How to run a pipeline review

What a pipeline review is for

A pipeline review has one job: find the deals that will not close the way the CRM says they will, early enough to do something about it. It is not a status update, it is not coaching, and it is not the forecast call. Teams that blur those three end up with a ninety-minute meeting where every rep reads their deals aloud, nobody challenges anything, and the number at the end is the same number the CRM already showed.

Separate the three meetings and give each a different input. The pipeline review inspects open opportunities against stage exit criteria and produces actions. The forecast call takes the output of that review and turns it into a commit. Coaching happens in the one-to-one, on the skill gap the review exposed, not in front of the team.

The review also protects your data. Every deal that gets inspected is a deal whose close date, amount and stage get corrected in the moment, which is the only pipeline hygiene routine that survives contact with a busy quarter. Automated field-completeness reports tell you what is missing; a review tells you what is wrong, which is a harder and more valuable thing to find.

Scope it by impact. In a book of 120 open opportunities, the top 20 by value plus everything with a close date inside the quarter is usually 30 to 40 deals, and that is an hour of real inspection. Everything else moves through hygiene rules and exception reports. If the review tries to cover all 120, it degrades into reading the pipeline report out loud, and the deals that needed challenging get four minutes each.

A weekly rhythm that holds

Fix the cadence before you fix the content. A rhythm that survives a bad quarter looks like this. Monday: reps update their own deals, close dates and next steps before a published cut-off. Tuesday: the manager runs the pipeline review with the team or in one-to-ones. Wednesday: the manager submits a commit. Thursday: the leadership forecast call. Nothing in the review gets re-litigated on Thursday, because Thursday only sees the output.

Publish the cut-off and mean it. If reps can edit deals after the manager has prepared, the manager arrives with a stale view and spends the meeting rediscovering changes. A frozen snapshot taken at the cut-off also gives you week-over-week movement, which is the single most useful artefact in the whole routine: what moved forward, what slipped, what appeared, what disappeared.

Length depends on the motion. A transactional team closing thirty deals a month needs 30 minutes on exceptions and a lot of automation. An enterprise team with eight deals per rep per quarter needs an hour, and should spend most of it on three or four deals in depth. The failure mode in both cases is the same: equal time per deal regardless of value or risk.

Monthly, add a wider pass. Look at aged pipeline, coverage by segment, and the deals nobody has mentioned in six weeks. Quarterly, review the stage definitions themselves. Our pipeline review checklist lays out the weekly agenda and the monthly additions, and the forecast call agenda covers the separate meeting that follows.

What to inspect each week

Inspect evidence, not opinion. For each deal on the list, the manager should be able to see, without asking: the last meaningful buyer activity and its date, the stage and how long the deal has been in it, the close date and whether it has moved, the named economic buyer, and the next scheduled step with a date and an attendee list. A deal missing the last two is not a late-stage deal regardless of what the stage field says.

Then run four filters over the whole book. Deals with a close date in the past. Deals that have not changed stage in more than one and a half times your average stage duration. Deals with no buyer-side activity in 21 days. And deals where the amount changed by more than 25% without a note. Those four filters find most of the rot, and they run as CRM reports, so the meeting starts with a list instead of a discovery exercise.

Watch the shape as well as the contents. A book where 70% of the value sits in the last stage and the close dates all land in the final two weeks of the quarter is not a healthy pipeline, it is a set of deals that have been pushed. Deal velocity by stage tells you whether the book is actually moving or just being re-dated.

Give reps the same view you use. If the manager inspects on fields the rep cannot see or does not understand, the review becomes an interrogation and the rep starts managing the fields instead of the deal. A shared deal review template keeps both sides working from one structure.

Questions that surface risk

Most review questions can be answered with a confident guess, which is why they surface nothing. Replace them with questions that require evidence a rep either has or does not have.

