The Cathedral of Cherry-Picks: Why Data-Driven is a Lie

The Cathedral of Cherry-Picks: Why Data-Driven is a Lie

The Relic on the Wall

Greg is leaning over the mahogany table, his knuckles white, pressing a clicker so hard I think the plastic casing might actually snap. The hum of the projector is the only thing filling the silence of the 17th floor. On the wall, a single bar chart glows with the intensity of a holy relic. The bar representing our Q3 engagement is towering, bathed in a triumphant shade of emerald green, while the competitors’ bars are rendered in a shameful, muted grey. ‘As you can see from the data,’ Greg says, his voice vibrating with a certainty that feels almost violent, ‘this is the clear path forward. The numbers don’t lie.’

I’m sitting three seats down, nursing a lukewarm coffee that tastes like burnt paper, watching a dust mote drift through the laser beam. I know about the 47 other slides Greg left in the ‘Draft’ folder. I know about the 117-page raw export from the analytics team that showed a 37% drop in retention among our primary demographic. But those numbers aren’t here. They haven’t been invited to the party. Greg isn’t being data-driven; he’s being data-dressed. He’s found a costume that fits the decision he made three weeks ago while staring at a Scotch on the rocks in a hotel bar.

Data Theater: The Selected vs. The Unseen

Emerald Bar

Chosen Metric (Positive Spin)

vs.

37% Drop

Retention Reality (Hidden)

Lying with Directions

There is a peculiar kind of nausea that comes from watching someone use a precision instrument to perform an act of total fiction. It reminds me of last Tuesday when I was walking near the library and a tourist, looking hopelessly lost with a paper map, asked me the way to the old pier. I pointed him left, toward the 47th street bypass, even though I knew the 17th street cut-through was half a mile shorter. Why did I do it? Because 47th street has those beautiful weeping willows and the better view of the skyline. I wanted him to have the experience I thought he should have, rather than the one he actually asked for. I lied with directions. Greg is lying with decimals.

We live in an era where ‘data-driven’ has become a corporate shibboleth, a phrase we chant to ward off the demons of accountability.

– The Performance

We live in an era where ‘data-driven’ has become a corporate shibboleth, a phrase we chant to ward off the demons of accountability. If a project fails but we had a spreadsheet to back it up, it wasn’t a mistake; it was a ‘statistical anomaly.’ But if we follow a gut feeling and fail, it’s a career-ending lapse in judgment. This has created a culture of data theater. We aren’t looking for the truth; we are looking for the most compelling story we can build out of the fragments of reality that don’t hurt our feelings.

The Prison Librarian’s Architecture

My friend Rio J.-P. understands this better than most. Rio is a prison librarian, a man who spends his days navigating a 17,000-volume collection of lives that were, in one way or another, the result of a very specific kind of data processing. Rio tells me that every inmate he works with has a perfectly logical, data-assisted narrative for why they are where they are. They don’t ignore the facts of their crimes; they just curate the context. They highlight the 7 good deeds they did that year and minimize the 17 seconds that changed everything.

‘People think the library is about facts,’ Rio told me once while shelving a tattered copy of a legal thriller. ‘But it’s really about architecture. How do you stack the things you know to keep the ceiling from falling in on your sense of self?’

In the boardroom, the ceiling is held up by Greg’s emerald-green bar chart. No one asks about the Y-axis. No one asks why the data starts on July 7th instead of June 1st. To ask would be to break the spell. We are all participating in a collective hallucination where the numbers are the masters and we are merely the humble servants of logic. But the logic is hollow. It is a ‘data-assisted’ performance.

Support Versus Illumination

Real data is messy. It is inconvenient. It is the 237 rows of a CSV file that contradict your favorite hypothesis. It is the realization that your $7,777 ad campaign didn’t actually drive growth, but merely happened at the same time a celebrity accidentally mentioned your brand. When we say we are data-driven, we usually mean we are using data like a drunk uses a lamppost: for support rather than illumination.

“Data is the makeup we put on the face of a guess.”

