The projector hummed. It was the only sound in the room for a full 15 seconds, a low B-flat of corporate anxiety. On the screen, two columns of numbers glowed under the header ‘Project Sparrow: A/B Test Results.’ Maria, who had spent the last 45 days of her life nurturing this data, stood silently. The numbers were unequivocal. Option A had a 15% higher conversion rate, a 25% lower bounce rate, and cost 5% less to implement. It was a statistical knockout.

Then, the VP leaned forward, tenting his fingers.

‘Good work, team. Really thorough.’ A pause. ‘My gut tells me B is the future, though. It feels more… expansive. Let’s find the numbers to support that.‘

“

And just like that, a month of meticulous work turned into a prop. The data wasn’t a tool for discovery anymore; it was an inconvenient script that needed a rewrite. This isn’t a rare event. It’s the silent, grinding engine of most organizations. We pretend to be data-driven, but what we really are is data-decorated. We use data like a drunk uses a lamppost: for support, not for illumination.

This behavior erodes institutional trust faster than anything I’ve ever seen. It teaches the most curious people on your team that their curiosity is worthless. It trains them that the real skill isn’t analysis, but the political theater of justifying a foregone conclusion. The outcome is a culture of deep, corrosive cynicism, where everyone knows the game is rigged but has to keep playing their part.

Owen F.T.’s Ceramic Bowl

I think about Owen F.T., a third-shift baker I met once who was obsessed with the science of his sourdough. He had spreadsheets that would make a CFO weep. He tracked hydration levels to the gram, ambient humidity, and fermentation times in 15-minute increments. He believed every loaf was a product of pure, hard data. One day, I watched him meticulously weigh his flour and water, check his charts, and then pour the entire mixture into a heavy, antique ceramic bowl with a noticeable crack. I asked him why he used it when all his tests showed a modern stainless steel bowl provided a more consistent temperature and a 5% better rise. He looked at the bowl, a distant expression on his face.

‘My grandfather used this one,‘ he said. ‘The thermal mass of the ceramic creates a better crumb structure.’ He said it with the confidence of a man stating a fact. He had never actually tested that theory. He just felt it was better, and his technical-sounding explanation was the data-decoration he needed to make his feeling legitimate.

“

It reminds me of that social dance when you’re trying to leave a party. You decided 25 minutes ago that you were ready to go. You’re done. But you can’t just announce it. That would be disruptive. So you linger, waiting for the perfect, justifiable exit. You’re not gathering data on whether you should stay; you’re hunting for a socially acceptable excuse to execute a decision you’ve already made. ‘Oh, look at the time!’ is the corporate equivalent of ‘Let’s find the numbers to support that.’ It’s a performance of rationality.

A Solution in Search of a Problem

I’m not immune. Years ago, I championed a ridiculously complex software feature. I was convinced it would be a game-changer. My gut screamed it. The initial data was… ambiguous. Some metrics were up, some were down. So I did what we all do. I cherry-picked. I focused on a single, obscure engagement metric that had jumped by an impressive 235% since we launched the beta. I built entire presentations around this one number. I called it the ‘leading indicator of future success.’ We ignored the red flashing lights from the customer support tickets and the plummeting daily active user counts. We decorated our gut feeling with that one beautiful number. The project was eventually killed, but not before it burned through a budget of $25,575 and an immense amount of team morale.

That failure taught me something essential about the nature of evidence. True, objective evidence doesn’t care about your narrative. It just is. It’s an unblinking eye. A lot of security is about this principle. You don’t install a camera hoping to see one specific thing; you install it to see what is actually there. We once had a recurring inventory discrepancy in a warehouse, a loss of about 145 units a month. Everyone had a theory. Theft. Supplier error. Shipping damage. Each department had its own narrative and was quietly looking for data to support it. Instead of arguing, we installed a single, high-resolution poe camera over the receiving station. It wasn’t there to prove anyone right. It was there to see. After 25 hours of footage, we found the ‘thief’: a faulty conveyor belt was dropping one specific product into a recycling bin every 45 minutes, hidden from view. The data from the camera didn’t have a gut feeling. It didn’t have a political agenda. It just showed what was happening.

And yet, I know that if the VP from that meeting had been in charge, he might have watched the footage and said, ‘Yes, but my gut tells me the real problem is employee morale. Let’s find footage of people looking unhappy.’ That’s the maddening part. The most objective data in the world is useless in the hands of someone who has already decided what it is supposed to mean. Data is fundamentally useless for making big, future-facing decisions.

This is why the process matters more than the numbers themselves. A culture that values challenging assumptions, that celebrates the person who proves the popular hypothesis wrong, is a culture that can actually use data. A culture that only rewards the confirmation of existing beliefs is just a story-telling club with spreadsheets. Every business needs rigorous, impartial data to avoid just gambling with investor money.