Instead of asking whether the buyer has budget, ask who signed the last purchase of this size and what that approval took. Instead of asking whether they like us, ask what the buyer said when the rep raised price, in the buyer's words. Instead of asking about the timeline, ask what happens inside the buyer's business on the day after they sign, and what is driving that date. Instead of asking about competition, ask what the buyer would have to tell the other vendor to end that conversation.

Two more are worth asking in every late-stage deal. What has to be true for this to close this quarter that is not true today? And who inside the buyer could stop this, and when did we last speak to them? A rep who cannot answer the second question has a champion, not a decision path.

Ask for the buyer's language, not the rep's summary. "They're excited" is a rep's word; "they said the current process costs them two days a month" is the buyer's. Qualification frameworks like MEDDPICC and BANT are useful here only as a checklist of what evidence to look for, not as a score to recite. A deal that scores well on paper and has no economic buyer on a call is still a deal at risk, and the review exists to say so out loud while there is still time in the quarter to fix it.

Stage exit criteria

A review is only as good as the stage definitions underneath it. If stage three means different things to two reps, no amount of inspection produces a comparable book. Write each opportunity stage as an exit test in the buyer's behaviour, not the seller's activity. "Demo delivered" is a seller action and can be true of a deal going nowhere. "Buyer has confirmed the problem, the approximate budget and who signs" is an exit test.

Keep it to five or six stages. Every extra stage is another judgement call and another place for deals to sit. For each one, write the exit criteria on a single line, name the required field or artefact that evidences it, and make that field required to advance. Common set: qualified opportunity with a named problem and buyer; validated need with a technical or functional confirmation; economic buyer engaged with budget confirmed; proposal delivered with a mutual close plan; and contract in legal.

Attach probabilities to stages only if you are prepared to calibrate them against your own closed-won history each quarter. An untested 60% at stage four is a number that flatters the pipeline. Many teams get more from dropping stage probability entirely and running forecast categories, where a manager explicitly places each deal in commit, best case or pipeline, and owns that call.

Enforce the criteria in the CRM rather than in the meeting. Required fields, validation rules and stage-gated approvals do the boring half of the job, which leaves the review free for judgement. Whatever you write down, publish it, and re-read it in the quarterly review before you argue about any individual deal.

Handling stalled deals

Stalled deals are the main reason pipeline coverage figures lie. A book carrying six months of deals nobody has touched looks like three times coverage and behaves like one and a half. The fix is a written rule that removes deals without requiring a fight in the meeting.

Set dead deal criteria in advance and apply them mechanically. A workable default: no buyer-side response in 30 days and no scheduled next step means the deal moves to a nurture status and leaves the open book; a close date pushed three times means the deal returns to the stage its evidence actually supports, not the stage it has been sitting in. Neither is a loss, and neither goes on the rep's record as one. That last point matters, because when closing a stalled deal feels like an admission of failure, reps keep them open, and the book fills with fiction.

Give reps one structured attempt before the rule fires. A short note to the economic buyer that names the silence and offers to close the file, sent once, resolves a surprising number of these either way. What you are buying is a decision, not a yes.

Then track what happens to the removed deals. If a third of them come back and close within two quarters, your window is too short. If almost none do, you can tighten it and get a cleaner book. Run that check once a quarter rather than adjusting the rule deal by deal, because a rule that bends on request is not a rule and the pipeline goes back to being a graveyard with a good coverage number on top.

Coverage math in the review

Coverage answers one question: is there enough open pipeline to hit the number at your historical win rate? The arithmetic is open pipeline value divided by the quota for the period. A team with $6M open against a $2M quarterly target has three times coverage.

The target ratio is not a rule of thumb, it is the inverse of your win rate with a margin. A team converting 25% of qualified pipeline needs 4x to break even on the maths and typically carries 4.5x to 5x to absorb slippage. A team converting 40% needs 2.5x. Copying somebody else's 3x is how teams end up structurally short or sitting on pipeline nobody works. Work out your own pipeline coverage ratio from your own win rate before you set a target.

