A Single Source of Truth Is a Mirage
What you really need is someone who knows which number to trust.
Everyone talks about building a “single source of truth.”
It sounds responsible. Clean. Smart.
If all your metrics live in one governed place, you eliminate confusion. Everyone aligns. Every number makes sense.
But in practice? That’s not what happens.
Because the truth isn’t in a tool.
It’s in people.
You don’t need a single source of truth.
You need someone who knows where the truth lives today.
And that’s rarely just one place.
Warehouses Lag. Businesses Move Fast.
The data warehouse is the go-to solution for chaos.
Centralize everything.
Standardize metrics.
Model the logic.
Visualize in a dashboard.
Ship it.
But by the time it’s “done,” the business has already changed.
A new pricing plan launches.
Sales changes comp.
Marketing tweaks lead scoring.
Support changes how they log churn.
Your model is stale. Your metric is broken. And nobody agrees on what’s “right.”
Even if the data is clean, it’s often wrong in ways that matter. Should that customer count as churned? Is that upsell real or just a billing quirk? Truth requires context—and that doesn’t live in your warehouse.
Startups Run on Partial Truths
At any given time, key metrics live in different systems:
ARR in Salesforce
Usage in your product database
Churn in Stripe
Customer health in someone’s head
You will never unify everything perfectly. That’s not a failure. That’s reality.
The real skill is knowing which system to trust for this specific question. That’s not a function of tools. It’s judgment. Business context. Pattern recognition.
And that lives in people.
Great operators aren’t trying to clean up every system. They’re navigating the mess. They know where the gotchas are. Which metrics are good enough. Which ones to ignore. They know when to double-click, and when to move on.
It Starts and Ends with the Human
That person? They’re your real source of truth.
They’ve lived through every redefinition of ARR. They know why gross margin changed last quarter. They can tell you when a number is technically correct but functionally useless.
They don’t just pull numbers.
They understand the story behind them.
At Intercom, we changed how we defined ARR multiple times. But the person we trusted to explain ARR didn’t change. That person was our source of truth—even when the data wasn’t.
What to Actually Optimize For
This should change how you hire and how you operate.
Hire analysts and operators with judgment.
People who ask better questions.
Who notice when a number doesn’t smell right.
Who can explain what matters—and what doesn’t.
Reward those people. Build around them.
Because the job isn’t to build a perfect data stack.
It’s to make good decisions.
The Mirage of Perfection
A single source of truth sounds like a data problem.
But it’s a people problem.
The idea that you’ll model your way to one perfect, always-accurate metric?
It’s a mirage.
Stop chasing it.
Next time someone says, “We need to build a single source of truth” — and kicks off a 6-month project to do so — send them this instead.
Then go find the person who actually knows what’s going on.
They’re your real source of truth.