Activity stages versus evidence stages
Most pipelines are a list of things the salesperson did: called, demoed, sent a proposal. That’s a to-do list with a currency symbol on it. The forecast built on it measures effort and calls it intent, which is why it’s wrong in the same direction every month.
An evidence-based pipeline flips the question. Not “what have we done?” but “what has the buyer done that tells us this is real?” The stage names can stay the same. The entry criteria change completely.
Stage by stage
| Stage | Activity version (don’t) | Evidence version (do) |
|---|---|---|
| Lead | We got a name | Source logged; fits the profile we wrote down |
| Qualified | We had a call | Buyer named the problem in their words; budget range and decision-maker confirmed |
| Demo | Demo done | Decision-maker attended; buyer described what changes for them if this works |
| Proposal | Proposal sent | Buyer agreed a decision date; next step owned and dated on both sides |
| Won | Verbal yes | Signed; handover checklist complete; delivery owner named |
| Retained | — | 90-day review booked; referral asked; expansion opportunity logged |
The rule that makes it work: deals move back
Evidence expires. A decision date that passed with no decision is no longer evidence. If deals can only move forward, every stage becomes a graveyard of deals that used to be real, and the forecast inherits all of them. Build the pipeline so a deal drops back a stage when its evidence lapses, and review the ones that dropped every Monday. That one rule does more for forecast accuracy than any reporting layer on top.
What it does to the numbers
The pipeline gets smaller and the forecast gets right. Demo-to-close goes up, not because anyone sold better but because deals that were never going to close stop being counted as demos. In Case 02 the equivalent — pitch-to-placement, their lead to close — went from 16% to 30%+, and the first month of it was purely definitional.