Marketing Analytics Pulse

43%. That is a typical MQL-to-SAL acceptance rate across mid-market B2B funnels, and it is one of the most misread demand waterfall conversion rates in B2B marketing. Most teams see it and conclude that roughly half their leads are garbage. That conclusion is wrong, and acting on it is expensive.

The number does not measure lead quality. It measures agreement. When marketing counts MQLs from one dashboard and sales counts accepted leads from another, the gap between them is definitional, not factual.

The waterfall only works when all five conversion rates — lead to MQL, MQL to SAL, SAL to SQL, SQL to SQO, SQO to closed-won — are calculated from the same cohort definition in the same system. A 14% end-to-end win rate on marketing-sourced pipeline means something precise. The same number measured across inconsistent stage definitions means nothing.

The hardest design decision is cohort membership: when a lead re-engages after five months of silence, does it enter a new cohort or continue in the original one? The join logic in Snowflake or BigQuery has to encode that answer explicitly, because the model will produce a number either way and never tell you which rule it used.

Stage-by-stage visibility is what lets a CMO reverse-engineer from a revenue target to required pipeline volume to required marketing investment. That is credible math. Everything else is rounding.

Go find your MQL-to-SAL rate. Then ask whether marketing and sales calculated it from the same records.

Revenue attribution was the presenting problem. The real problem was sitting one layer underneath it.

We were brought into a B2B org where marketing and finance had stopped agreeing on pipeline numbers. Not recently. For over a year.

Every quarter, the same QBR scene: marketing's slide showed strong contribution, finance looked at Salesforce, and the two figures did not reconcile. Nobody called anyone a liar. They just stopped trusting the system.

The assumption going in was that the attribution model was misconfigured.

What we actually found was simpler and harder to fix. Marketo and HubSpot had each been used at different points in the company's history, leaving duplicate contact records and overlapping attribution windows that no one had ever formally retired.

Salesforce was assigning closed revenue to the last sales-logged activity. The marketing platform was claiming influence across every touchpoint. Neither methodology was wrong.

They were just never designed to talk to each other, and no data model existed to translate between them.

The rebuild started before we touched a single dashboard. We sat with marketing and finance together and documented what pipeline contribution actually meant for this business: which stages counted, which attribution window applied and why, what an MQL meant to both teams in the same room.

Once that was settled, opportunity data moved from Salesforce into BigQuery, modeled in dbt, and surfaced in Looker with logic both sides had signed off on.

The QBR conversation changed. Not because the numbers got bigger. Because both teams were finally reading from the same page.

These systems break when the tools get configured before the definitions get written. The infrastructure inherits the ambiguity and amplifies it until someone finally has to go back to the beginning.

This is what we build at Marqeu: the strategy and the implementation, together.