A stage should describe a buyer situation, not how long a record has existed.
Write observable criteria
For every stage, name the event, required evidence, owner, and next action. Avoid criteria such as “feels promising.”
Handle reversals
Allow an opportunity to move backward when evidence changes. Keep the reason and date so forecasts do not hide the change.
Review quality
Flag records lacking exit evidence or a current next action. Sales leadership decides whether the stage model needs revision.
Write verifiable evidence
For discovery, evidence might be a dated conversation naming problem, stakeholders, and an agreed next step. “Had a good call” cannot be audited. Keep criteria short and observable.
Measure quality
Sample records for exit evidence, age, owner, and next action. Reversals and long pauses reveal weak criteria. If the model adds work without improving forecast conversations, propose a smaller evidence set for leadership approval. Test crm stage exit criteria on several opportunities that recently changed stage. For each one, capture the required evidence, decision owner, and next action, then note whether the record actually met the gate.
When a stage has frequent exceptions, distinguish a missing field from a legitimate alternate path. Update the rule only after the pipeline owner reviews examples and confirms that reports will retain their meaning.
