Insights · Pipeline
Why pipeline coverage benchmarks no longer hold, how to calculate what you actually need from your win rate, and what a pipeline creation system looks like when it is designed rather than accumulated.
Umar Gill, F.ISP
Managing Partner, Compose Sales Partners
Last reviewed August 2026 · Next review February 2027
The short version
✓
Pipeline coverage is not a benchmark, it is a calculation. Required coverage equals one divided by your win rate.
✓
3x assumes you win a third of what you qualify. Almost nobody in enterprise technology operates there now.
✓
The share of AEs at quota has fallen from 74 percent in 2012 to 48 percent in 2026. That is a decade-long trend, not a bad year.
✓
Adding volume makes it worse. Unqualified pipeline eats the scarcest thing you have, which is qualified selling hours.
✓
A library of plays is not a system. Plays need a taxonomy, execution standards, operationalisation and traceability.
✓
A sales methodology will not fix a pipeline creation problem. It is a conversion instrument answering a demand question.
Pipeline coverage is not a benchmark. It is a calculation, and the arithmetic moved while the rule of thumb stayed put.
Every pipeline conversation I have joined in the last three years has followed the same shape. The number is at risk. Coverage is reported. Coverage looks acceptable. The quarter still misses. Then somebody says generate more pipeline, and the machine starts up. More outbound, more campaigns, a partner push, an event brought forward.
What has struck me over 15 years is how little the geography matters. The same conversation happens in London, in Frankfurt, in Boston, in Singapore. Same underlying problem, and it has sharpened every year I have been doing this.
The uncomfortable part is that the coverage number was never acceptable in the first place. It was measured against a rule of thumb that stopped being true a decade ago.
Umar Gill, F.ISP
Managing Partner, Compose Sales Partners
Single-year benchmark swings are mostly noise. The long series is not. The Bridge Group has run the same biennial study of the Account Executive role since 2006, now in its tenth edition, most recently across 158 B2B companies. Same firm, same method, ten rounds. That makes it the most reliable series available on this question, and it is worth reading four of its lines together.
Share of AEs at quota
74% › 48%74% in 2012, 66% in 2022, 51% in 2024, 48% in 2026.
Time to full productivity
6.2 mthsUp from 5.7 months in 2023. The highest in the study’s history.
Experience required at hire
3.7 yrsUp from 2.7 years in 2022. Companies buying in experience to compensate.
Quota to OTE ratio
4.6xUp from 4.2x in 2024. Quotas rising against earnings.
Quotas rising against earnings. Ramp lengthening. Companies buying in more experience to compensate for it. And a smaller share of people clearing the bar each time the study runs. None of that is a coaching problem.
Pipeline coverage is not a benchmark, it is a calculation. Required coverage equals one divided by your win rate. Three times coverage assumes you win a third of what you qualify, which is where the rule came from, and almost nobody in enterprise technology operates there now.
Win rate benchmarks currently sit well below that, though they move around depending on how the denominator is defined. Winning by Design flagged enterprise win rates falling into the 17 to 20 percent range back in 2023. Ebsta and Pavilion reported an average of 19 percent in their 2025 GTM Benchmarks, alongside 78 percent of sellers missing quota.
One caution on that source, and it matters more than the figure. Ebsta and Pavilion reported 29 percent the year before. A ten point single-year swing in a market-wide win rate is almost certainly a change in what was counted rather than a change in the market. Use the direction. Do not put a ten point market swing in front of a board.
Worked example
A team carrying a £10m quarterly new ACV quota at a 19 percent win rate on qualified opportunities.
Pipeline actually required
£52.6m£10m ÷ 0.19
Pipeline reported
£35mWhat the board sees
Coverage against the 3x rule
3.5xWhich looks healthy
Shortfall against the real requirement
£17.6mInvisible until week ten
That team is £17.6m short and reporting green. It will find the gap in week ten of the quarter, roughly nine weeks after anything could have been done about it.
A sales play is not a deck, and a library of plays is not a system. A pipeline creation system has a play taxonomy, three content layers, execution standards, and traceability from first touch to closed.
