Data Work Is Political

The people doing the least data work have the most opinions about it

“Just get me the number.” That’s the lie.

Because the second you do, they ask where it came from. Then they want to know if it includes refunds. Then someone else says the old number was higher. Then the VP says to “take it offline.”

What they’re really saying is: We weren’t prepared for the truth to have a source.

The work of aligning definitions, untangling lineage, and documenting ownership isn’t boring. It’s political. You’re not formatting tables. You’re drawing battle lines.

Why “source of truth” is a threat

Every metric has a power base. Churn rate. Active users. Gross margin.

Who defines it gets to decide what’s “real.” Who controls it decides who gets blamed.

So when a data team walks in and says, “We’ve cleaned up the lineage and created shared semantics,” they’ve accidentally stepped into a turf war. Because what sounds like governance to you sounds like exposure to them.

They don’t want one churn rate. They want five—so they can pick the one that best suits the narrative.

This is why your data model never gets approved. This is why your lineage diagram collects dust. This is why your Confluence gets ghosted.

The architecture is a mirror of the politics

Every undocumented pipeline is a backchannel. Every “temporary” Tableau extract is a shadow government. Every stale dashboard is someone’s past promotion still lingering in production.

And you want to rewrite it? Good luck.

You think you’re cleaning things up. But you’re erasing informal alliances. You’re dissolving the ambiguity that gave them cover. You’re redistributing power—without asking.

No wonder they fight you.

Stop pretending this is neutral

If your job is to clarify truth, you’re going to make people uncomfortable. Because clarity is not just technical. It’s existential.

It forces:

• Marketing to own the bounce rate
• Finance to justify their margin games
• Product to admit the feature didn’t move the needle
• Execs to stop hedging with “directionally correct”

This is why data work gets deprioritized until it’s urgent. Because truth is only convenient when it matches the story already sold.

How to survive it

Treat lineage work like stakeholder mapping. Treat definitions like contracts. Treat ownership like succession planning.

You need influence, not just access. Buy-in, not just documentation. And a spine, because the closer you get to clarity, the more resistance you’ll face.

Data isn’t the new oil. It’s the new politics. And your ability to navigate that determines whether your team ships trusted intelligence—or just prettier dashboards for the next reorg.