pepto Bismol insight today, I would be immediately asked for the data to validate it.
Prove it or it can’t be true. What then?
Conduct a multi-market survey asking people how they feel about diarrhea? On a scale from 1 to 10, how responsible do you feel for your stomach aches? Would a human even answer that questions honestly? Or know how to answer?
Look at GWI for an over-index in over-eating guilt? The closest I could get would be generic attitudes towards health and wellness. (I just tried).
Look at in-coming traffic data to PeptoBismol.com? Who the F goes to PeptoBismol.com when they’re about to crap themselves (prospective, new consumers)? Who the F goes to PeptoBismol.com after they’ve crap themselves (existing, repeat consumers)?
Another sin is conflating data with numbers.
As the old saying goes, just because it can be quantified, doesn’t mean it’s important and vice versa.
History is data. A good movie is data. Overhearing a conversation is data. A joke is data.
Even so, data may hold all the clues.
But it will never hold a truly great answer.