– The Core Deception

This erosion of rigor creates catastrophic blind spots. When we cherry-pick the metrics that make us look good, we lose the ability to see the iceberg until we are already underwater. I’ve seen companies ignore a 57% increase in customer complaints because their ‘Net Promoter Score’-a metric specifically designed to be easily manipulated-stayed flat. They chose the comfort of the metric over the reality of the resentment.

It’s the same impulse that leads people to ignore the fine print on a nutrition label or the complex chemistry of what they put in their bodies. We want a simple answer. We want to be told that the oil we’re using to fry our eggs is ‘healthy’ because it has one specific antioxidant, ignoring the 7 other ways it might be inflammatory. True clarity requires seeing the whole picture, even the parts that make you want to change your mind. For instance, when you’re navigating the confusing landscape of dietary choices, you need the kind of raw, unfiltered transparency found at avocado oil for cooking where the charts don’t hide the inconvenient truths about smoke points and fatty acid profiles just to sell a specific bottle.

The Weight of Truth

Rio J.-P. sees the same thing in the prison library. Inmates will come in looking for books on ‘The Law of Attraction’ or specific case law that they believe will prove their innocence on a technicality. They have 17 different reasons why the evidence was misinterpreted. They are data-driven in their pursuit of a specific outcome. They ignore the 377 pages of trial transcripts that tell a different story because those pages don’t serve the narrative of the ‘Safe Room’ they’ve built for their egos.

We do this because the truth is heavy. To be truly data-driven would mean being willing to be wrong 77% of the time. It would mean walking into a meeting and saying, ‘The data suggests that my favorite project is a waste of money and we should kill it immediately.’ Have you ever heard anyone say that in a meeting? I haven’t. In 17 years of corporate life, I’ve seen 47 people get promoted for spinning a failure into a ‘learning opportunity,’ but I’ve seen zero people get promoted for admitting the data proved them wrong before they spent the budget.

The Dashboards of Deception

7

KPIs Discussed

70

KPIs Appendix

This isn’t just about business; it’s about our relationship with reality. We treat data like a buffet where we only take the shrimp and ignore the wilted kale. We create ‘Dashboards of Deception’ that monitor 77 different KPIs, but we only talk about the 7 that are flashing green. The other 70 are relegated to a ‘technical appendix’ that no one reads because it doesn’t fit the ‘High-Level Overview’ that the executive team wants to see.

The Architect of Ignorance

I think back to that tourist I misdirected. I felt a twinge of guilt as he walked away, but I also felt a strange sense of power. I had decided his path. I had used my ‘expert’ knowledge of the city to manipulate his experience. Greg feels that same power. He isn’t just presenting data; he’s presiding over our perception of the truth. He is the architect of our collective ignorance.

If we want to fix this, we have to stop asking ‘What does the data say?’ and start asking ‘What is the data hiding?’ We have to become comfortable with the 17% of the results that don’t make sense. We have to look for the outliers, the anomalies, and the awkward truths. We have to stop using numbers as props and start using them as scalpels.

Rio J.-P. once told me that the most honest book in his library is a dictionary. ‘It doesn’t try to tell a story,’ he said. ‘It just gives you the pieces. It’s up to you not to lie about what they mean when you put them together.’

As Greg clicks to the final slide-a massive ‘Q&A’ text over a photo of a mountain climber reaching the summit-I realize that no one is going to ask a question. We are all too tired, or too invested in the emerald-green lie, to point out that the mountain in photo is actually in the wrong hemisphere for the metaphor he’s trying to use. We just nod. We are ‘data-driven.’ We are headed for the cliff, but at least we have a 117-slide deck that says the view on the way down is spectacular.

Next time, I’ll tell him about the shortcut through the alley, even if the walls are covered in graffiti and the smell of old trash lingers in the air. It’s not a pretty walk, but it’s the truth of the city. And in a world of curated emerald-green bars, a little bit of ugly truth is the only thing that might actually save us from our own well-documented delusions.

I’ll probably see that tourist again, or someone like him, wandering near the library. Next time, I’ll tell him about the shortcut through the alley, even if the walls are covered in graffiti and the smell of old trash lingers in the air. It’s not a pretty walk, but it’s the truth of the city. And in a world of curated emerald-green bars, a little bit of ugly truth is the only thing that might actually save us from our own well-documented delusions.

The logic is hollow, but the design is solid.