How to run a sales pipeline that tells the truth
Most pipelines are a list of wishes. What each stage should mean, when a deal is real, and the one rule that keeps the forecast worth reading.
A pipeline has one job: let you look at it and know what is actually going to close. Most fail at that, and they fail the same way. Deals go in on optimism, nothing ever comes out except by closing, and within a quarter the total is a number everybody quietly ignores.
The fix is not a better layout. It is agreeing what a stage means and being willing to remove things.
A stage is an observable fact, not a feeling
The single most useful discipline is defining every stage by something the buyer did, not something you hope. "Interested" is a feeling and cannot be checked. "Asked for pricing" is a fact, and two people would agree on it.
| Weak stage | Why it rots | Observable version |
|---|---|---|
| Interested | Nobody can disagree with it, so nothing ever leaves | Replied asking a question about the product |
| Qualified | Means something different to every rep | Confirmed budget holder and a timeframe |
| Proposal | Often means "I intend to send one" | Proposal sent, on a date you can name |
| Negotiation | Frequently one unanswered email | They came back on price or terms |
When to create an opportunity
After a real reply from a real person, expressing something you could describe to a colleague. Not when you add a promising company, not when you send a sequence, not when someone opens an email four times.
Creating deals earlier feels productive and costs you the one thing the pipeline is for. A pipeline holding forty deals where eight are real is less useful than one holding eight, because now you have to work out which eight every time you look.
The hygiene rule
If a deal has not changed stage in 30 days, it is not moving. Close it, or write down the specific thing you are waiting for and who owes it.
This is the rule nobody enjoys, and it is the whole game. A pipeline that only ever grows is a list of wishes, and its main effect is to hide the deals that genuinely are moving underneath the ones that are not.
- Closing a stalled deal does not delete anything. The contact, the company and the whole history stay.
- A deal that comes back to life is a new opportunity, and that is the honest way to record it.
- Reps are not punished for closing deals as lost. They are punished, eventually, for a forecast nobody believes.
Value it once, then leave it alone
Put a realistic number on a deal when you create it and resist adjusting it upward as you get more excited. Deal values that drift up as a deal ages are the clearest sign a pipeline is being managed by mood.
If the scope genuinely changes, change the value and know that you did. That is different from rounding up because a call went well.
Reading it
Once stages mean something, three questions become answerable at a glance:
- Is there enough at the top? A healthy pipeline is widest at the earliest stage. If it bulges in the middle, you stopped prospecting a month ago and have not felt it yet.
- Is anything stuck? A stage where deals arrive and never leave is a broken step in your process, not a run of bad luck.
- What is actually closing? Only the late stages, weighted honestly. The early ones are hope with a date attached.
Do
- Define every stage by something the buyer did.
- Create an opportunity only after a real reply.
- Close anything that has not moved in 30 days.
- Set the value once, realistically.
- Review the whole pipeline weekly, briefly.
Don't
- Do not create a deal for every promising company. That is a target account, not an opportunity.
- Do not advance a stage because time passed. Stage means what is true.
- Do not keep a deal open because closing it feels like giving up. It is the same outcome, recorded honestly.
- Do not use the pipeline as a task list. A reminder is a task, not a deal.
How many stages should a pipeline have?
Four to six. Fewer and it tells you nothing, more and reps guess between neighbouring stages, which puts noise into every report built on it.
Should I reopen a lost deal that came back?
Create a new opportunity instead. Reopening rewrites history and makes your cycle times and win rates meaningless.
What if a deal is genuinely slow rather than dead?
Then it has a specific next step and a date. Write those down. "Slow" with neither is the same as dead, and pretending otherwise is how a pipeline stops being believed.