A CEO showed me a board slide in January that said win rate 41%. Two weeks later his VP of Sales sent the same board a slide that said 22%. Nobody had lied. They had pulled two different reports out of the same HubSpot instance.
The 41% came from a report filtered to deals that reached the proposal stage. The 22% came from every opportunity created in the quarter. Both were sitting in the same CRM, both were labelled "win rate," and the two numbers were about to drive completely opposite decisions. The CEO was ready to hire two more AEs on the strength of 41%. The VP was ready to cut the SDR budget on the strength of 22%.
This is the most common broken metric I run into. Not pipeline coverage, not forecast accuracy. Win rate. It looks like the simplest number in the business, which is exactly why nobody audits it.
Here is how to calculate it so the number means something, what the real 2026 benchmarks are, and what to do when yours is bad.
What win rate actually measures
Win rate is the share of sales opportunities that end in a closed-won deal. The formula everyone quotes:
Win rate = Deals won / Total deals × 100
That formula is fine. The trouble is the denominator. "Total deals" can mean at least five different things, and each one produces a different number for the same team in the same quarter.
Win rate is not one number. It is a family of numbers, and most teams report whichever one flatters them.
Before you compare your win rate to a benchmark, to last quarter, or to another rep, confirm both sides use the same denominator and the same definition of a qualified opportunity. Almost nobody does this.
The five denominators
All opportunities created in a period. Includes deals that are still open. Gives you the lowest number and the least useful one in-quarter, because young deals have not had time to close.
All opportunities closed in a period. Wins divided by wins plus losses, where both closed in the same window. This is what most CRM reports default to. It is the standard and it is defensible, but it mixes deals created across many different quarters, so it lags your actual pipeline quality by a full sales cycle or more.
Qualified opportunities only. Strips out deals that never should have entered pipeline. Raises the number substantially. Across the datasets I have seen, roughly 21% average becomes about 29% when you filter to qualified only.
Proposal-stage or later. Only counts deals where you put a price on the table. This is the vanity version. It can read close to 47% for a team whose real number is in the low twenties.
Created-cohort, matured. Take every opportunity created in Q1, wait until Q4, then see how many of them closed won. This is the only version that tells you whether the pipeline you built was any good.
I run two of these and ignore the rest. Closed-period win rate for trend, cohort win rate for diagnosis. If you only have room for one on the board deck, use closed-period and put the sample size next to it.
The no-decision problem nobody wants to solve
Here is the one that costs the most money.
Most B2B teams do not close deals that stall. They leave them open. A deal goes quiet in March, the rep keeps pushing it to next month, and by September it is still sitting in Negotiation with a close date that has moved eleven times. It was dead in April. Everybody knows it. Nobody wants to mark it lost because losses show up on the scoreboard and stalls do not.
That behaviour inflates your reported win rate by 10 to 15 percentage points, because those dead deals never make it into the denominator.
How much your reported win rate rises when stalled deals are left open instead of being closed as no-decision. This is the single biggest source of fake win rate in B2B CRMs.
The fix is unglamorous and takes about an hour. Create a closed-lost reason called "no decision." Write a rule: any deal with no logged buyer activity for 45 days past its expected close date gets auto-closed as no decision. Build it as a HubSpot workflow or an n8n job that runs weekly. Tell the reps it is not a mark against them and it does not affect comp on activity.
Then watch what happens to the number. Every team I have done this with sees the same thing. Win rate drops between 8 and 14 points in the first month, everyone panics, and then the forecast starts being right for the first time in a year.
One client, a Series A team in Berlin selling into mid-market logistics, went from a reported 34% to a real 21% the week we turned this on. The CEO's first reaction was that we had broken the CRM. His second reaction, about six weeks later, was that he had been hiring reps against a fantasy number and had been running 40% over on headcount plan for three quarters.
What a good win rate actually looks like in 2026
Benchmarks are dangerous, because as I just showed, nobody is measuring the same thing. But directionally, here is where the market sits.
The Ebsta and Pavilion GTM benchmark work, which covers hundreds of thousands of opportunities and tens of billions in pipeline, put average B2B win rates around 19% in the most recent cut, down from 29% the year before. That drop is real and it is market-wide. Buying committees got bigger, budgets got scrutinised harder, and cycles stretched.
The other consistent finding: 78% of B2B sellers missed quota, up from 69% the year prior. If your win rate is down and your team is missing plan, you are not uniquely bad. You are average. That is not a comfort, but it should change what you conclude from the number.
Win rate scales inversely with deal size, and this is the pattern that matters most for planning:
If you are moving upmarket, your win rate is supposed to fall. A team that goes from $30K deals to $150K deals and holds win rate flat has almost certainly changed how it counts, not how it sells. I have seen this misread as success on more than one board deck.
Source also matters more than most teams admit. Inbound demo requests convert at multiples of cold outbound. Referral and partner-sourced deals convert higher than either. If your blended win rate moved and you did not segment by source, you have learned nothing. You may just have changed your channel mix.
The segments that make win rate useful
A blended company-level win rate is a pulse check. It tells you whether the patient is alive. It tells you nothing about what to do.
The value is in the cuts. These are the five I build for every client, in this order:
The stage-exit cut is the one people skip and it is the one that pays. A 20% win rate can be built two completely different ways. Either you qualify loosely and lose most deals early, which is cheap, or you qualify well and lose deals at proposal, which is expensive because you burned a solutions engineer and three weeks of AE time on every one.