Two types, and confusing them is why sales play libraries bloat until nobody opens them.
Anchor plays
Twelve month life
Tied to your ICP and core value proposition. In the businesses I have seen, many forget that their core propositions need a play too. Those who do this properly always align their main plays around the core value propositions. To sustain this well, it is usually just three to five plays, no more. They carry most of the sourced pipeline and justify real investment in content, training and coaching.
Tactical plays
Expiry date set at launch
Competitive displacement, product launch, event follow-through, install base expansion, regional plays, a sales campaign or something needed for a particular customer type. These tactical plays are the ones that make sales plays relevant to the business every day. Time-boxed at six to twelve weeks with an expiry date set at launch. Tactical plays that never get retired are how you end up with the library nobody opens.
The bill of materials for a sales play has three layers, and the order matters because each one derives from the one above it.
Customer facing
What the buyer sees
Umbrella message, then ICP, then buying persona. Content is derived from this, not written alongside it. FocusVision research cited by Highspot puts buyers at roughly thirteen content pieces consumed before first contact with a seller. If your economic buyer and your technical evaluator get the same narrative across those thirteen touches, you have one layer rather than three.
Seller facing
Usually the thinnest part of the build
The internal kit: entry trigger, target account list, qualification bar, discovery questions, objection handling, defined next step. This is what sellers actually consume, and it is where most play builds run out of budget. A beautiful customer-facing layer with nothing underneath it is a campaign, not a play.
Operationalisation
What has to be in place before launch
The plumbing, and the difference between a play that launched and a play that was announced. Pricing and CPQ configured so the offer can actually be quoted. Campaign codes and source identifiers created so the pipeline can be traced. The marketing campaign linked to the play rather than running beside it. A named route to market, because plays without one default to the AE, which is the most expensive channel you own. Interlock agreed with product, marketing, partner and inside sales on who does what. And the field genuinely told, in the rhythm they already attend, rather than an email nobody opened.
Five execution standards. Without them you cannot tell a play that failed from a play that was never really run, which means every post-mortem turns into an argument about effort.
A named owner
An entry trigger
An activity standard
An exit definition
A sourced pipeline target
Every play carries a source identifier from first touch through to closed. If you cannot attribute sourced pipeline to a specific play then you cannot evidence which ones work, and if you cannot do that you will never retire the ones that do not. Sales operations owns this. It is not an analytics nicety, it is the mechanism that lets the system improve instead of just accumulate.
Play development, launch training, coaching to the execution standard, and deal reviews conducted against play criteria rather than a rep’s narrative. A play launched without a coaching plan is effectively dead within a quarter, however well it landed on the day.
A demand generation cadence reviews pipeline creation against target, by play and by route. It is not the forecast call, and running them as one is the most common structural error I see.
Run them as one meeting and creation loses to closing every week, because closing is louder and closer.
Forecast call
Looks backwards
In-quarter deals
Commit and upside
What closes, and when
Demand gen cadence
Looks forwards
Sourced pipeline vs target
By play, by route
Marketing and partner present
Win room
Looks sideways
Named must-win pursuits
Standing, cross-functional
Convened by significance
The win room should be standing from the start of the year, not assembled in week ten when somebody notices a must-win slipping. Convened by exception, it only ever convenes when it is already too late.
Sales, marketing and partner should not hold separate pipeline plans. One plan, one sourced pipeline target, split by route, each route with its own number and its own owner.
Separate plans are how three functions each hit their own number while the business misses.
Plays and pipeline · PACE
None of this assembles itself.
Pipeline creation is the first part of PACE, the delivery framework inside the COMPOSE Sales Model® that covers plays, academy, content and engagement. It is the part that turns a go-to-market strategy into opportunities that actually appear.
We build the whole system: the play taxonomy, the full bill of materials across all three layers, the execution standards, the traceability, and the demand generation cadence that keeps it running after launch.
Built with your sellers, tested on live deals, and handed over working.
