How to run a sales pipeline review in 30 minutes

The default pipeline review works through the list from top to bottom, giving each deal a moment. It feels thorough and it is the least efficient possible use of the time, because most deals in any pipeline have not changed since the last meeting and nothing said about them will change anything. The result is a meeting where the interesting cases arrive last, when attention has gone, and where the main output is that everybody has recited their list.

It also has a subtler cost: it trains the team to prepare a narrative rather than to surface problems. If you know you will be asked about every deal in turn, the sensible preparation is to have something reasonable to say about each one, and "reasonable" tends to mean optimistic. A review built around exceptions asks a different question — what changed, and what is not moving — and that question cannot be answered with a general impression. Thirty minutes is enough for a review of exceptions and nowhere near enough for a review of everything, which is why the format matters more than the duration.

Start with deals that moved stage since the last review, and keep it brief — this is the good news, it takes a couple of minutes, and its real purpose is to confirm that the record matches reality. A deal somebody describes as progressing that has not moved stage in the system is the first useful finding of the meeting, and it usually means either the record is stale or the progress is imagined.

Second, deals stuck longer than they should be for the stage they are in. This is where most of the time goes and where the meeting earns its place. Third, the forecast for the next 30 days: which specific deals are expected to close, and what has to happen for each. Not a total — a list, because a total conceals whether it depends on one large deal that is not actually ready. Fourth, blockers that genuinely need somebody else to act: a price approval, an introduction, a decision about terms. Anything that can be resolved by the person raising it is not a blocker and does not belong in the meeting.

"Stuck" needs a definition agreed in advance, or the review becomes a negotiation about which deals count. The workable version is a threshold per stage, because stages legitimately take different amounts of time — a week in negotiation is normal, a week between qualified and proposal sent is usually not. Set each threshold from your own history rather than from a general rule, and accept that with a small number of deals these thresholds are rough. Rough and written down still beats precise and argued about.

The reason to fix the definition beforehand is that it changes who is doing the work. When stuck is defined, the list arrives at the meeting already assembled and everyone can look at the same set of deals. When it is not, the first ten minutes go on deciding what to discuss, which is the least valuable conversation available. It also removes the personal element: a deal appearing on the stuck list is an arithmetic outcome rather than somebody's judgement about a colleague's performance, and that makes it considerably easier to discuss what is actually wrong.

Stage conversion rate — the share of deals entering a stage that leave it forwards — is the first. It localises the problem. If most qualified deals reach proposal and most proposals die, the issue is in the proposal or in who is receiving it, and no amount of additional lead generation will fix it. If few leads become qualified, the problem is upstream in who is arriving. This is the number that stops a team from responding to every disappointing month with the same instruction to find more leads.

Average time in stage is the second, and it is the early-warning half. Conversion rates only resolve once deals close, which can be months; time in stage moves immediately. A stage whose average duration is lengthening is telling you about a problem that has not yet shown up in any outcome. Two cautions worth keeping in mind: both metrics depend entirely on stages being updated promptly, so they are measurements of your record-keeping before they are measurements of your sales process. And with a small number of deals, a single unusual case moves an average considerably — look at the direction over several months rather than reacting to one figure.

A review that produces a shared understanding and no assignments will produce the same list next time, with the same deals stuck slightly longer. Each deal discussed should end with one of a small number of outcomes: a named action with a date, a decision to disqualify, a decision to move it to nurture, or an explicit decision to leave it alone until a stated date. That last one is legitimate and under-used — deciding not to act, on purpose, is different from forgetting, and it stops the same deal being rediscussed weekly with no new information.

The hardest and most valuable outcome is disqualification. A review where nothing is ever removed is a review that only adds, and the pipeline becomes a list of everything anyone has ever spoken to. This is where a manager's contribution matters most, because the person who has invested weeks in a deal is the last person able to judge it dispassionately. Making it normal to close deals in the meeting — treated as good hygiene rather than as failure — is what keeps the remaining numbers meaningful.

Weekly is the interval that fits most small businesses, for a practical reason rather than a ceremonial one: it is short enough that stage movement is still fresh in memory and long enough that something has usually changed. Fortnightly reviews tend to spend their first stretch reconstructing what happened, and daily ones do not have enough new information to justify the interruption.

The review also functions as a deadline for record-keeping, and that is a substantial part of its value — the CRM gets accurate shortly before the meeting because the meeting reads from it. That only works if updating happens before, not during. A review that begins with everyone correcting stages has already spent its most useful minutes and, worse, is reading from data that was wrong an hour earlier for anybody who looked. Keeping it to 30 minutes helps here too: a short meeting with a fixed shape is one people prepare for, and a long meeting with a loose shape is one people attend.