Most B2B sales processes lose deals in the same three places. None of them is the one founders blame. Here is the diagnostic that finds yours.
//THE SHORT VERSION → A sales process is not a CRM. It is the defined sequence a deal moves through, and most B2B processes leak revenue silently because no stage is measured. → Before you optimize anything, settle one question: do you have a process that leaks, or no process at all? The answer changes the entire work plan. → Map your deals to seven stages and measure the conversion between each against your own prior quarters. External stage benchmarks cannot transfer, and chasing one is how founders fix the wrong stage. → Deals rarely die where they appear to die. The deal lost at proposal was usually never qualified at discovery. → Late-stage stalls are mostly a committee problem, not a price problem. In a 2024 survey, Gartner found 74% of B2B buying teams experience unhealthy internal conflict, with buying groups running from five to 16 people. → Sales cycles are getting longer for 57% of sales teams, so slippage is now a structural condition to design around, not a bad quarter to wait out. → CRM hygiene is not admin. A process you cannot see in the pipeline report is a process you cannot optimize. |
Every founder who has scaled past founder-led selling eventually hits the same wall. Revenue is coming in, but it is lumpy. Some quarters land, some miss, and nobody can say exactly why. The instinct is to blame lead volume or hire another rep. The real problem is usually upstream of both: there is no sales process, or there is one on paper that nobody actually runs.
This guide is the diagnostic. It walks the seven stages, shows you where deals actually die, and gives you the three fixes that carry most of the outcome. But it starts somewhere most guides skip, because getting this first question wrong wastes two quarters.
First, answer the only question that changes the work
There is a fork here, and choosing wrong is expensive. Optimization assumes a process exists and leaks. Rebuilding assumes there is no real process to fix.
Optimize when reps broadly follow the same stages, the CRM reflects most deals accurately, and the problem is a measurable drop at one or two stages. That is a tuning job.
Rebuild when there is no shared definition of the stages, every rep sells their own way, the CRM is an unreliable record, and the founder is still the only person who can reliably close. There is nothing to optimize because there is no process, only a collection of individual habits.
// the honest test Pull stage-to-stage conversion for your last four quarters. If you can produce those numbers and you believe them, you have a process worth optimizing. If you cannot, because the data does not exist or cannot be trusted, you are rebuilding first. Most founders discover which camp they are in about ninety seconds into trying. |
Rebuilding means defining the stages, documenting what good looks like at each, wiring it into the CRM, and enforcing it, before any fine-tuning is worth attempting. Everything below assumes you have cleared that bar.
Why sales processes leak revenue silently
A leaking sales process does not announce itself. Revenue still comes in. Deals still close. The leak shows up as something vaguer: a win rate that drifts down, a sales cycle that creeps longer, forecasts that miss in ways nobody predicted. Because no single stage is measured, the loss is spread thin across the whole funnel and never attributed to a cause.
This is the difference between a process that exists and a process you can see. Most founders have the former. Deals move, reps work, the CRM fills up with activity. But ask three questions most cannot answer: what percentage of discovery calls become qualified opportunities, where in the funnel do deals stall the longest, and which stage has the worst conversion. Without those answers, every fix is a guess.
The silent leak is expensive precisely because it is silent. A process losing ten points of conversion at one stage can cost more than an entire underperforming channel, and it never shows up as a line item. It just looks like sales being hard.
You cannot fix a stage you do not measure. Most B2B revenue leaks are not a lead problem or a rep problem. They are a measurement problem wearing a sales costume. |
There is also a headwind underneath all of this that did not exist a few years ago. In Salesforce’s seventh-edition State of Sales report, drawn from more than 4,000 sales professionals, 57% say the sales cycle is getting longer. Longer cycles mean more stages where a deal can quietly stall, and more time for a buying committee to lose momentum. Slippage is no longer an anomaly to explain after the fact. It is a structural condition to design the process around.