Most clients start with one anchor play, built end to end and proven in live deals before anything scales.
The questions and objections that come up every time this argument is made in a room. Most of them are reasonable.
Required pipeline coverage equals one divided by your win rate on qualified opportunities. At a 19 percent win rate you need 5.3x, not 3x. At 25 percent you need 4x. Coverage is a calculation, not a benchmark, and using somebody else’s number means measuring yourself against their win rate rather than your own.
Only if you win a third of what you qualify, which is where the 3x rule came from and where almost nobody in enterprise technology now operates. Win rate benchmarks currently sit closer to 17 to 20 percent, which puts required coverage at five times or more. A team reporting 3.5x against a 19 percent win rate is substantially short while showing green.
Published enterprise B2B win rates currently sit in the 17 to 20 percent range, though figures move considerably depending on how the denominator is defined. Winning by Design flagged 17 to 20 percent in 2023, and Ebsta and Pavilion reported 19 percent in their 2025 GTM Benchmarks. Use the direction rather than any single number, and calculate your own from qualified opportunities.
Sometimes true and worth testing first. The test is your no-decision rate. If 40 to 60 percent of your losses are to no decision then you are in the normal range and the problem is qualification and business case, which is entirely yours. If most losses are to named competitors, that is positioning and product. Most organisations do not separate the two in loss reporting, which is why the argument never resolves.
This is the strongest objection on the list and there are markets where it holds. In genuinely high-velocity motions with short cycles and high win rates, coverage really is the binding constraint. The distinction is cycle length. Past roughly ninety days, with committees past a handful of people, the arithmetic inverts, because a wasted opportunity slot now costs you a quarter rather than a week.
Half right. Adoption is a management issue. But a framework captured as optional CRM fields has been designed as documentation rather than as a gate, and management pressure does not convert one into the other. If a deal can reach a late stage with those fields blank, the system permits it.
Usually accurate and usually beside the point. If marketing is measured on sourced volume and sales on closed revenue, the two are optimising against different definitions of a good opportunity, and both can hit target while the business misses. The fix is not better collaboration. It is one sourced pipeline definition, agreed qualification criteria at the handover, and one plan with one number split by route.
Somewhere between eight and fifteen at once for a ramped AE in enterprise technology, depending on cycle length and complexity. That ceiling is why unqualified pipeline is not neutral: every slot filled by something that was never real is a slot unavailable to something that was.
Every function can be individually correct and the system can still fail. That is what makes this a design problem rather than a performance problem.
Bertuzzi, M. (2026). AE Models, Motions & Metrics: 2026 Research Report (10th ed.). The Bridge Group. 158 B2B companies. Cited with permission. bridgegroupinc.com
Dixon, M. and McKenna, T. (2022). The JOLT Effect. DCM Insights. Analysis of 2.5 million recorded sales conversations. See also “Stop Losing Sales to Customer Indecision”, Harvard Business Review, June 2022.
Ebsta and Pavilion, 2025 GTM Benchmarks Report. Win rate figure flagged as methodologically unstable above.
Forrester, The State of Business Buying (January 2026).
Gartner B2B buying research.
6sense, 2025 Buyer Experience Report (n=4,510).
Winning by Design industry analysis (2023).
FocusVision research on buyer content consumption, cited by Highspot.
Browse more perspectives from our team of operators and advisors.
The best first step is a simple conversation about where your sales organisation stands, and what to fix first. No pressure, no pitch.
Compose Sales Partners
6th Floor
37 Lombard Street
London
EC3V 9BQ
Compose Sales Partners
20-22 Venture West
Greenham Business Park
Newbury
RG19 6HX
COMPOSE Sales Partners, the COMPOSE Sales Partners logo, COMPOSE Sales Model®, the COMPOSE Sales Model diagram and COMPOSE Sales Kick-Off Model are trademarks used by Compose Sales Partners Limited under licence. PERFORM, RISE and PACE are proprietary frameworks of Compose Sales Partners Limited. All other brands and product names are the trademarks of their respective owners.