Then make the number honest. Count only deals that have passed a real qualification gate, exclude anything with a close date outside the period, and exclude the stalled deals your rules have flagged. Many teams run both a gross and a qualified coverage figure and manage to the second. The gap between the two is itself a useful metric: a widening gap means the top of the funnel is producing volume that does not survive inspection.

Review coverage by segment and by rep as well as in total. A team can sit at a comfortable 3.4x overall while two reps carry 1.2x and will miss. Our pipeline coverage calculator does the arithmetic including the win-rate-derived target, and the ratio itself is worth reviewing monthly rather than weekly, since it moves too slowly for a weekly meeting to act on.

Feeding the forecast call

The review earns its place by making the forecast call short. If the forecast call re-inspects deals, the review failed. The handoff is three things: a commit number the manager owns, a list of the deals inside it with the evidence behind each, and a named upside set with what would have to happen for each to land.

Require managers to declare changes against last week rather than restating the whole book. What came into commit and why, what left and why, what slipped and to when. That framing takes ten minutes per manager and surfaces the pattern that matters: a manager whose commit is stable but whose contents churn every week does not have a forecast, they have a rolling guess.

Keep the record. Store each week's commit and compare it to actuals at quarter end, by manager. Forecast accuracy is a manager-level skill and it improves quickly once people know the history is kept. Without the record, the same optimistic manager is optimistic every quarter and nobody can say so with evidence.

Do not let the tooling replace the judgement. AI forecast signals are useful for flagging deals a manager has not questioned, but the commit should stay a human call with a name attached. If you are deciding whether this whole routine needs dedicated software, our guide to sales forecasting software covers what those products add once the process already exists, and the wider sales forecasting software category lists the tools we track.

When tooling helps

Most teams should run this in the CRM they already pay for. Pipedrive Growth at $39 per seat per month and HubSpot Sales Hub Professional at $90 per seat per month annually, with a one-time $1,500 onboarding fee, both include forecasting and the reports the four filters above need. Salesforce Sales Cloud lists Pro Suite at $100 and Core at $195 per user per month billed annually, with forecasting from Pro Suite upward. All checked September 2026.

The first thing worth buying is usually not a review tool, it is whatever gets data into the CRM without rep typing. Scratchpad publishes Solo at $19 per user per month billed annually and Team at $49, which is cheap enough to test against the hours reps spend on updates. Activity-capture products such as Nektar quote per user and do not publish figures; Nektar does list a data backfill add-on at $8 per user per month of history, as of September 2026.

Dedicated inspection and forecast platforms sit a level up and none publish list prices. Clari, now sold through Salesloft after the merger closed in December 2025, Aviso, and Gong all quote per seat plus a platform fee. For benchmarks rather than prices, Vendr reports a median Gong contract of $55,333 a year across 1,141 purchases and a median Aviso contract of $73,662 a year. Ask each vendor to price light viewer seats separately.

In a demo, ignore the executive dashboard and make them run a manager's Tuesday review live: pull the four filters, change a close date, record an action, and show last week's snapshot beside this week's. Do not buy while stage definitions are still unsettled, under roughly fifteen reps, or when the real problem is that nobody runs the meeting. If you are between two of these, the Clari vs BoostUp comparison covers the dedicated-forecast end, and the CRM software category covers the layer underneath.

Buyer checklist

• Separate the pipeline review, the forecast call and coaching into three meetings with different inputs.
• Publish a weekly cut-off after which reps cannot edit deals, and take a frozen snapshot at that moment.
• Build four CRM reports: past close dates, stage age over 1.5x average, no buyer activity in 21 days, amount changed over 25%.
• Write single-line exit criteria for each stage and make the evidencing field required to advance.
• Scope each review to the top deals by value plus everything closing in the period, not the whole book.
• Set dead deal criteria in writing (30 days silent with no next step) and apply them without a debate each week.
• Derive your coverage target from your own win rate, then track qualified coverage by rep and by segment.
• End every review with dated actions and a named owner, then check them at the start of the next one.
• Store each week's commit and compare it to actuals by manager at quarter end.

Related

All 2 Forecasting guides