A sales pipeline is not a list of opportunities, it is a sequence of commitments. Every stage should state what the buyer has agreed to and what you have verified. Without an observable entry criterion and a verifiable exit criterion, the pipeline turns into stage scenery and the forecast into a belief. Five stages, defined by buyer actions, are enough to steer.
We walk into the same scene with almost every team we work with. A spotless CRM, neatly aligned columns, cards sliding from left to right. And nobody in the room who can say why a given deal moved to the next column.
You do not judge a sales pipeline by how tidy it looks. You judge it by how solid the criteria are that let a deal cross from one stage to the next. This guide is about designing those stages, not running them: how to inspect them every week is covered separately in our piece on running a pipeline review that moves deals.
What a sales pipeline actually is
A sales pipeline is a view of the opportunities that are genuinely in play, each one sitting at a specific point in your sales process. Every line carries an estimated value, a date, an owner and a next commitment from the customer.
The word that matters is “genuinely”. A pipeline is not a wish list, and it is not a contact database either. It is an inventory of deals that have earned their place.
A well-built pipeline answers three questions without anyone opening a spreadsheet. Where does each deal stand? What has been confirmed with the customer? What is missing to move forward?
Everything else, the charts, the colour coding, the weekly dashboards, is presentation. You can switch CRMs as often as you like; the problem will stay exactly where it is, in how your stages are defined.
Pipeline, funnel and forecast: three different tools
Teams blur the three all the time, and that is precisely when decisions start being made blind.
The funnel describes a flow. It is an aggregate view of conversion, from first contact to closed-won. It tells you where volume drops off. It does not help you manage a single deal.
The pipeline describes a state. It is a snapshot of open opportunities and where each one sits. It is the working tool of the people who own the deals.
The forecast describes a projection. It is the commitment on what will close, and it depends directly on the quality of your stages. A reliable sales forecast does not come from smarter weighting. It comes from sharper definitions.
When the pipeline is fuzzy, the forecast becomes an article of faith. And once the forecast is an article of faith, leadership stops trusting the pipeline. The relationship is over before the quarter ends.
Why a stage without criteria is worthless
A stage is a border. If nobody knows what entitles a deal to cross it, it is just decoration in the CRM.
Pipelines rarely collapse overnight. They go stale. A deal parked at the same stage for weeks, a rep who “has a good feeling about this one”, a manager who waves it through because the quarter looks thin. Each of these says the same thing: the border means nothing.
Look at it from another angle. When two reps put two near-identical deals in different stages, that is not a discipline problem. It is a definition problem. The criterion does not exist, so everyone makes up their own.
And a vague criterion does more than create a mess. It breeds optimism. A misplaced deal climbs the pipeline, props up the end of the quarter, then vanishes without explanation. Nobody lied, and yet everybody was misled.
A stage without an exit criterion is a promise nobody keeps.
Building your pipeline stages, one at a time
The rule fits in one sentence. Every stage has an observable entry criterion and a verifiable exit criterion. Entry states what must be true for a deal to enter the stage. Exit states what the buyer has agreed to do to move on to the next one.
Two principles before the detail.
The exit criterion must be a fact, not an impression. “The customer confirmed the scoping workshop date in writing” is a fact. “We really clicked on the call” is not.
Crossing a border must involve an action by the buyer. As long as only your team is doing things, the deal is not progressing. It is just moving around.
Stage one: identification
Entry: a problem that fits your target profile has been spotted at a named account or contact, with a traceable source.
Exit: a first conversation has happened and a qualifying call is booked or done.
What does not count: a list import, a page visit, a purchased email address. Those are hypotheses, not opportunities. They belong in your prospecting, not in your pipeline.
Stage two: qualification
Entry: a direct conversation has taken place with someone connected to the problem.
Exit: the customer has named the problem in their own words, a rough scope is on the table, the contact’s role in the decision is known, and a dated next step exists.
Classic disqualifiers: the inbound quote request with no conversation behind it, the contact who will not take a call, the deal where nobody knows who decides.
Stage three: discovery
Entry: qualification is done and the next meeting is confirmed.
Exit: the customer has shared what the problem costs them, buying criteria are explicit, the people involved in the decision are identified, and an evaluation timeline is agreed.
This is the stage that does the most damage when it is rushed. Incomplete discovery is never recovered later. You pay for it at the end of the cycle, when price becomes the only topic left. When several people weigh in on the decision, mapping that group is part of discovery, as we explain in selling to a committee, not a person.
Stage four: validation
Entry: discovery is complete and the customer has agreed to evaluate a solution.
Exit: the customer has signed off on the expected value, the proposed solution is scoped, and you have access to the economic buyer or the procurement process.
