Most B2B SaaS pipelines collapse to three stages, qualified, proposal, closed. Every deal that dies between "problem confirmed" and "contract signed" gets logged as lost in legal or lost to no decision. The pipeline did not lose the deal, it failed to model it. A pipeline that cannot describe what happens between confirmation and signature cannot tell you why a deal stalled, which means it cannot tell you what to do next. The five-stage model below has survived contact with three live B2B SaaS implementations, and it survives because each stage maps to a distinct buyer behaviour and a distinct seller action.
Why this matters now
B2B buying has not gotten easier, it has gotten more crowded. Adamson, Dixon, and Toman argued in Harvard Business Review that solution selling has run out of room, because procurement teams now arrive with their own diagnosis and shortlist. The seller's remaining lever gets pulled before the buyer ever runs procurement: reframe the problem the buyer thought they had, and earn the right to define the outcome. A pipeline with three stages cannot do that work. The deal-stage architecture is the simplification mechanism, when stages are named for buyer behaviour and seller job, the team has a forcing function for what to do next. When the stages are named "qualified, proposal, closed," the team has a filing cabinet.
Stage one: Educate, vision-selling, not feature-selling
The educate stage is where the buyer encounters a frame for the problem they did not have language for yet. The seller's job is not to demo the product: it is to give the buyer a vocabulary for the operating problem the product addresses, and a point of view on why it will get worse before it gets better. Buyer behaviour: a willingness to take a meeting where nothing is sold. Seller action: a reframe: a one-page articulation of the problem, anchored in the buyer's market, stage, and commercial pressure. Exit criteria, the buyer says, in their own words, the problem you came to discuss. If they cannot, you are not out of educate.
Most teams skip this stage and call the first call "discovery." Discovery without education is a checklist of pain points the buyer answered with whatever they thought you wanted to hear.
Stage two, Discover, the working session that earns the right to propose
Discovery is not a script of qualifying questions. It is a working session, ideally with one of the buyer's operating leaders in the room, where the seller maps the buyer's workflow against the frame from educate. Buyer behaviour: they bring data, they bring a colleague, they cancel a meeting to take this one. Seller action: pressure-test the frame, name the two or three places it holds and the one place it does not, write back to the buyer within twenty-four hours. Exit criteria, the buyer agrees, in writing or on a call, that the problem as described is the problem worth solving. Without that, you cannot price a solution to it.
Stage three, Value, the workshop that closes the deal
This is the stage most teams skip, and skipping it is why deals die in legal. The instinct after discovery is to send the proposal, pricing, terms, scope, redlines, and let the buyer's procurement function do the rest. The proposal lands in a committee that has not seen the frame from educate, has not been in the discover working session, and has no commercial reason to defend the deal. The deal dies on the procurement desk and the seller calls it a legal problem.
The fix is a value workshop with the buyer's full committee, before the proposal goes out. One deck. One hour. The seller is selling the vision of the outcome: what the buyer's revenue motion looks like in twelve months, what the renewal curve looks like in twenty-four, what advocacy and expansion patterns they unlock: not the software, not the seat count. The workshop is the moment the committee aligns on the outcome, which is what makes the proposal a formality instead of a negotiation. Buyer behaviour: the full committee shows up. Seller action: present the outcome, not the price. Exit criteria, the workshop was on the calendar and happened. You cannot move out of value until the workshop is done.
Why this stage is missing from most pipelines
Nobody on the team owns it. Marketing owns the top of the funnel, sales owns the close, and the operating decision the buyer is actually making, "will this deliver the outcome we agreed to", happens in a no-man's-land between them. The value stage names that gap and gives it an owner.
Stage four: Setup, contracting, kickoff, the stages CS inherits
Setup is the stage where the deal is commercially won but operationally undefined. The contract is in legal review or signed pending kickoff scheduling. Buyer behaviour shifts from evaluating to onboarding: implementation timelines, change management, handoff from buying committee to operating team. Seller action: make the kickoff feel inevitable. Confirmed kickoff date, named owners on both sides, a one-page implementation plan the customer success owner has already read. Exit criteria, contract signed and kickoff scheduled. The cleanliness of the inheritance determines whether the first ninety days are a smooth onboarding or a recovery project.
Stage five: Closed Won, a single, defended definition
Closed won is one of the most abused stage names in B2B SaaS pipelines. In an inherited HubSpot or Salesforce instance it sometimes means contract signed, sometimes kickoff scheduled, sometimes revenue recognized, and frequently whatever the rep needed it to mean to hit the quarter. The five-stage model defines it as one thing, contract signed and counter-signed, implementation scheduled, customer success owner named, revenue event recognizable to finance. Anything earlier is still in setup. The stage definition is a forcing function for clean reporting, and clean reporting is the only thing the forecast depends on.
Pattern from the field
A Series B B2B SaaS team in DACH came to us with a four-stage pipeline, new, qualified, proposal, closed. Win rate from proposal was running below thirty percent, and the team was treating that as a pricing problem. It was not, it was a stage-definition problem. We ran the five-stage model against ninety days of historical deals and found that roughly two-thirds of the deals lost from proposal had skipped what we now call the value workshop. We rebuilt the pipeline against the five stages, made the value workshop a hard exit criterion, and rewrote the deal stage definitions in HubSpot with a one-sentence definition, entry criterion, exit criterion, and owner per stage. The first quarter on the new pipeline shifted win rate on stage-five deals to a number the founder was willing to put in a board deck.
Resolution, a five-stage pipeline playbook
For any B2B SaaS team about to redesign or audit a sales pipeline:
- Name the stages for the buyer's behaviour and the seller's action. Not "qualified" or "proposal": educate, discover, value, setup, closed won. The name has to tell the rep what the next move is.
- Write a one-sentence definition for each stage. Add an explicit entry criterion, an explicit exit criterion, and a named owner. If a stage cannot be defined in one sentence, it is doing too much.
- Make the value workshop a hard exit criterion. No deal moves to setup without a workshop on the calendar with the buyer's full committee. Do not weaken the forcing function.
- Audit ninety days of historical deals against the new stages. Map every won and lost deal to the stage it actually died in. The gap between legacy labels and reality is the diagnostic.
- Rebuild the HubSpot or Salesforce pipeline against the new stage definitions. One pipeline, five stages, one owner per stage. Delete legacy stages that no longer map.
- Train the team on the stage transitions, not the stage names. The team needs to know what moves a deal from educate to discover, not how to spell educate.
- Report on stage-to-stage conversion, not aggregate win rate. A pipeline whose stages map to buyer behaviour gives you a velocity reading at every transition. That is what a forecast is supposed to be.
Where Checkpoint comes in
Pipeline architecture is RevOps work that pays for itself in the first quarter. The five-stage model is one of the standard frames we run on inherited and greenfield pipelines at Checkpoint, and the audit-and-rebuild cycle on a typical HubSpot pipeline takes four to six weeks. If your pipeline is collapsing to three stages and the team is calling stalled deals "legal problems," the fix is a stage architecture that forces the work before the proposal goes out. Talk to us before you redesign the pipeline a second time.