Don't Let Data Drive

If you’ve spent any time in an analytics organization, in a business school, or scrolling through a LinkedIn feed, you’ve seen this quote:

Without data, you’re just another person with an opinion.
W. Edwards Deming

Opinions are a dime a dozen; data is what sets you apart from the armchair executives out there. If you want a seat at the table, you bring the numbers.

Everyone knows that, right? At least everyone with an MBA.

Well, I’ll see your Deming and raise you a Drucker:

Most books on decision-making tell the reader: “First find the facts.” But executives who make effective decisions know that one does not start with facts. One starts with opinions.
Peter Drucker

Hmmm. Dueling quotes. A management theory Battle Royale. The gloves are off, the cards are down…let’s see who takes it all home. Deming versus Drucker, facts versus opinions – who will win?

Neither. Both. Nothing’s set in stone.

Touché.

A word of disclaimer here. When dealing with the natural sciences, physics and chemistry and biology, data can drive – absolutely. You don’t need persuasion to measure acceleration; you don’t need an understanding of psychology to test the tensile strength of a crossbeam. Kirkaldy's Testing Works was absolutely justified in saying that your opinion of the sturdiness of cement has no bearing on reality.

But few of us have the luxury of right and wrong.

The vast majority of us work in the tertiary sector, and thus exist in shades of grey¹. Business, academia, politics, even medicine – these aren’t hard sciences. They may be based on hard sciences, and there are objective elements… but as soon as you introduce those pesky little factors called humans, you’re squarely in the soft-science zone.

Here, facts don’t matter nearly as much as we’d like them to.

Facts and opinions are not opposite ends of a spectrum; they are multi-dimensional Venn diagrams that intersect and interweave and fold in on each other like Möbius strips.

Now, to be fair, both quotes at the beginning of the article are taken out of context. Deming wasn’t actually as dead-set on data as all those print-outs tacked on cubicle walls would have you believe, and Drucker didn’t mean that you don’t need facts – just that you also need opinions to tease out the relevant information from the irrelevant.

Neither management guru was as party-line as the quotes at the beginning of this chapter make them sound, and we shouldn’t be either.

Facts and opinions; logic and gut. By these powers combined, you unleash an unstoppable force – human judgement.

Organizations that deal with humans need humans at the wheel

Good decisions have both an objective and a subjective component. Data alone can’t make a decision. What we call instinct usually has some sense of rationality built in, whether we can articulate it or not.

Why do we accept a black-box model’s output, trained on four years of historical data, but not the senior vice president’s gut feel, trained on twenty years in the industry? We don’t understand the process through which either result was derived.

But it’s safer, career-wise, to side with the algorithm. It gives us plausible deniability. If we do what the numbers tell us, regardless of how flawed those numbers are, we can’t be blamed.

But data and numbers are tools. They’re hammers and nails. If you want to build something solid, they’re necessary – but not sufficient. We still need carpenters.

Organizations that deal with humans need humans at the wheel, not just ones and zeros and regression models. Don’t get me wrong, data should still be in the front seat – but slide 'em on over to the passenger’s side and let them navigate.

Judgement, not data, should drive.

¹ ...and black, and navy-blue pinstripe ↩