The seven-stage diagnostic
Before you can optimize anything, you need a map. Most B2B sales, though not all, move through the same seven stages. Product-led motions, procurement-led public sector deals, and channel sales run differently, but for a founder-led services or SaaS business the sequence below will be recognisable.
Naming the stages and measuring the handoff between each is the whole diagnostic. Note what the table does not contain: target conversion rates. That omission is deliberate, and the next section explains why
// Table 01 · The seven stages and what to measure at each
Stage | What has to be true to advance | What to measure at this handoff |
1. Lead | Fits the ICP and has shown intent | Share of leads accepted by sales as worth a call |
2. Discovery | Problem, budget, and timing confirmed | Discovery calls that become qualified opportunities |
3. Demo / proposal-fit | Solution mapped to their specific problem | Demos that produce a requested proposal |
4. Proposal | Scope and price agreed in principle | Proposals that reach active negotiation, and days spent waiting |
5. Negotiation | Terms, procurement, and stakeholders aligned | Deals reaching a decision, won or lost, versus going quiet |
6. Close | Signed. Won or lost, recorded honestly | Win rate on qualified opportunities, with no-decisions counted as losses |
7. Onboard | Customer live and seeing first value | Time to first value, and whether onboarding friction shows up in renewals |
Stage seven earns its place even though it sits after the money. Onboarding is where deal quality gets audited: a customer who struggles to reach first value is usually a deal that was qualified loosely at stage two, and that shows up later as churn rather than as a sales problem.
To run the diagnostic, pull the last four quarters of closed deals, assign each to the stage where it entered and the stage where it actually died, and calculate the conversion between every step. The stage with the steepest drop is where your revenue is leaking. Most founders are surprised by which one it is.
Two details make or break this exercise. First, assign each dead deal to where it truly died, not where the CRM last recorded it. A deal marked closed-lost at proposal that never had a confirmed budget died at discovery. Second, four quarters is the minimum. Anything less and seasonality and small numbers will hand you a false signal.
Why there is no benchmark to compare yourself against
The obvious next question is what good looks like. It is worth being direct: there is no external answer that can tell you anything true about your process, and this is not a gap in the available research. It is structural.
A stage conversion rate is a ratio between two definitions. “Qualified opportunity” means one thing at a firm that requires a confirmed budget holder on the call, and something entirely different at a firm that counts any booked demo. Those two companies can run identical sales motions and report conversion rates that differ by a factor of five, purely because they drew the stage boundary in different places. Neither number is wrong. Neither one transfers.
Add the second variable and it compounds: conversion moves with deal size and segment, so even a company using your exact stage definitions will not produce a comparable figure unless it also sells at your price point to your buyer. Any published benchmark is an average across companies that disagree on both dimensions.
This matters more than it sounds. If you anchor on a borrowed number, you will conclude your discovery stage is healthy when it is not, or panic about a stage that is performing fine for your deal size and segment. Either way you optimize the wrong thing for a quarter, and you find out two quarters later.
Your baseline is your own last four quarters. The only benchmark that can tell you something true about your process is the one your process produced. |
The practical version: run the diagnostic twice before you act on it. One quarter of data tells you a number. Two consecutive quarters tell you a direction, and direction is what you can actually manage.
Where deals actually die
Ask a founder where deals are lost and most will say the close, the moment the prospect went quiet or chose a competitor. The data almost never agrees. The close is where the loss becomes visible, not where it began. The deal that dies at proposal was usually never qualified properly at discovery. The terminal event and the root cause sit at opposite ends of the funnel.
Two leaks are consistent across B2B services and SaaS firms.
The first is the discovery-to-qualified gap: reps run discovery calls that feel productive but never confirm budget, authority, or timing, so the pipeline fills with deals that were never real.
The second is the late-stage stall, where deals do not get rejected, they get quietly deprioritized. Founders read this as a price objection. It usually is not. In a survey of 632 B2B buyers conducted in 2024, Gartner found that 74% of buying teams demonstrate unhealthy conflict during the decision process, meaning members hold conflicting objectives, disagree on the right course of action, or get overruled from outside the group. The same research puts buying groups at five to 16 people spanning as many as four functions, and finds that groups reaching consensus are 2.5 times more likely to report a high-quality deal.
