How to run a sales pipeline review meeting

A pipeline review should take forty minutes and cover six deals, not ninety minutes and forty. The change that makes that possible is the order you take them in: start from the deals carrying most of the uncertainty rather than walking the list by stage or by size. In a typical quarter that is three or four names, and everything else is rounding.

Below: why stage order fails, which deals actually deserve the time, an agenda you can use on Monday, and how often to hold it.

Why walking the list fails

The default review walks every open opportunity, usually newest or largest first, asking for an update on each. Two things go wrong. The meeting runs out of time somewhere in the middle of the list, so the last third gets three minutes. And attention is spread evenly across deals whose influence on the quarter is wildly unequal.

How unequal is a question with an arithmetic answer. A deal's contribution to the uncertainty of the quarter is p × (1 − p) × value² — probability times its inverse, times value squared. Two consequences fall out of that, and neither is obvious.

• Uncertainty peaks at 50%, not at the bottom. A deal at 90% or at 10% is close to decided. A large deal sitting at 50% is the least decided thing you own.
• Value is squared. A deal twice the size carries four times the doubt, which is why concentration is far sharper than coverage ratios suggest.

52% of the uncertainty in forty deals sits in three of them. Those three are the only ones whose movement changes the answer this quarter. The other thirty-seven are the meeting you have been having.

"Which deals are at risk" is the wrong filter, and it is the one most reviews use. It surfaces the shaky small ones while a large deal sitting at 60% looks healthy. The right filter is which deals move the number most if they go either way.

An agenda that fits in forty minutes

1. The number, as a range (3 minutes). Where the quarter lands, the odds against target, and how that has moved since last time. Not a point estimate: a range and a probability. If it moved, say which deals moved it.

2. The deals that drive the spread (20 minutes). Three or four names, in order of how much uncertainty they carry. For each: what has to be true for this to close, what would tell us it will not, and what the owner needs. This is the meeting. Everything above and below it is logistics.

3. Deals that changed category (8 minutes). Anything that moved stage, slipped a month, or had its probability changed since the last review. Changes carry information; static deals mostly do not.

4. Overdue deals (5 minutes). Anything still open in a month that has closed. Either it moves or it is lost, and leaving it open quietly inflates next month's pipeline.

5. Commitments (4 minutes). Each rep leaves with one or two deal names, not a spreadsheet. Write down what the team is committing to, because next month's review starts by comparing that to what happened.

How often to hold it

Weekly for the drivers, monthly for the whole thing. The three or four deals carrying the uncertainty deserve a fifteen-minute check every week, because that is where the quarter is decided. The full review, including the parts of the pipeline nobody has looked at, is a monthly job.

Quarterly is too slow to act on anything and too long to remember what was said. Daily pipeline meetings are activity theatre: deals do not change daily, and asking as though they do teaches the team to invent movement.

The thing that actually matters more than the interval is that each review closes the previous one out. If you committed to a number last month, this month's meeting starts with what closed against it. Without that loop the meeting is a status update, and status updates do not improve forecasts.

Three things to stop doing

• Stop cleaning the CRM before the meeting. The forecast needs the handful of deals that move the number, and you can get those from the reps in the room in fifteen minutes. Hygiene matters for coverage reporting; it does not need to be perfect for a forecast.
• Stop asking "will it close?" You get a yes. Ask what has to be true for it to close, and what would be the first sign it will not. Those have answers that can be checked next week.
• Stop using stage probability as the number. Ask the owner for their own figure. It is worse at first and much better within two quarters, because a person's judgement can be scored and a stage's cannot.

None of this needs software. A spreadsheet with deal, value, the owner's own odds and the close month will support the whole agenda, and the record it builds over a few months is worth more than any report your CRM produces.