The stage changed after the demo.
That sounds reasonable until someone asks what changed for the buyer.
Did they bring the decision-maker into the next meeting? Did they confirm the problem in their own numbers? Did they agree on a date, a buying process, or even a next conversation?
No. The demo went well. People nodded. One person said the product looked interesting. The rep moved the opportunity from Discovery to Proposal because leaving it where it was felt pessimistic.
Nothing happened. The CRM simply became more confident.
This is how a pipeline turns into a mood board. Every label looks precise, yet two reps can use the same stage to mean completely different things. The manager sees a proposal-stage opportunity. One rep means that pricing has been reviewed with the economic buyer. Another means that a PDF was sent into the building.
Those are not the same deal.
A stage is a claim
A sales stage makes a claim about what is true right now. "Discovery complete" should mean the team knows enough to decide whether the opportunity deserves more time. "Proposal" should mean the buyer understands the proposed change and is participating in a decision, not merely that the seller generated a document.
Salesforce describes exit criteria as the specific, measurable requirements a prospect must meet before moving forward. HubSpot makes a similar point: each step needs defined actions and exit criteria if a team wants to measure whether its process works.
The word measurable can make this sound like a RevOps workshop involving a large spreadsheet. It does not have to be.
What did the buyer say, do, share, schedule, or approve that was not true before?
That question shifts the stage away from seller activity and towards buyer evidence.
"We ran the demo" is an activity.
"The operations lead brought finance into a 30-minute pricing review next Tuesday" is evidence.
"We sent the proposal" is an activity.
"The buyer reviewed the proposal, raised two commercial issues, and agreed to resolve them with procurement by Friday" is evidence.
Activities matter. They just do not prove progress on their own.
Build stages around evidence, not theatre
Most pipelines have some version of qualification, discovery, evaluation, proposal, negotiation, and close. The labels matter less than the agreement behind them.
| Stage | Weak reason to advance | Evidence that earns the next stage |
|---|---|---|
| Qualified | They took a call | A relevant problem is confirmed, the account fits, and both sides agree there is something worth examining |
| Discovery | We asked our questions | The current process, consequence, stakeholders, urgency, and decision path are understood well enough to test a solution |
| Evaluation | They liked the demo | The buyer has connected the solution to its use case and involved the people needed to judge it |
| Proposal | We sent pricing | Scope and commercial terms have been reviewed with someone who can influence the decision, with a dated next step |
| Negotiation | Legal has the contract | The remaining issues are named, each has an owner, and both sides are working towards an agreed decision date |
| Commit | The rep feels good | The buyer has confirmed the approval path, final action, and intended signature date |
Do not copy that table into the CRM unchanged. A two-call agency sale and a nine-month enterprise software purchase should not share identical gates.
Instead, take one recently won deal and one that stalled. Reconstruct what actually happened. Which buyer actions separated real movement from polite participation? Where did the team know less than the stage implied? The answers will give you a first draft that belongs to your sales motion.
Use four pieces of evidence
A workable exit criterion usually contains four things.
1. A buyer action
The buyer commits time, introduces a stakeholder, shares data, tests the product, reviews a proposal, or starts an internal process.
This protects the pipeline from meetings that were pleasant but inert. Interest can be genuine without being strong enough to support a purchase.
2. Evidence in the buyer's words
The CRM should hold more than a checked field called "pain confirmed." What is happening? Who feels it? What does it cost or prevent? Why has the team not fixed it already?
The exact answer matters because a generic summary can make a weak problem sound qualified. "Reporting is manual" is vague. "The regional managers rebuild the forecast every Thursday because site-visit notes arrive after the call" gives the team something it can test and solve.
3. A mutual next commitment
"Follow up next week" is not a next step. It is a reminder for the seller.
A real commitment names what happens, who is involved, and when. The next meeting may still be cancelled. That is fine. The point is that both sides have agreed on the work required to keep evaluating.
4. A reason not to advance
Every stage needs a way out as well as a way forward.
No problem, no owner, no access to the right people, or no plausible decision path may mean the deal should remain where it is, move to nurture, or close. Keeping it alive because the logo looks attractive does not protect pipeline. It hides the answer.
Do not turn the CRM into airport security
Exit criteria can go wrong in the other direction. A team begins with a sensible desire for consistency and ends with 19 mandatory fields before a rep can move an opportunity.
That does not produce better evidence. It produces creative data entry.
Keep the gate small enough to use during a normal week. Ask for the few facts that change how the team sells or forecasts. Where possible, capture those facts from the meeting itself rather than asking the rep to reconstruct the conversation hours later.
The manager should also be able to inspect the evidence. "Champion identified" is less useful than a short note explaining who the person is, what they want, and what they have done. A checkbox records the conclusion. The note lets somebody challenge it.
This becomes especially important when a rep leaves or a deal changes hands. The new owner should not inherit a row of green fields and a mystery. They need the buyer's problem, commitments, objections, stakeholders, and decision path.
Run the 10-minute pipeline test
Pick five opportunities that moved stages in the past fortnight. For each one, ask the rep:
- Change: What became true for the buyer?
- Record: Where is that evidence recorded?
- Commitment: What did the buyer agree to do next?
- Exit: What would make us move the deal back or close it?
This is not a trap. If a capable rep cannot answer consistently, the stage definition is probably doing too little work.
Listen for answers built entirely around seller activity: we called, presented, followed up, sent, or checked in. Then ask for the buyer half of the sentence.
We presented, and the department head asked us to model the rollout for 12 sites.
We sent pricing, and procurement returned its required terms with a review booked for Thursday.
We followed up, and nothing changed.
That last answer is useful too. It stops a routine email from masquerading as progress.
Make the forecast inherit the evidence
Forecasting problems often look mathematical. Sometimes they begin much earlier, with language.
If "Proposal" can mean anything from "the rep is preparing a quote" to "the CFO has approved the commercial structure," a stage-weighted forecast cannot rescue the ambiguity. The percentage is being applied to several different realities.
Clear exit criteria narrow those realities. They make pipeline reviews less theatrical because the conversation can move from "How do you feel about it?" to "What evidence do we have, and what is missing?"
They also make coaching more useful. A manager can see whether the rep is struggling to uncover consequence, reach the decision group, secure a mutual next step, or recognise that the deal has stopped. "Be more confident" is not a coaching plan. "Before we build the proposal, get the finance lead into the evaluation and confirm how the decision will be funded" is.
The CRM stage still matters. It is the headline.
The evidence underneath it is the article.
Quick answers
What are sales stage exit criteria?
They are the observable requirements a deal must meet before it moves to the next pipeline stage. Good criteria describe buyer evidence, not only tasks the seller completed.
Should every company use the same sales stages?
No. The stages should match how customers actually buy from that company. A useful starting structure can be adapted, but the evidence, people, and commitments will differ by sales motion.
How many exit criteria should a stage have?
Use the smallest set that gives the team a truthful picture and changes what it does next. If reps need to complete a long form to advance a deal, the process is likely to create compliance rather than clarity.
What if the buyer will not provide the evidence?
That is evidence too. The deal may be too early, missing the right stakeholder, or not important enough to the buyer. The answer is not automatically to close it, but it should not be promoted on hope.
Research note: This guide builds on public pipeline and sales-process guidance from Salesforce, HubSpot, Gong, and Close. LogicNotes added the buyer-evidence model, stage examples, four-part criteria, and pipeline-review test.
Keep the evidence behind the stage.
LogicNotes turns sales conversations into structured CRM updates, so managers can see why a deal moved without asking reps to rebuild the meeting from memory.
See LogicNotes for teams