Sales Pipeline Stages That Two Reps Would Score the Same Way
If two reps would put the same deal in different stages, every number downstream (conversion, coverage, forecast) is measuring opinion.
Pipeline stages look like administration and are actually the foundation every other revenue number sits on. Conversion rates, cycle time, coverage and forecast are all computed from stage membership, so if two reps would score the same deal differently, all four are measuring opinion.
This is for whoever owns the definitions, which in most companies is nobody.
The Rule: Observable Exit Criteria
A stage needs a criterion that either happened or did not, with no judgment involved.
Bad: "Qualified; prospect shows interest." Interest is not observable, every rep reads it differently, and the same deal sits in different stages depending on who owns it.
Good: "Qualified; we have confirmed the budget owner, a timeline, and a named problem the prospect has agreed is worth solving." Three facts, each verifiable, each answerable yes or no by someone who was not on the call.
The test for any definition: could a manager reading the CRM notes, without talking to the rep, confirm the deal belongs in that stage? If not, rewrite it.
A Five-Stage Default
Adapt rather than adopt, but this shape works for most B2B motions.
- Engaged, a two-way conversation exists. They replied and said something real. Exit: a discovery conversation is scheduled.
- Qualified, budget owner identified, timeline established, problem named and agreed. Exit: all three are recorded, not inferred.
- Validated; the solution has been evaluated against their actual situation. A demo on their data, a pilot, or a technical review. Exit: they have seen it work for them specifically.
- Committed, verbal agreement plus a defined path to signature with dates and named owners on both sides. Exit: that plan exists in writing.
- Closed, signed.
Five is usually right. Three loses the resolution you need to find where deals die; eight produces stages nobody can distinguish and reps who guess.
Stages Belong to the Buyer, Not Your Process
The most common design error is naming stages after what your team does: "Demo Delivered," "Proposal Sent."
Those describe your activity, which you control, so they measure effort rather than progress. A deal can move to "Proposal Sent" because a rep sent a proposal to someone who never asked for one.
Name them after what the buyer has done or confirmed. "Proposal Sent" becomes "Commercial Terms Agreed". The first is your action, the second is their position, and only one of them predicts a close.
One Direction, and Say What Happens When It Goes Wrong
Deals should move forward or close. A deal sliding backwards means the earlier stage was scored wrong, which is useful information you should capture rather than hide.
Write down the handling explicitly: a deal that loses its champion or its timeline goes back to Qualified with a note, or closes as lost with a reason. Either is fine. What is not fine is leaving it in Committed because moving it feels like an admission.
And set a hard rule on staleness: any deal with no contact in 60 days closes as lost. Reopening takes ten seconds if it comes back, and a pipeline of forty opportunities where twelve are live makes every metric fictional.
Get Both Sides to Sign the Definitions
The stage where marketing hands to sales is where most organisations argue permanently, and the argument is always the same: marketing hit its number and sales says the leads were unqualified.
Both are usually right, because nobody wrote down what qualified means. A definition signed by whoever owns marketing and whoever owns sales converts a recurring argument into a fact either side can check.
This is the highest-value hour available to a revenue operations function, and it requires no tooling at all.
What the Stages Give You Once They Are Real
Stage-to-stage conversion, which localises a problem instead of telling you "the funnel is down." A drop between Qualified and Validated is a different issue from a drop between Committed and Closed, and only one of them is a pricing conversation.
Cycle time per stage, which surfaces the stage where deals silently sit. A deal spending 60 days in Validated is telling you something a close-rate number hides entirely.
And a coverage number that predicts rather than describes, because it is measured at a stage that correlates with closing.
Where to Put Them
In the CRM, visible at the point a rep changes a stage, not in a document nobody opens. If the definition is not on screen at the moment of the decision, it does not exist operationally.
Review them once a year. Stage definitions drift as the motion changes, and a definition that no longer matches how you sell is worse than none because it produces confident wrong numbers.
Frequently asked questions
How many sales pipeline stages should I have?
Usually five. Three loses the resolution you need to find where deals die, and eight produces stages reps cannot distinguish, so they guess, which puts noise into every metric computed from stage membership.
What makes a good pipeline stage definition?
An observable exit criterion that either happened or did not. The test is whether a manager reading the CRM notes, without talking to the rep, could confirm the deal belongs in that stage.
Should pipeline stages be named after my sales process?
No. "Demo Delivered" and "Proposal Sent" describe your activity, which you control, so they measure effort rather than progress. Name stages after what the buyer has done or confirmed, "Commercial Terms Agreed" rather than "Proposal Sent".
What should happen to deals that go quiet?
Close them as lost with a reason after 60 days with no contact. Reopening takes ten seconds if they come back, and a pipeline where most opportunities are dead makes coverage, conversion and forecast all fictional.
Want help putting this to work?
Talk to a grobot strategist about wiring this into your stack.
Talk to a Strategist →Running outreach for a book of clients? See how benefits agencies run a whole book on one record.
