Your data warehouse is lying to you about pipeline

A former dbt Labs engineer explains the attribution gaps hiding in plain sight.

The data warehouse is supposed to be the source of truth. But for most companies, it's hiding critical attribution problems.

We spoke with a data engineer who spent two years building data models at a high-growth SaaS company. She discovered that the standard pipeline attribution model was systematically undercounting certain revenue sources by 30%.

The problem wasn't the warehouse. It was the model.

Attribution is hard. Multi-touch attribution is harder. And most companies aren't investing enough rigor in how they're connecting revenue to the activities that actually generated it.

"Everyone trusts the warehouse because it looks official," she says. "But garbage in, garbage out."

The fix isn't complicated, but it is tedious. It requires detailed documentation of how different touchpoints map to pipeline stages, regular audits, and a willingness to challenge your own assumptions.

For revenue leaders, the lesson is simple: audit your attribution model. You're probably wrong about where your revenue is coming from.