Sales Stage Exit Criteria Template: How to Write Rules Reps Actually Follow

What makes an exit criterion good or useless?

A good criterion describes something the buyer did that cost them effort. A useless one describes something the rep did. "Proposal sent" is useless because a rep can send a proposal to someone who never asked for it. "Buyer scheduled the security review" is good because the buyer spent internal capital to make it happen.

That single test fixes most stage definitions. Read each criterion and ask who had to act. When the answer is your own team, the criterion is measuring effort and calling it progress.

What does the exit criteria template look like?

One row per stage with the buyer evidence, the artifact that proves it, and the CRM field that records it. Keep it to two or three criteria per stage. One criterion is trivially satisfiable. Five is paperwork, and reps respond to paperwork by parking deals in an early stage until the week they close. That behavior looks like hygiene failure and is actually a design failure, and it wrecks stage conversion data.

Why does the artifact column matter more than the criterion?

Because a criterion without an artifact is an opinion and an opinion cannot be inspected. Two reps will both mark "buyer confirmed a problem" as satisfied, and only one of them will have notes containing a sentence the buyer actually said.

The artifact requirement changes the conversation in a review. Instead of asking whether a deal is really at stage four, a manager asks to see the redline document. The question is answerable in ten seconds and it is not a judgment call, so it does not turn into an argument about rep confidence.

The artifact also gives you something to audit at scale. Sampling twenty deals per quarter and checking whether the named artifact exists tells you the real state of stage discipline, which no dashboard will.

How do you enforce criteria without a validation rule war?

Validate facts, inspect judgment. A required field linking the economic buyer contact record before stage four is a fact and a reasonable hard stop. A required text box asking the rep to explain buyer intent produces "confirmed" typed forty times and a team that resents the process.

Three enforcement layers work together:

- Hard validation on the small number of fields that are objectively checkable. - A weekly report of deals that skipped a stage or moved two stages in one update. - A quarterly artifact audit on a random sample.

The skipped-stage report is the highest-value item on that list. Deals that jump from qualified to negotiation in one save are usually deals a rep worked outside the process, and those are worth understanding whether they close or not.

What does good stage design do for the forecast?

It makes stage conversion rates comparable, which is the entire reason stages exist. When "stage four" means the same thing across reps and across quarters, the historical conversion rate from stage four to closed won becomes a usable number. When it does not, weighted pipeline math is applied to categories that have no consistent meaning, and the output looks precise while being arbitrary.

Exit criteria also make aging interpretable. ORM groups opportunities with a machine learning model and predicts a close curve for each group, and those curves run from one to eighty weeks, with most of the expectation landing before week twelve and very few groups carrying expectation past fifty-two weeks. A deal sitting well past its group's curve is a real anomaly only when the stage it sits in means something.

Set the definition of activity alongside the criteria. ORM counts a change in stage, close date, or amount as meaningful activity, and applies a twelve-month rule for most customers. Expect the results to be uncomfortable at first: across ORM's customer base, more than 10% of pipeline has gone twelve months without being touched.

What should you do about the data quality objection?

Ignore it, because it is almost always wrong. Every team believes their data is uniquely bad and that it is the reason they cannot run the business the way they want. Everyone has bad data. What matters is whether it is consistent, because consistent data supports accurate prediction even when it is messy.

That is the strongest argument for stage exit criteria. Criteria do not make the data clean. They make it consistent, which is the property that a forecast model can work with. A team where every rep applies the same imperfect rule produces better forecast accuracy than a team where half the reps apply a perfect rule and the other half apply nothing.

Revise the criteria once a year, or after a structural change such as a new segment or a pricing model shift. Revising more often is tempting and it makes conversion rates incomparable across periods, which costs more than the improvement is worth. Fix the enforcement instead. That is usually where the problem actually is, and cleaner stages are what make pipeline coverage mean anything at the portfolio level.