Most pipelines fail in the same way. They are set up carefully, used properly for about six weeks, and then quietly become a place where deals go to sit.
The cause is nearly always a definitional one. The stages describe how the salesperson feels about the deal rather than something the buyer has done — and feelings are not auditable, so nothing ever forces a deal to move or leave.
Stages should describe buyer actions
Compare two sets of stages.
The common version: New → Contacted → Interested → Qualified → Proposal → Negotiation → Closed
A version that stays accurate: New → Contact made → Need confirmed → Proposal sent → Decision pending → Closed
The difference is that every stage in the second set has an observable event behind it. "Interested" is an opinion. "Need confirmed" means the buyer told you what problem they are solving and roughly when — you either have that or you do not.
The test for any stage: could two people look at the deal and disagree about whether it belongs there? If yes, the stage is a feeling, and it will silently corrupt your forecast.
Leads from ads, forms and WhatsApp arrive scored and assigned. Pipelines, follow-ups and forecasts keep every deal moving.
Start free See the demoWrite the exit criterion, not the stage name
For each stage, write one sentence describing what must be true for the deal to leave it. Put it where the team can see it.
- New → someone has attempted contact
- Contact made → the buyer has responded, in any form
- Need confirmed → they have described the problem and a rough timeframe
- Proposal sent → a specific price for a specific scope has been delivered
- Decision pending → the proposal has been acknowledged and a decision date exists
- Closed → won, lost, or explicitly not now
That list takes ten minutes to write and removes most of the ambiguity that makes pipelines rot.
Four to six stages
Fewer than four and everything piles into one bucket, so you cannot see where deals die.
More than six and the distinctions become too fine to apply consistently. Two salespeople put the same deal in different stages, the data stops meaning anything, and — more damaging — updating the CRM becomes tedious enough that people stop doing it.
Every stage you add is a small tax on every deal. Charge it only where the information earns its keep.
Every deal needs a scheduled next step
The single highest-value rule you can enforce.
A deal with no next action is not in your pipeline in any meaningful sense. Nobody is doing anything about it. It sits there inflating the total and giving you a forecast built partly on deals nobody has touched in a month.
Make it explicit: every open deal has a scheduled next action with a date. No next action, and the deal is either closed as lost or moved to a nurture list.
Teams resist this at first because the pipeline total drops. That drop is the point — the number was wrong before.
Let deals go backwards
A deal can move back a stage. Someone who was ready to sign and has gone quiet because their budget was reassigned is no longer at Decision pending, whatever your proposal says.
Pipelines that only move forward accumulate deals stuck at the late stages, which is exactly where over-optimism does the most damage to a forecast.
Probabilities, if you use them, must be checked
Assigning "Proposal sent = 60%" is fine, but only if 60% of deals at that stage actually close. Most businesses inherit these numbers and never verify them.
Once a quarter, take every deal that reached each stage six months ago and calculate what share closed. That is your real probability. Use it.
An unverified probability produces a forecast with a decimal point on it that is no better than a guess — and worse, because people trust it.
Read the pipeline for stalls, not just totals
The useful question is not "how much is in the pipeline". It is where do deals stop.
Look at the average time a deal spends in each stage, and the share of deals that never leave it.
- Piling up at Contact made → your follow-up is too slow or too weak
- Piling up at Need confirmed → you are proposing to people who were never going to buy
- Piling up at Proposal sent → your proposals are not answering the real question, or the price is wrong
- Piling up at Decision pending → you are not talking to the person who decides
Each of those has a different fix, and the pipeline is the only place the difference is visible.
Record why you lost
Closed-lost with a reason is one of the most valuable fields in a CRM and one of the least filled in.
Keep the list short enough to be used: price, timing, chose competitor, no budget, no decision, no response. Six options in a dropdown, not a free text box.
After fifty losses you will have a real answer to "why do we lose", and it is usually not the one the team believes. Teams tend to say price. The data often says slow follow-up or talking to the wrong person.
Two habits that keep it honest
A pipeline decays quietly. Two small rituals stop it.
A weekly ten-minute sweep. Once a week, filter for deals with no activity in fourteen days. For each: move it, close it, or schedule something. This is not a management review — it is hygiene, and it takes ten minutes once people are used to it.
A monthly look at closed-lost. Read the reasons. Not the totals, the actual deals. Patterns show up quickly and they are rarely what the team says in conversation.
Forecasting without fooling yourself
The number most teams report is the weighted pipeline: each deal's value times its stage probability, added up.
It is useful, with two caveats that are almost always ignored.
It is only as good as your probabilities, and those need checking against real outcomes, not inherited from whoever set up the CRM.
It says nothing about timing. A weighted pipeline of ₹40 lakh tells you nothing about this month if half of it is waiting on decisions scheduled for next quarter.
A more honest pair of numbers: committed (deals with a proposal out and a decision date inside the period) and everything else. Committed is what you forecast. Everything else is what you work on so that next month has a committed number.
The stage most teams are missing
Somewhere between "proposal sent" and "closed" there is usually an invisible stage: waiting on someone we have not spoken to.
Deals stall here constantly. The contact is convinced; their finance head, their partner, their head office is not. The salesperson keeps following up with the person who already agrees.
If you find deals sitting in Decision pending for weeks, add a field rather than a stage: who decides, and have we spoken to them? In many businesses that single question explains most of the stalled pipeline.
A pipeline you can set up this afternoon## A pipeline you can set up this afternoon
- Five stages, each defined by something the buyer did
- One written exit criterion per stage, visible to the team
- Every open deal has a scheduled next action
- Deals with no next action get closed or moved out
- A short closed-lost reason list
- A quarterly check of stage probabilities against what closed
None of that needs a consultant or a bigger CRM. It needs the discipline to define stages by evidence rather than optimism — which is the only thing that keeps a pipeline honest once the novelty wears off.
Frequently asked questions
How many stages should a pipeline have?
Between four and six for most businesses. Fewer than four and you cannot see where deals stall; more than six and the distinctions stop being meaningful and the pipeline stops being updated.
Should each stage have a fixed win probability?
Only if you check it against reality. A probability nobody has verified produces a forecast that looks precise and is not.
Why do deals sit in one stage forever?
Usually because the exit criterion is vague, so nobody can say whether it has been met. Define stages by an observable buyer action and the problem mostly disappears.
What should happen to a deal with no next step?
It should not stay in the pipeline. A deal with no scheduled next action is not a deal, it is a hope, and counting it inflates every forecast you produce.
