Direct answer
A pipeline full of dead deals makes every forecast wrong, because open value is counted that nobody is working. To clean it up: flag deals by age and last activity per stage, review the flagged list in one short meeting with the deal owners, and give each deal one of three outcomes: re-engage with a dated next step, close as lost with a reason, or move to a nurture list outside the pipeline. Then add a light recurring check so stale deals are caught monthly instead of once a year. This is a different job from redesigning your stages (see CRM pipeline stage design) and from automating routine tidying (see manual CRM cleanup automation).
Signs your pipeline is polluted
You rarely need a formal audit to know. The symptoms are usually visible in the pipeline view itself:
- Total open value is far larger than anything the team believes it can close, and nobody trusts the forecast
- Deals sit in early stages with a created date from a previous year
- Reps answer 'what is the next step on this one?' with silence or 'I should call them'
- Deals have no open task, no scheduled activity and no recent note
- Stage conversion looks oddly good for late stages and terrible for early ones, because dead deals stay at the front and only the successful ones ever move
- Close-lost is almost never used, so the only way a deal leaves the pipeline is by winning
Set age and activity thresholds by stage
One threshold for all stages does not work, because a healthy early-stage deal moves faster than a late-stage one waiting on a buyer's procurement. Build the thresholds from your own history instead of copying someone else's numbers:
- Take your won deals and look at how long they typically spent in each stage. That is your normal range for the stage
- Flag a deal for review when its time in the current stage is well beyond that range. A common starting rule is a multiple of your typical time, which you then adjust after the first review
- Add a second flag independent of stage: no logged activity (call, email, meeting, note) for a set number of days that you choose. Two flags catch different failures
- Add a third flag for deals with no future-dated task or next step, since a deal without a next step is stalled by definition
- Treat a past-due expected close date as its own flag
Choose the numbers, then test them
Set the initial thresholds, filter the pipeline, and look at the first twenty flagged deals. If most are clearly dead, the thresholds are about right. If half are live, they are too tight. If nothing is flagged in a pipeline everyone knows is dirty, they are too loose. Adjust once and write the final numbers down.
Run the review meeting
Do not ask reps to clean their own deals silently. They will close the obvious ones and quietly leave the awkward ones. A short, focused review works better:
- Export or save the flagged view before the meeting so everyone works from the same list, sorted by owner
- Give each owner a fixed time per deal, roughly a minute, and ask one question: is there a real conversation with a real next step?
- Decide on the spot: re-engage, close-lost, or nurture. 'Leave it for now' is not an outcome unless a dated next step is entered during the meeting
- A manager or an ops person updates the records live or right after, so decisions do not evaporate
- Keep a record of what was closed and why, so you can look back at patterns
Close-lost reasons: capture them properly
Closing stale deals is only useful if the reason is recorded, otherwise you delete information along with the clutter. Use a short picklist rather than free text, for example: no response, chose a competitor, no budget, timing, not a fit, bought nothing, went with in-house, duplicate or invalid. Keep it to a handful of options that reps can choose without thinking, and include one for 'stale, no contact' so the reason is honest. Reasons only help if they are required at close in your CRM, so check whether your platform can make that field mandatory on a lost status. If it cannot, agree it as a team rule and check it in the monthly review.
Re-engage or close: a decision rule
The hardest calls are deals that were once real. A workable rule keeps this from turning into opinion:
- Re-engage if the last real two-way contact was recent enough that the buyer would remember you, and there was a stated need. Send one specific message and enter a dated follow-up
- Give re-engagement one attempt window. If there is no reply by the end of it, close as lost or move to nurture
- Close as lost if there was never a real two-way conversation, the contact left the company, or the need has been met elsewhere
- Move to nurture if the timing is the only obstacle and the buyer said so. Keep them as a contact with a follow-up date, not as an open deal inflating the pipeline
- Do not reopen closed deals to hide them from a report. If the deal comes back to life, create a new deal or reopen it with a note, following one consistent convention
What the cleanup does to your forecast
Expect open pipeline value to fall, sometimes sharply, and warn stakeholders before it happens. That drop is not lost revenue; it is the removal of value that was never going to close. Two side effects to check afterwards: stage conversion rates will change because the denominators change, and time-in-stage averages will improve because the oldest outliers are gone. Do not compare pre- and post-cleanup reports as if they measure the same thing. Note the cleanup date in your reporting notes so a later reader knows why the trend line steps.
Keep it clean
A one-off cleanup decays. What keeps it from decaying is a small routine, not a heroic project:
- A saved 'stale deals' view built on the thresholds above, reviewed by each owner monthly
- A rule that every open deal has a future-dated task, with a view that lists the ones that do not
- A weekly pipeline meeting that starts with the oldest deals rather than the newest
- Optionally, an automated reminder to the owner when a deal crosses a threshold. Check what your platform supports; the reminder is a nudge, and the decision should stay with a person
When it is bigger than a checklist
If deals are stale because the stages themselves do not describe how your buyers progress, or because nobody owns the pipeline, cleaning only postpones the problem. LATYNEX does not promise that anything will close automatically or that a pipeline will stay clean without an owner. Where a structured review helps, the CRM Cleanup & Sales Workflow Setup package (€1,990, fixed scope) covers one CRM and one principal pipeline; larger or multi-pipeline work is scoped separately. See CRM & Sales Workflow Implementation for what is reviewed and how it is delivered.
Questions
How old is too old for a deal to stay open?+
There is no universal number. Compare a deal's time in its current stage with how long your won deals typically spent there, and flag anything well beyond that range. Then check activity: a deal with recent two-way contact and a dated next step is alive whatever its age.
Should we just auto-close deals after a set number of days?+
We would not close deals automatically without a review. An automated flag or reminder is a reasonable nudge, but closing loses information unless a person records why. If you do automate closing, log a distinct close-lost reason such as 'stale, no contact' so the data stays honest, and check what your platform supports.
Will deleting stale deals damage our reports?+
Close them rather than delete them. Closed-lost deals keep their history and reason, and reports on win rate stay meaningful. Deleting removes the record entirely, which you generally cannot undo, so treat deletion as a last resort for genuine junk such as test or duplicate deals.
How often should we repeat the cleanup?+
Do a full review once, then a lighter monthly pass on a saved stale-deals view. Teams with longer sales cycles can stretch the interval; teams with short cycles may want it more often. Let the number of flagged deals guide you: if the list is short, the routine is working.
Who should own the pipeline cleanup?+
The deal owner decides each deal, and one named person, usually a sales manager or ops lead, owns the process and the thresholds. Without that named owner the cleanup tends to happen once and fade.