The Number That Should Stop You Cold
Jordan Crawford ran 1.6 million GTM datasets through a rigorous quality filter and 12 survived. Not 12 percent. Twelve datasets — total.
If your outbound is underperforming, the instinct is to fix the copy, swap the sequence, hire a better SDR, or test a new channel. Almost nobody's first move is to question whether the underlying data ever met a minimum viable threshold. That's the problem.
Crawford's finding isn't an outlier — it's a stress test that most revenue teams have never run on their own stack. The implication is direct: you may be optimizing a machine that's running on bad fuel.
What Separates Signal from Noise
The 12 datasets that survived share specific structural properties. Crawford names the criteria, which means you can actually audit your current providers against them. The categories of failure in the other 1.6 million aren't random — they cluster around stale firmographic data, intent signals with no clear behavioral anchor, and contact records that technically exist but don't reflect real buying context.
The practical test is harder than it sounds. Most RevOps teams inherit a data stack rather than choose one. Providers get grandfathered in, list vendors are renewed out of habit, and intent data subscriptions persist because nobody wants to run the experiment of turning them off. The result is outbound volume built on a foundation nobody has pressure-tested.
The Audit You Should Run This Week
Before your next campaign build or sequence refresh, apply a blunt filter to every data source in your stack:
- Recency: When was this record last verified? Intent signals decay fast — weeks, not quarters.
- Behavioral anchor: Does the signal tie to a specific, observable action, or is it an inferred score from a black-box model?
- Match rate: What percentage of records actually reach a real inbox or connect to a real decision-maker? If your provider can't tell you, that's your answer.
- Source transparency: Can the vendor trace where the data originated, or is it aggregated from aggregators?
If a data source can't answer those four questions cleanly, it shouldn't be informing your ICP targeting — regardless of what the contract cost.
The uncomfortable takeaway: most pipeline problems diagnosed as messaging failures are actually data quality failures. Fix the foundation before you spend another dollar on outbound execution.