Every pipeline review eventually hits the same slide: a column of deals that have not moved since last month. The usual response is a round of "just checking in" emails. Most of them go unanswered, the column stays the same length, and next month it is longer.
A stalled pipeline is not one problem. It is a set of deals stuck for different reasons, and each reason needs a different move. Triage means sorting them before you act. This is a method you can run in one working session with a CRM export and a quiet room.
First, define stalled
"Stalled" needs a rule, or the conversation turns into opinions about individual deals. A workable definition has two parts:
- No stage change for longer than your normal time in that stage. If deals usually spend two weeks in discovery, a deal at six weeks is stalled.
- No two-way contact in the same period. An email you sent does not count. A reply, a meeting or a call does.
You need both. A deal can sit in one stage for a long time while legal reviews a contract, and that is not stalled. A deal can change stage because someone tidied the CRM, and that is not progress.
If you do not know your normal time per stage, use a rough number for now and note it as an assumption. Precision matters less than applying the same rule to every deal.
Pull the list, then stop looking at value
Export every open deal that meets the rule. Include the owner, stage, last two-way contact, the last thing the buyer said, and the next step recorded in the CRM, if any.
Now hide the deal value column. Value pulls attention toward the biggest numbers, and the biggest numbers are often the ones least likely to close. Triage first. Prioritise second.
Sort every deal into one of five buckets
Read each row and put it in exactly one bucket. If you cannot decide, the deal goes in the fifth.
- Waiting on us. The buyer asked for something, a proposal, a security answer, a revised quote, and we have not sent it. This is the most common bucket and the most embarrassing. The fix is internal.
- Waiting on them, with a date. The buyer said they would come back after a board meeting or budget cycle. Not stalled in any useful sense. Set a task for the date and leave it alone.
- Lost the thread. The contact went quiet, changed role, or left. The deal needs a new route in, not another email to the same address.
- Never qualified. Looking back at the notes, there was no budget, no timeline or no real problem. The honest move is to close it and stop counting it as pipeline.
- Unknown. Nobody can say what is going on. This bucket tells you about your CRM discipline as much as the deal.
The sorting takes longer than you expect the first time, mainly because the CRM notes are thin. That is a finding in its own right.
One move per bucket
Each bucket has a default action. Owners can override with a reason, but the default keeps the session moving.
- Waiting on us: send the thing within two working days, with a short note that owns the delay. No "just checking in".
- Waiting on them: a dated task, no outreach before the date.
- Lost the thread: find a second contact at the account, or a relevant reason to re-open with the original one. A new piece of information beats a nudge.
- Never qualified: close as lost with the real reason. If you want to keep in touch, move the contact to a nurture list with their consent in order.
- Unknown: the owner has one working day to find out which bucket it belongs in.
Write the reason, not just the stage
Whatever you decide, write the bucket and the reason into the CRM. Next month's triage starts from those notes, and the trend across months tells you where the process breaks. A pipeline where half the stalls are "waiting on us" has a capacity or process problem, not a demand problem.
What this does to the forecast
Triage usually shrinks the pipeline on paper, because never-qualified deals get closed. That can be an uncomfortable conversation with a board or a founder who likes the bigger number. Have it anyway. A smaller pipeline you believe is worth more than a larger one you have to caveat every time you present it.
It also makes the forecast easier to defend. Deals in the "waiting on them, with a date" bucket have a real next event. Deals in the "unknown" bucket should not be in the forecast at all until they move.
Where the Pipeline Crew fits
The Pipeline Crew does the preparation for this session on a weekly cadence, so you start from a sorted list instead of an export.
Its output is a pipeline summary, a stalled-lead list, owner priorities and handoff recommendations. In practice that means the list above, grouped, with a suggested owner and a drafted next step for each line. The crew does not decide what to close, and it does not contact anyone.
People approve three things before anything changes: who owns a lead, any update written to a CRM record, and any outbound follow-up. If a draft follow-up is wrong, you edit or reject it, and it stays a draft. The full list of what waits for a person is on the agent handoffs page.
One limit worth stating plainly. The crew's diagnosis is only as complete as the data it can see. If reply data is not connected for a source, the report marks it as unavailable rather than treating silence as "no reply". Connected tools are listed on the integrations page, including which CRM connections are still coming soon.
Run it this week
- Set the stall rule: time in stage plus no two-way contact.
- Export and hide value.
- Sort into the five buckets, one per deal.
- Apply the default move and write the reason in the CRM.
- Book the same session for next month and compare bucket counts.
For a step-by-step version, see the how-to guides. To see how many outcomes each plan runs at once, see pricing. And if the stall starts earlier, on the page that is supposed to create the pipeline, run a free audit of your site first.