A deal that reaches this stage without access to whoever controls the budget is not validated. It is on hold, and you cannot invoice on hold.
Stage five: commitment
Entry: value and scope are validated on the customer’s side.
Exit: the contract is signed, or a signature date is set with commercial terms accepted.
Resist the urge to add a “negotiation” stage. Negotiating is not a state; it is what happens inside the final stages. Add columns and you add blind spots.
The usual mistakes
Too many stages
Every stage adds a border to define and police. Beyond a handful, the team stops enforcing criteria and just drags cards around. The right number is the one you can recite from memory, one sentence per stage. If you need a diagram to remember your own stages, you have too many.
An inflated pipeline
A pipeline in which no deal ever moves backwards ends up meaning nothing. If the exit criterion lets a deal through without any action from the customer, every deal creeps forward a little each week. The pipeline becomes an optimistic queue, and the forecast a shared piece of fiction.
No exit criterion
Many teams document what it takes to enter a stage, then let each rep decide when to leave it. That is exactly backwards. Exit criteria protect the pipeline. Entry criteria just fill it.
Criteria nobody can check
“The customer is motivated.” “The need is confirmed.” Statements like these cannot be disproved, so they are useless. A good criterion can be checked by opening the CRM and reading the last interaction. That is also why data entry has to stay light: a criterion nobody fills in does not exist, a point we cover in getting your team to use the CRM.
The stage that never goes backwards
No honest pipeline moves forward all the time. If your stages forbid moving down, they have stopped measuring. They are just flattering you.
How to read the health of your pipeline
Three readings are enough to tell whether your stages hold up.
Stage-to-stage progression. How many deals cross each border over a given period. A sharp drop between two stages points straight at the one whose exit is poorly defined.
Time in stage. Entry date, exit date, gap. If deals pile up at the same border, the criterion for crossing it is too vague to ever trigger. The same measure shows where your cycle stretches, the starting point for any effort to shorten your B2B sales cycle.
Regressions. A share of deals moving backwards is healthy. It shows reps dare to downgrade a deal that went wrong, and that your criteria are tough enough to be failed. A pipeline with no regressions is a pipeline that suits everyone a little too well.
If you want to go one step further, add a fourth reading: the share of opportunities with no dated next step. It measures the team’s execution discipline directly. It is often the most revealing number on the board, and the least comfortable one to put on screen.
Reviewing these numbers has its own rules and its own ritual. Well-built stages are the precondition. Reviewing them properly is the method.
Where to start
Do not rebuild your pipeline in a big offsite workshop. Start where it hurts most.
Write exit criteria before entry criteria. It feels counterintuitive, but it is the order that works. First define what proves a stage is finished, then work out what a deal needs to get in.
Test against past deals. Pull the won and lost opportunities from recent months and count how many would have met your criteria. The result rarely sparks a debate.
Block the move. If a criterion is not filled in, the deal cannot change stage. If your tool cannot enforce that, set up a short check-in that deals only with stage changes.
Treat regressions as information. When a deal moves back, the question is not “who got it wrong” but “which criterion did we misjudge”.
Bring the team along. Clear stages will not survive a sales team built the wrong way round. Structure and pipeline get fixed together, not one after the other.
A pipeline is built, not declared
Write your exit criteria this week. Test them on past deals. You will see exactly where your process leaks, and why the forecast keeps missing.
Rethinking your pipeline is not an admission of failure. It is what makes everything else hold.
Frequently asked questions
- What is the difference between a sales pipeline and a funnel?
- A funnel describes conversion across a population, from first touch to signature. A pipeline describes the state of each open opportunity, one by one, with its stage, value and next commitment. The funnel tells you where volume leaks; the pipeline is what you manage deals with day to day.
- How many stages should a sales pipeline have?
- As few as you can get away with, often five. Each stage should mark a change in the nature of the relationship with the buyer, not a shade of progress. If you cannot explain a stage in one sentence from memory, it is one too many and nobody will enforce its criterion.
- What entry and exit criteria should we use?
- Facts, never feelings. A good exit criterion shifts responsibility to the buyer: they confirm a date, share the cost of the problem, sign off on scope, sign the contract. As long as the next move is still yours, the stage is not done, however good the last call felt.
- Can a deal move backwards in the pipeline?
- Yes, and it has to be able to. A deal moving back proves your criteria are demanding enough to be missed and that reps are willing to downgrade a deal that went sideways. A pipeline that never lets deals regress stops measuring anything: it just records the team's hopes.
- How can I tell if my pipeline is inflated?
- Look for three symptoms: deals with no dated next step, close dates that slip for no stated reason, and opportunities that advance while the buyer has done nothing. All three point to the same root cause, exit criteria that do not require any action from the buyer.