Most revenue forecasting leans on the same handful of metrics: pipeline coverage, stage progression, activity volume. They're easy to report on a dashboard. They're also, on their own, surprisingly weak predictors of whether a quarter actually lands.
Why the obvious metrics under-perform
Coverage and stage progression describe the shape of the pipeline, not the quality of the decisions being made inside it. A deal can sit in "late stage" because it's genuinely close, or because no one has had the honest conversation that would move it to closed-lost. The dashboard can't tell the difference; the number looks identical either way.
Signals that tend to matter more
- Multi-threading depth. How many people at the buying organisation have engaged, not just how many meetings happened. Single-threaded late-stage deals are disproportionately the ones that stall.
- Change in deal velocity, not just stage. A deal moving stages quickly early and then stalling is a different (and more predictable) risk pattern than one that's been consistently slow.
- Specificity of the next step. Vague next steps ("follow up next week") correlate with weaker forecast accuracy than concrete, mutually agreed actions with a date attached.
- Manager conviction, independently gathered. Not the rep's self-rated confidence, but what a manager believes after their own review of the deal, gathered without the rep's number anchoring it first.
A practical implication
None of these require new technology to start tracking. They mostly require deciding to look at a different layer of the data that usually already exists in the CRM or call records. The harder part isn't measurement; it's resisting the pull toward the metrics that are easiest to put on a slide, in favour of the ones that are actually diagnostic.
That trade-off, legible versus predictive, shows up constantly in revenue operations, and it's rarely resolved in favour of the metric that's harder to explain in a QBR.