Read those numbers together and the late-stage stall stops being mysterious. Your deal is not sitting still because your price is wrong. It is sitting still because five to sixteen people who disagree with each other cannot reach consensus, and nobody inside the account is doing the work of getting them there. If your process has no mechanism for surfacing that conflict and helping a champion resolve it, you are relying on the buying committee to organise itself. Most do not.
This is why optimizing the close stage in isolation rarely works. By the time a deal reaches close, its fate was mostly decided three stages earlier. The leverage is upstream.
The three highest-leverage fixes
Not all stages are worth equal attention. Three carry most of the outcome.
01 | Fix discovery first. Discovery is where deal quality is set. A rigorous discovery confirms three things before anything advances: the prospect has a real, urgent problem, they have budget authority or a path to it, and there is a timeline that makes this quarter realistic. Reps who skip this in the rush to demo fill the pipeline with deals that feel alive and convert at a fraction of the rate. Tightening discovery is almost always the single highest-leverage change, because it improves the win rate feeding every stage downstream. Given what the Gartner data shows about buying groups, discovery should also establish who else has to agree. A discovery call that identifies the problem but not the other four to fifteen people who get a vote has done half the job. |
02 | Make the demo about them. The most common demo failure is a feature tour. A high-converting demo is not a walkthrough of your product; it is a narrow demonstration of the two or three things that solve the specific problem discovery uncovered. When the demo mirrors the prospect’s own words back to them, conversion to proposal climbs. When it is the same canned sequence every prospect sees, it signals that nobody was listening. |
03 | Engineer the close, do not hope for it. A clean close is built, not wished into being. It means mutual action plans that name every step to signature, early identification of the economic buyer and procurement, and a follow-up cadence that does not depend on a rep’s memory. Most late-stage stalls are momentum failures, not price objections. Removing friction from the path to signature recovers deals that would otherwise drift into no-decision. |
// find your leak Not sure which of the three is costing you most? The mrktrs diagnostic maps where your pipeline actually breaks and what to fix first. One question, one recommendation, sixty seconds. No sales call required. |
Worked example: where the leak actually was
// illustrative examples, not client case studies Northpoint and Vantage are composite scenarios built from patterns we see repeatedly across B2B firms at this stage, not named mrktrs clients. The figures illustrate how the diagnostic reads in practice and are realistic for this kind of fix, not verified results. |
Two firms, same revenue target, same apparent problem: not enough closed deals. The diagnostic sends them in opposite directions.
01 | The firm that blamed the close. Northpoint, a B2B services firm, was convinced its reps could not close. Running the seven-stage diagnostic told a different story. Their demo-to-proposal and proposal-to-close rates were steady across four quarters. But discovery-to-qualified had fallen to roughly half its level from the two quarters before, when they started booking demos off any expression of interest. The internal comparison was the tell: healthy later stages sitting behind a collapsed early one. The fix was a hard qualification checklist before any demo. Demo count fell, qualified-opportunity conversion recovered toward its earlier level, and the same team closed more revenue from fewer deals. |
02 | The firm that blamed the leads. Vantage, a SaaS company, was about to spend heavily to fix what it called a lead-quality problem. Discovery and demo conversion were steady quarter on quarter. The steep drop was between proposal and close, where time-in-stage had more than doubled against the prior year and deals quietly slipped. The cause was not bad leads. Deals were reaching proposal with a single contact and no visibility into the rest of the buying group. Building a standard close sequence, naming the economic buyer and procurement before a proposal went out, and giving the champion something to circulate internally recovered deals that had been dying in silence. The planned acquisition spend was never needed. |
Same symptom, opposite root cause, and in both cases the instinct was wrong. Neither could have known that without measuring every stage against its own history.