Same headline number, opposite problem, opposite fix. If most of your losses happen after proposal, you have a value and pricing problem, not a targeting problem. If they happen at discovery, you have a targeting problem and your ICP definition needs work before you touch the pitch.
How to actually build this in your CRM
Assume HubSpot, because that is what most of the teams I work with run. The same logic maps to Salesforce or Attio with different field names.
Two implementation details that matter more than they look.
First, lock the source field. If reps can edit it, they will, usually after the deal closes and usually toward whatever source the comp plan rewards. Set it on creation, make it read-only after, and let RevOps override with an audit trail.
Second, show the sample size. A rep with a 60% win rate on five deals has told you nothing. On thirty, they have told you something. I put the count in parentheses next to every win rate on every dashboard, and it has stopped more bad decisions than any other single formatting choice I have made. If a cell has fewer than 20 closed deals behind it, grey it out.
If you want the full build pattern for reporting that executives actually read, I wrote that up separately in the sales dashboard guide.
What actually moves win rate
Once the number is honest, the interventions that work are boring and they are mostly upstream of the sales conversation.
Fix targeting before you fix pitching
The single biggest win rate lever is not talking to accounts you were never going to win. Most teams' win rate problem is a pipeline quality problem wearing a sales skills costume.
Run this: take the last 12 months of closed deals, split by fit criteria (headcount, industry, tech stack, funding stage), and find the cells where you win above 30% and the cells where you win below 10%. The below-10% cells are usually 30 to 40% of your pipeline by count. Stop working them for a quarter and watch what happens to the blended number.
Multithread earlier, not later
Deals with three or more engaged contacts on the buyer side close at meaningfully higher rates than single-threaded ones. This is the most reliable pattern in the data I have looked at across client CRMs. The catch is that reps add contacts late, usually after the deal is already in trouble, which does not help.
Make contact count at each stage a gate, not a report. A deal cannot move to proposal with one contact on it. I go deeper on the mechanics in the multithreading guide.
Qualify out loudly
The fastest way to raise win rate is to have fewer bad deals in the denominator, and the only sustainable way to do that is to make qualifying out a celebrated act rather than a failure. If your pipeline reviews only ask "how do we win this," reps will never disqualify anything. Add a standing question: which two deals are we killing this week?
That single change moved one client from 19% to 27% over two quarters without any change to the sales pitch, headcount, or product. They just stopped spending time on deals that were never going to close and put those hours into the ones that were.
Do not chase win rate at the expense of volume
Worth saying plainly, because the metric can be gamed. A team that only works its best-fit accounts will post a beautiful win rate and miss its number, because it left revenue on the table in the medium-fit segment. Win rate is a ratio, and ratios can be improved by shrinking the denominator in ways that hurt you.
Watch win rate alongside total won revenue and pipeline coverage. If win rate goes up and won revenue goes down, you over-corrected.
Not sure your win rate is real?
We audit the CRM, rebuild the reporting, and show you the number your board should actually be looking at. Usually takes a week.
Book an audit →FAQ
What is a good sales win rate?
For B2B SaaS in 2026, roughly 20 to 30% measured against all closed opportunities is normal, and the market average sits closer to 19 to 21% after the last two years of compression. Under $50K ACV you should be at 25 to 35%. Above $250K, 12 to 22% is healthy. Any benchmark you read is worthless unless it states its denominator, so treat public numbers as a rough sanity check and benchmark mainly against your own prior cohorts.
Should I include no-decision deals in my win rate?
Yes. Excluding them inflates the number by 10 to 15 points and hides the most fixable problem you have. A buyer who stalls out is a loss with extra steps, and it usually points at a qualification or business-case failure that a competitive loss does not. Track them as a separate closed-lost reason so you can see the split, but keep them in the denominator.
Win rate vs close rate vs conversion rate: what is the difference?
Win rate is usually opportunity-level: of the deals that reached a decision, how many did you win. Close rate is often used the same way, though some teams mean leads to customers by it. Conversion rate normally refers to a single stage-to-stage step, like MQL to SQL. The terms get used interchangeably and that is exactly why your CEO and your VP of Sales end up with different numbers. Write the definitions down once and put them in the CRM field descriptions.
How often should I review win rate?
Monthly for the trend line, quarterly for the segment cuts. Anything more frequent is noise unless you close hundreds of deals a month. Most SMB and Series A teams do not have the volume for weekly win rate to mean anything, and reacting to weekly swings is how you end up rewriting the pitch every fortnight.
How do I improve a win rate that is stuck in the low teens?
Look at where deals die before you look at how reps sell. If most losses happen at or before discovery, you have a targeting problem and should tighten ICP and lead qualification. If they happen after proposal, you have a value or pricing problem and should work on business case and multithreading. Adding more pipeline to a low win rate just scales the waste, which is the default response and almost always the wrong one.
Where to go from here
Win rate is the cheapest diagnostic in the business, and it is broken in about four out of five CRMs I open. Getting it honest costs a few hours of workflow config. Acting on it, which means killing the accounts you were never going to win, is the harder part and the part that pays.
If you want help auditing the number and rebuilding the reporting around it, that is a lot of what we do in CRM and RevOps engagements. If the fix turns out to be automation heavy, closing zombie deals, syncing sources, enforcing stage gates, that lives in AI and automation work. And if the answer is that your targeting is the real problem, that is a go-to-market conversation.
Either way, get in touch and we will look at your actual numbers rather than the ones on the slide.