Ask ten sales teams how many pipeline stages they use and you will get answers between three and fourteen. The teams with fourteen are usually unhappy. The teams with three are usually guessing at their forecast. Neither problem is solved by picking a number off a blog post, including this one — but there is a test that resolves it.

The test: does the stage change because the buyer did something?

A stage should represent a change in the buyer's position, not a change in yours.

"Proposal sent" is a stage about you. You sent a document. The buyer may not have opened it. Nothing about their intent has changed, so a pipeline full of deals in Proposal Sent tells you nothing about what will close.

"Proposal accepted in principle" is a stage about them. Something happened on their side that moves the deal materially closer.

Apply that test to your current stages and most lists get shorter quickly. The stages that fail it are usually internal admin steps that felt important when someone designed the process.

What too many stages costs you

The cost is not conceptual, it is behavioural. Every extra stage adds a decision the rep has to make, and reps resolve ambiguity by not updating.

You can spot this from the data. Symptoms of an over-specified pipeline:

  • Deals cluster in two or three stages while the rest sit near-empty
  • The gap between "last activity" and "stage changed" grows to weeks
  • Two reps put identical deals in different stages
  • Reps batch-update everything the day before the forecast meeting

That last one is the tell. If your pipeline is accurate on Thursday afternoons and fiction the rest of the week, the stage model is too fine-grained for the effort people are willing to spend.

There is a second-order cost too. With fourteen stages, each holds few enough deals that conversion rates between them are statistically meaningless. You end up with a beautifully detailed funnel built on samples of four.

What too few stages costs you

Three stages — roughly Open, Quoted, Closed — is easy to keep current and tells you almost nothing.

The specific loss is forecasting. If half your pipeline sits in one enormous middle stage, you cannot distinguish a deal that had one exploratory call from one that is waiting on a signature. Your weighted forecast is then a single average applied to wildly different situations, which is a more confident way of guessing.

You also lose the ability to diagnose. When conversion drops, a coarse pipeline tells you it dropped. A well-shaped one tells you it dropped between first meeting and requirements agreed, which points at discovery rather than pricing.

A starting shape

For most B2B teams selling something that takes more than one conversation, five to seven stages works. A reasonable default:

  1. New — captured, not yet contacted
  2. Contacted — a real conversation happened, not a voicemail
  3. Qualified — need, budget and authority understood well enough to invest time
  4. Proposal — commercials are on the table
  5. Negotiation — they want it, you are agreeing terms
  6. Won / Lost — terminal

Every one of those, except Proposal, is a statement about the buyer. Proposal survives because in most businesses it does coincide with a real change in the relationship, but it is the stage most worth scrutinising in your own process.

Note what is not there: no "Follow-up", no "Nurture", no "On hold". Those are not stages, they are situations that can occur in any stage. If a deal goes quiet in Qualified, it stays in Qualified with a follow-up date. Creating a parking stage produces a bucket that only ever grows.

Getting the terminal stages right

The stages that matter most for reporting are the ones at the end, and they are the ones teams treat most casually.

Mark explicitly which stages count as won and which as lost, rather than inferring it from position. This sounds pedantic until you try to compute a conversion rate and discover that "Closed" contains both. In Buzzflo's CRM a status carries won and lost flags for exactly this reason, so conversion reporting and the moment a lead becomes a customer are both unambiguous.

Then capture why a deal was lost, with a small fixed set of reasons — price, timing, competitor, no decision, disqualified. Free text will not aggregate. Five options will, and "no decision" turning out to be your largest bucket is one of the more useful things a pipeline can tell you.

Naming, and why it matters more than you think

Stage names get read hundreds of times a week and shape how reps think. Two rules:

Name the state, not the activity. "Requirements agreed" beats "Discovery call". The first describes where the deal is; the second describes something you did, which may or may not have achieved anything.

Use your own vocabulary. If your business calls them enquiries rather than leads, or the clinic-facing team thinks in terms of doctors rather than contacts, the software should match. Fighting a tool's vocabulary is a small, permanent tax on adoption — which is why terminology in Buzzflo is configurable across the whole product rather than fixed.

Reviewing the model

Pipelines drift. A quarterly review with three questions is enough:

  • Which stage holds deals longest, and is that real or is it a parking bucket?
  • Are there stages where fewer than five deals passed through this quarter? Merge them.
  • Did any two people disagree about where a deal belonged? That is an ambiguous stage definition, not a training problem.

Change the model when the answers say so, but not more than quarterly — every change breaks historical comparability, and a stage model you keep re-cutting produces no usable trend at all.

What to do differently on Monday

Export your pipeline and count deals per stage. If more than 60% sit in one stage, you have a bucket rather than a model. If any stage holds fewer than five deals a quarter, you have detail you are not using.

Then take each stage name and ask the test: did the buyer do something? Delete or merge every stage where the honest answer is no. You will almost certainly end up with fewer stages than you started with, and a forecast you believe more.

If you are rebuilding the pipeline anyway, the CRM module lets you define stages, mark won and lost, and add your own fields without waiting on anyone.

sales pipeline softwarecrm software

Buzzflo puts this into practice — see crm & lead management, take the product tour, or compare plans.