CRM hygiene is the instrument panel
None of this works if the CRM does not reflect reality. CRM hygiene is not a software project; it is the discipline that makes the process visible. If deals sit in the same stage for weeks with no activity, if stage definitions are fuzzy enough that two reps file the same deal differently, or if close dates are aspirational rather than evidenced, the pipeline report is fiction and every optimization built on it is too.
Three things make a CRM usable for optimization. Stage definitions that describe what must be true to advance, not just a label. Required fields at each stage so a deal cannot move forward without the evidence that it should. And a single, enforced definition of what counts as an opportunity, so the top of the funnel is not inflated with leads that were never real.
There is a capacity argument here too, not just a data one. Salesforce’s research puts reps at 60% of their time on non-selling tasks. Adding CRM requirements without removing anything makes that worse and guarantees the discipline decays within a quarter. Required fields should replace existing admin, not stack on top of it. If the hygiene work is not paid for out of existing non-selling time, it will not survive contact with a busy quarter.
A process you cannot see in the pipeline report is a process you cannot optimize. |
The metrics that matter, and the ones to ignore
Sales generates an enormous amount of data, and most of it is noise.
The four that matter: stage-to-stage conversion (where the leak is), sales-cycle length (how fast deals move, and where they stall), win rate on qualified opportunities with no-decisions counted as losses (the health of the whole process), and slipped deals (the count that pushed to a later period, the clearest early warning of a forecast miss).
The four to ignore, or at least never optimize toward: raw activity counts, which measure effort rather than outcome; total lead volume, because more bad leads is not progress; individual rep leaderboards in isolation, which hide process problems behind personalities; and any dashboard metric that cannot be tied to a stage or a dollar.
Given that cycles are lengthening for most teams, sales-cycle length deserves more attention than it usually gets. Track it per stage, not just end to end. An overall cycle stretching from 60 to 90 days tells you something is wrong. Knowing that all thirty of those days accumulated between proposal and close tells you what to fix.
A sales process that is defined, measured against its own history, and tuned at its three highest-leverage stages does not just close more deals. It makes revenue predictable, which is the thing every founder past the referral plateau is actually chasing. The leak was never the lead volume. It was the quiet gap between stages that nobody was measuring.
// next step We run this diagnostic for a living. If you want the seven-stage map built against your own four quarters, and a ranked list of what to fix first, that is what we do. mrktrs owns growth for B2B firms between $1M and $20M ARR. |
Frequently asked questions
What is the ROI of sales process work?
It lifts revenue without buying more leads, so a few points of conversion recovered at discovery or close drops straight to the bottom line on pipeline you already have. Unlike acquisition spend, the gain compounds: a tighter process makes every future lead more valuable, not just this quarter’s.
Can this fix bad leads?
Partly, and the distinction matters. A rigorous discovery stage is a filter that stops poorly qualified leads inflating the pipeline and dragging down win rate. It will not turn the wrong audience into buyers, which is a targeting and positioning problem upstream. But it will stop bad leads consuming sales capacity and distorting your numbers, which is often mistaken for a lead-quality problem when it is really a qualification one.
Should we adopt MEDDIC, BANT, or another qualification framework?
Any of them beats no framework, and the choice matters far less than the enforcement. A framework is only a structured way to make discovery confirm problem, authority, budget, and timing before a deal advances. If your CRM does not require the evidence at the stage gate, adopting a named methodology changes the vocabulary and nothing else.
Which stage should we fix first if we can only fix one?
Discovery, unless your own data contradicts that. It sets the quality of everything downstream, so a gain there compounds through five subsequent stages, while a gain at close applies only to deals that already survived. The exception is when the diagnostic shows healthy early conversion and a severe late-stage stall, which points at a missing close motion rather than a qualification problem.
How does this differ from pipeline coverage?
Coverage asks whether you have enough qualified pipeline to hit the number. This asks whether the pipeline you have converts as it should, stage by stage. They fail differently: weak coverage means you miss because there was never enough at the top, while a leaking process means you miss because deals died in a gap you were not measuring. Coverage is a volume question, process is a conversion question, and a healthy quarter needs both.