Cleaning a Pipeline That Has Gone Stale

Pipelines fill up because closing a deal as lost feels final and leaving it open costs nothing today. Multiply that by a year and the board shows two hundred thousand euros in play, of which perhaps a fifth is real. Our guide to pipelines, stages and the deals you are actually going to close ends with maintenance, because a board nobody trusts is worse than no board — people stop looking at it and start keeping their real list somewhere else.

Find the deals that stopped

Sort by last activity rather than by stage or value. What you want is every deal where nothing has happened — no message, no note, no stage change — for longer than your normal sales cycle. For most service businesses that threshold is somewhere between three and six weeks.

Two other lists are worth pulling at the same time: deals whose expected close date has passed, and deals with no expected close date at all. The second group is usually larger than anyone expects and is the reason forecasts drift.

Three outcomes, no fourth

1. Revive it — there is a real next step and someone will take it this week. Set the date and move on.
2. Move it to a recovery pipeline with a review date, for deals that are genuinely just early rather than dead.
3. Mark it lost with a reason. This is the right answer far more often than people are comfortable with.

"Leave it and see" is not an option, because that is precisely how the board got into this state. A deal you cannot describe a next step for is not in your pipeline; it is in your imagination.

“

Marking a deal lost does not lose it. It moves it somewhere you might actually go back for it.

— be digital ai team

Lost is recoverable, forgotten is not

The reluctance to mark deals lost comes from treating it as deletion. It is the opposite: a deal closed as lost, with a reason and a date, becomes a queryable record you can revisit in six months. A deal left drifting in stage two becomes invisible within weeks, because everyone learns to scroll past it.

Recording the reason as you close it takes seconds and is what makes the list worth revisiting, as covered in learning from lost deals.

Bulk actions, carefully

Clearing a hundred stale deals one at a time will not happen, so bulk actions are how this gets done. They are also how a board gets destroyed by a mis-clicked filter.

Two precautions. Check the count before you act — if you expected forty deals and the selection says four hundred, the filter is wrong. And do the reason and the stage change in one operation rather than closing everything first and intending to annotate later, because later does not arrive. Where the bulk change also triggers automations, make sure you know what those will send; a hundred deals moved to lost should not produce a hundred farewell messages.

The weekly habit that prevents it

A quarterly clean-up is a symptom. The fix is five minutes a week: open the list of deals with no activity in fourteen days and deal with each one. At that scale it is three or four deals, and each takes ten seconds — the same review folded into running a weekly business review with your copilot.

If the same deals keep appearing on that list, the problem is upstream. Either your stages let deals sit without an owner or a next step, or your follow-up is manual when it should be automatic, which is what automations that fire when a deal moves exists to solve.

Clean the contacts too

While you are in there, the same logic applies one level down. Duplicate contacts, records with no phone number, and people who unsubscribed months ago all distort your segments and your campaign lists, and none of them announce themselves. Half an hour on that once a quarter is worth more than most reporting work.

Judge the result by the forecast

After a proper clean, your pipeline total will drop, sometimes by half, and that is the point. The number was never real; it was just comfortable. What you have afterwards is a smaller figure you can act on — you can look at every deal on the board, say what happens next, and plan the month against something that is actually going to arrive.