Most early-stage B2B companies do not have a sales process; they have a collection of wins that happened in different ways for different reasons, which feels like a process because deals keep closing but is actually closer to a winning streak. The tell is that you cannot reliably explain why you won the last five deals or predict whether you will win the next five — each came together through a different sequence of events, driven largely by the founder's instinct and improvisation. That works until it does not: it cannot be taught to a hire, measured for weak points, or scaled beyond the founder, because there is no defined thing to teach, measure, or scale. Moving from random wins to a repeatable process means imposing a defined structure — a set of stages every deal moves through, with clear criteria for advancing — so that winning stops being a series of lucky improvisations and becomes a system that produces predictable outcomes. This guide lays out the six stages a repeatable B2B process needs and, more importantly, what makes them repeatable rather than decorative.
The distinction that matters is between having stages and having a repeatable process, because most companies that draw a pipeline still win at random. Stages alone — names on a board that reps apply by feel — do not make a process repeatable; what makes it repeatable is that each stage has a defined entry and a verifiable exit criterion, so that a deal's position reflects a real, checkable state rather than a rep's optimism. A repeatable process is one where two different people, looking at the same deal, would place it at the same stage, because the criteria are objective. That objectivity is what allows the process to be taught, measured, forecast, and improved — and its absence is why so many companies with a tidy-looking pipeline are still, underneath, winning by luck.
Why Random Wins Eventually Fail You
Winning at random feels fine while it works, which is exactly what makes it dangerous. The problems are all deferred. You cannot hire into a random process, because there is nothing to onboard a rep against — they cannot learn a sequence that does not exist, so they either improvise (badly, lacking the founder's instinct) or flounder. You cannot diagnose a random process, because without defined stages there is no way to see where deals are dying; the engine is a black box that either produces a win or does not. You cannot forecast a random process, because predicting outcomes requires knowing where deals are and how they typically progress, which a random process does not track. And you cannot improve a random process, because improvement requires isolating which part is weak, which requires parts. Random wins, in other words, defer every scaling problem to the moment you try to scale — at which point all of them arrive at once.
The Six Stages
A repeatable B2B sales process can be modeled in six stages. The exact names matter less than the principle that each has a defined exit criterion — a verifiable condition that must be true to advance.
- 1 · Target. Identify accounts that match the ICP. Exit criterion: the account verifiably fits the defined profile — not "seems interesting," but meets the firmographic and trigger signals.
- 2 · Engage. Open a real conversation with the right person. Exit criterion: a two-way dialogue with someone who has the standing to evaluate, not a one-sided outreach.
- 3 · Qualify. Confirm there is a real, prioritized problem and a path to a decision. Exit criterion: verified pain, a decision process, and a plausible timeline — not just polite interest.
- 4 · Diagnose & Demonstrate. Understand the specific problem deeply and show how you solve it. Exit criterion: the buyer agrees your solution addresses their actual, articulated problem.
- 5 · Commit. Align on terms and a mutual plan to close. Exit criterion: agreement on scope and price, and a shared, dated plan to signature with the real decision-makers engaged.
- 6 · Close. Sign and hand off cleanly to onboarding. Exit criterion: a signed agreement and a defined transfer to delivery, so the win does not leak at the finish line.
Notice that every exit criterion is something you can verify, not something you feel. That is the whole point: it is what turns the six stages from labels into a repeatable process.
Stages without exit criteria are just labels. The Startup Sales Engine Playbook gives you the six stages with the verifiable exit criteria that make a process actually repeatable — the architecture we install on build engagements. Download it and turn random wins into a system.
Get the Sales Engine Playbook →The Exit Criteria Are the Process
If you take one thing from this, take this: the stages are not the process — the exit criteria are. Any competent founder can draw six stages on a board in five minutes; what takes real work and real judgment is defining, for each stage, the specific verifiable condition that proves a deal has genuinely earned its way forward. Without that, the stages are theater: reps move deals based on hope, the pipeline fills with deals sitting in stages they never actually qualified for, and the forecast built on those positions is fiction. With well-defined exit criteria, the pipeline becomes an honest map of reality — every deal's position reflects a checkable state — and that honesty is what makes everything downstream possible. The difference between a company that wins at random and one that wins repeatably is rarely the stages they drew; it is whether they did the harder work of defining what it actually takes to leave each one.
Why This Matters Most for Plateaued Companies
The move from random to repeatable is most urgent for companies that have plateaued — that grew on founder-driven random wins and then stalled, unable to push past the ceiling of what improvised selling can produce. A plateau under healthy demand is very often a process problem in disguise: the company outgrew what random winning can sustain but never built the repeatable structure that would let it scale, so growth flattened at the limit of the founder's personal bandwidth and instinct. For these companies, imposing a repeatable process is not optional polish; it is the specific intervention that breaks the plateau, because it converts the founder's tacit winning approach into a system that others can run and that can be measured and improved. If your company is winning deals but has stopped growing, the random-to-repeatable transition is likely the lever — and identifying exactly where your process breaks down is precisely what a revenue diagnostic is built to surface.
The most common self-deception here is mistaking a clean-looking pipeline for a repeatable process. You can have six well-named stages and a CRM full of deals neatly sorted into them, and still be winning entirely at random — if the stages are applied by feel rather than verifiable criteria, the tidiness is cosmetic. The test is not whether your pipeline looks organized; it is whether two people would independently place the same deal at the same stage. If they wouldn't, you have a tidy random process, which is still random.
