Defining your pipeline stages. Criteria that actually move deals forward.

How to define pipeline stages: the short answer.

To define pipeline stages properly, each one should represent a specific, observable thing the buyer has done, not a task the seller has completed or a feeling the seller has about how the deal is going. "Sent proposal" is a stage. "Feels positive" is not, and yet a huge number of CRM pipelines are effectively built around the second kind of judgement because the stages are too vague to gate on anything concrete. Get the stage definitions right and forecasting, coverage and win-rate reporting all become more reliable, because they are built on data that means the same thing for every deal and every rep. If you are building or resetting a sales pipeline management process end to end, defining these stages properly is the foundation everything else sits on.

The standard structure most B2B teams land on, whatever they call the individual stages, runs: lead in or first contact, qualified, meeting booked or discovery complete, proposal or quote sent, negotiation, and closed won or closed lost. Five or six stages is usually enough. More than that and reps spend time updating stage names instead of selling; fewer than that and the pipeline stops telling you anything useful about where a deal actually sits.

How it works in practice.

Give every stage an exit criterion, not just a name.

The stage name tells a rep where a deal sits. The exit criterion tells them what has to be true before it can move to the next one. Without an exit criterion, "qualified" means whatever the rep decides it means that day, and every deal in that stage is really at a different point in the buying process even though the CRM shows them as equal.

A workable set of exit criteria looks like this:

Qualification frameworks like BANT (budget, authority, need, timeline) or MEDDIC (metrics, economic buyer, decision criteria, decision process, identify pain, champion), developed originally at PTC in the 1990s, give you a ready-made checklist for the "qualified" exit criterion rather than leaving it to individual judgement. You do not need to adopt either wholesale. Picking three or four of their questions and turning them into your own exit criterion works for most founder-led sales teams.

Match stages to what the buyer does, not what your team does internally.

A common mistake is building stages around internal admin: "Proposal drafted" instead of "Proposal sent and acknowledged", or "Contract prepared" instead of "Contract signed". Internal-facing stages measure your own workload, not the buyer's progress, and they let deals sit in a stage indefinitely because nothing external forces them to move.

Every stage name and exit criterion should be checkable against something the buyer said or did, ideally something you could point to in an email thread or call recording if someone challenged it.

Handling the handoff between SDRs and AEs.

If your process splits prospecting and closing across two roles, the stage where a deal changes hands needs its own explicit exit criterion, separate from qualification itself. "Meeting booked" is not the same as "meeting held and confirmed as a fit", and treating them as one stage hides the point where deals actually get lost between an SDR's pipeline and an AE's.

Track the conversion rate specifically across that handoff stage. A high volume of meetings booked but a low rate of meetings converting to a genuinely qualified opportunity usually means the SDR's qualification bar and the AE's definition of qualified are not aligned, which is a conversation to have directly rather than a problem the CRM can fix on its own.

Keep stage count between five and seven.

More granular stages feel like better visibility, but they usually create the opposite. Reps either skip stages they cannot be bothered updating, or a deal technically sits in the wrong stage for weeks because nobody remembered to move it through three intermediate steps. Five to seven stages, each with an unambiguous exit criterion, gives you enough resolution to forecast accurately without asking for more discipline than a busy sales team will actually maintain.

Rolling out redefined stages without losing pipeline history.

Changing stage definitions on a live CRM is not just a settings change. Deals already sitting in the old stages need mapping to the new ones, or forecast history breaks at the point of the change and nobody can compare quarter to quarter cleanly. Map every old stage to a new one before you switch, communicate the change and the reasoning in a single message to the whole team, and run the old and new definitions side by side in a spreadsheet for the first two weeks so reps have something to check against while the new habit forms.

A team that resists a stage redesign is rarely resisting the idea. More often they are worried about the extra admin of reclassifying a live pipeline by hand. Doing that reclassification centrally before launch, rather than asking each rep to work through their own deals in their own time, removes the main objection and gets you to consistent data faster.

Mapping stages to forecast categories.

Most CRMs separate the stage a deal sits in from the forecast category it rolls up into, usually something like pipeline, best case, commit and closed. Confusing the two is a common source of forecast noise: a deal can genuinely be in the negotiation stage while the honest view of whether it closes this quarter is still best case rather than commit.

Set a simple rule connecting the two. A deal only moves into the commit category once it has met the exit criterion for negotiation and the close date sits within the current forecast period. Without that rule, individual reps apply their own judgement about what counts as committed, and the forecast becomes a collection of personal opinions rather than a consistent read of the pipeline.

What good looks like.

A well-defined pipeline shows a fairly even distribution of deals across stages, not a pile-up at "proposal sent" that never seems to move. Stage-to-stage conversion rates are stable enough month to month that you can use them to forecast, and a manager reviewing the pipeline can look at any deal's stage and know, without asking the rep, roughly what has and has not happened yet.

This is also where pipeline stages connect to the wider reporting picture. As covered in my guide to SaaS KPIs, most B2B teams aim for three to four times pipeline coverage against target, and a 25% win rate on qualified pipeline is a commonly cited benchmark. Neither number means anything if the stages underneath them are not consistently defined: a "qualified" pipeline that is really a mix of genuinely qualified and barely touched leads will always undercount the coverage or win rate you actually have. Tracking whether reps actually keep to these definitions in practice is one of the metrics covered in CRM adoption metrics.

A quick health check worth running monthly: pull every deal in each stage and check the exit criterion against reality for a random sample of five. If more than one or two fail the check, either the criteria are not clear enough or nobody is enforcing them in pipeline reviews, and the fix is usually the review habit rather than the stage design itself.

Run a deeper audit quarterly rather than monthly: look at the full distribution of open deals across stages, not just a sample, and check whether it has shifted in a way that suggests a genuine change in the sales process, more deals stuck in negotiation, say, or fewer reaching proposal, versus a change that just reflects looser stage discipline creeping back in. The two look similar on a dashboard but need very different responses.

Pitfalls to avoid.

• Stages with no exit criterion. Without one, stage placement becomes a matter of individual judgement, and the pipeline stops being comparable across reps.
• Naming stages after internal tasks. "Proposal drafted" measures your team's workload, not the buyer's progress. Name stages after what the buyer has done.
• Too many stages. Beyond seven or eight, reps stop maintaining them accurately and the extra granularity adds noise, not insight.
• Redefining stages without remapping old deals. A relaunch that leaves existing pipeline unmapped breaks forecast history at the point of change and makes quarter-on-quarter comparison meaningless.
• Never revisiting the stage definitions. A pipeline built for one sales motion will not automatically fit a new product line, a new segment, or a longer sales cycle. Review the stages whenever the sales process changes meaningfully, not just at CRM setup.
• Letting dead deals sit in an active stage. A deal that has gone quiet for months is not still "in negotiation". Build a routine for moving stalled deals out, or the pipeline overstates what is actually live.