How to Tell You're Still Winning at Random
Because random winning disguises itself as a process, it helps to know the specific symptoms. The clearest is that you cannot answer "why did we win that deal?" with anything more precise than a story — each win has its own narrative, and no common pattern runs through them. A second symptom is that your win rate swings unpredictably from period to period with no identifiable cause, because there is no consistent process whose performance you could track. A third is that your forecast is routinely wrong in both directions, because the deals' stage positions do not correspond to any real, checkable state. A fourth is that your first sales hire is struggling despite being capable, because there was no defined process to onboard them against. And a fifth is that when a deal stalls, no one can say why or what to do, because there is no model of how deals are supposed to progress to compare against.
Any one of these is a sign; several together is near-certainty that you are winning at random behind a pipeline that looks organized. The reason to name the symptoms is that founders rarely conclude on their own that their process is random — it feels like a process from the inside, because deals do close. It takes the specific symptoms, or an outside diagnostic, to reveal that the closing has been happening through improvisation rather than system, and that the apparent process is a narrative laid over a streak.
What a Defined Exit Criterion Looks Like
To make the difference concrete, take the Qualify stage. In a random process, a deal "is qualified" when the rep feels good about it — they had a nice call, the prospect seemed interested, it has positive energy. None of that is checkable, so two reps would qualify differently and the same rep would qualify differently on different days. In a repeatable process, the Qualify exit criterion is specific and verifiable: the buyer has articulated a real, prioritized problem in their own words; there is an identified decision process and the people in it; and there is a plausible timeline tied to a real driver. A deal that cannot meet all three does not advance, regardless of how good the call felt. The difference is enormous: under the verifiable criterion, a deal in the Diagnose stage is genuinely a deal where qualification was confirmed, so the conversion rate from Qualify to Diagnose actually means something, and the forecast built on it is trustworthy.
Apply that same rigor to every stage and the process transforms. The exit criteria do not just organize the pipeline; they change rep behavior, because reps now have to earn each advance against an objective bar rather than their own optimism — which means deals that were never real get exposed early instead of lingering in late stages inflating the forecast. This behavioral effect is part of why well-defined exit criteria are worth the work to build: they do not merely describe reality more accurately, they force the team to confront it sooner.
Building the Repeatable Process Without Guessing
Defining the six stages and their exit criteria for your specific business is harder than it looks, because the right stages and criteria are derived from how your deals actually win — which requires both the data of your past deals and the judgment to read the real pattern in them rather than the story you tell yourself. The exit criteria, especially, are where inexperience shows: set them too loose and the process is random with extra steps; set them on the wrong signals and the process tracks activity instead of buyer commitment, producing a forecast that misses. Getting this right is one of those tasks where having built repeatable processes many times before compresses what would otherwise be quarters of trial and error into a deliberate, correct build — because the operator already knows which exit criteria predict closed-won and which merely feel like progress. The stages are simple to name and genuinely difficult to define well, which is exactly why the random-to-repeatable transition so often stalls when attempted from scratch.
A tidy pipeline isn't a repeatable process. If two people would place the same deal at different stages, you're still winning at random.RRClosers
Most early companies don't have a sales process — they have a collection of random wins that can't be taught, measured, forecast, or scaled. A repeatable process imposes six stages — Target, Engage, Qualify, Diagnose & Demonstrate, Commit, Close — but the stages aren't what make it repeatable. The verifiable exit criteria are.
The test of repeatability is whether two people would place the same deal at the same stage; if not, a tidy pipeline is still a random one. This transition matters most for plateaued companies, where it's often the specific lever that breaks the ceiling — and defining the exit criteria well is the hard, judgment-heavy part that operators who've built many processes get right the first time.
FAQ: Repeatable Sales Process for B2B Startups
Defined stages with verifiable exit criteria — a checkable condition that must be true to advance a deal. The test: two different people looking at the same deal would place it at the same stage, because the criteria are objective. That objectivity is what lets the process be taught, measured, forecast, and improved.
Target (verified ICP fit), Engage (a real two-way conversation with the right person), Qualify (verified pain, decision process, timeline), Diagnose & Demonstrate (buyer agrees you solve their actual problem), Commit (agreed terms and a dated mutual close plan), and Close (signed and cleanly handed off). The names matter less than each having a verifiable exit criterion.
They defer every scaling problem to the moment you scale. You can't hire into a random process (nothing to onboard against), can't diagnose it (no stages to see where deals die), can't forecast it (no defined progression), and can't improve it (no parts to isolate). It works until you try to grow, then all the deferred problems arrive at once.
No. A tidy pipeline isn't a repeatable process if the stages are applied by feel rather than verifiable criteria — that's a tidy random process, which is still random. The test is whether two people would independently place the same deal at the same stage. If they wouldn't, the organization is cosmetic.
A plateau under healthy demand is often a process problem in disguise: the company outgrew what random winning sustains but never built the repeatable structure to scale past the founder's bandwidth. Imposing a repeatable process converts the founder's tacit approach into a system others can run — frequently the specific lever that breaks the plateau.
Because the right criteria are derived from how your deals actually win, which takes both the data of past deals and the judgment to read the real pattern. Set them too loose and the process is random with extra steps; set them on the wrong signals and you track activity instead of buyer commitment. Operators who've built many processes know which criteria predict closed-won.