Diagnostics · Answered by Golden Section from more than 400 B2B software companies observed
Treat the retention as proof that the product works and look for the break in acquisition, which is usually one of three things. The segment you win in may be saturated or you may have drifted away from it; the pipeline may have depended on the founder, who stopped selling; or the team may be harvesting expansion because it is easier and pays the same. Do not respond by buying more leads. Split bookings into new-logo and expansion for eight quarters, map where new-logo pipeline is created and where it dies, and read the comp plan for what it actually rewards. At 110% the company still grows about 10% a year with no new customers, which is healthy but capped. Next step: rebuild the funnel for new logos alone, stage by stage.
Strong NRR earns the right to invest in acquisition, but only after you know which stage of the new-logo funnel broke. Fix the stage, then fund the channel.
| Metric | Value | What it means | Source |
|---|---|---|---|
| Growth from the base alone | about 10% a year | 110% NRR with zero new logos adds 10% of starting ARRSimple arithmetic; the cap on a business that has stopped acquiring | IllustrativeIllustrative arithmetic |
| Net revenue retention | 110%+ is a finding | 12-month dollar retention including expansionGS's 2026 re-test: underwrite 100%, treat 110%+ as a finding; the private median is 101% | Golden Section, publishedInvesting in Software addendum |
| Expansion share of new ARR | 40% at the median | expansion ARR ÷ total net new ARRRises to 67% above $100M ARR; the top line increasingly rests on the installed base | Golden Section, publishedInvesting in Software addendum |
| Segment coverage | at least 25% of SAM | sum of chosen segments as a share of serviceable addressable marketRule of thumb for whether target segments leave room to grow | Golden Section playbookCustomer Segmentation play |
High retention and stalled acquisition come from different machines, so the fix for one tells you nothing about the other. Retention says the customers you have got the value. A stall says the process that finds the next customer has broken, and that process has fewer places to hide than founders think.
Start with the segment. If the best customers share traits you can name, the customer segmentation work shows whether that group still has room or whether you have sold most of it. Then walk the sales funnel for new logos only; a stall at the top means pipeline creation stopped, often when the founder stepped back from selling, and a stall lower down means deals are arriving but dying. Finally read the sales compensation plan. A rep paid the same for expansion and for a new logo will rationally spend the quarter on the account that already likes you.
The risk of waiting is quiet. Expansion from a fixed base eventually slows, and a buyer pricing the company sees a business that stopped winning new customers.
A company at $5M ARR with 110% NRR has added four new logos in a year, against 20 the year before. The split shows the funnel is not empty: qualified opportunities are flat, but the close rate on them has halved. The comp plan pays the same rate on expansion as on new logos, and expansion now makes up most of each rep's attainment. The segmentation review also shows the company's best segment is two-thirds penetrated. The founder separates new-logo quota, adds a higher new-logo rate, and assigns one rep to an adjacent segment the customer data already favors. All figures are invented for illustration.
A company whose market is small by design and whose founder wants a profitable, slowly growing business may accept 110% NRR as the growth engine. That is a legitimate choice, but make it on purpose and tell the board rather than drifting into it.
From the Golden Section mistakes list, each paired with the play that prevents it.
The instinctive response to a stall is to try five new channels at once, which spreads force exactly when it needs concentrating.
Founders abandon a new-logo channel before it has had enough repetitions to show whether it works.
When the founder stops sourcing deals, inbound and partner-sourced opportunities that nobody owns are often the first pipeline to go.
In the order we would run them. Each is on its own page, most with a free Excel template.
Shows whether the segment you win in still has room or has been sold through.
Locates the stage where new-logo deals stopped converting.
Rebuilds the model of how many leads the new-logo target needs and what each one costs.
Makes the comp plan reward the new-logo work the company actually needs.
Measures whether new-logo spend is earning its return once the motion restarts.
Sales and marketing plays Segmentation, funnel, pipeline and comp all sit here, and a new-logo stall usually lives in one of them.
It produces roughly 10% annual growth from the base, which is healthy and still below the 25–40% range we treat as great for capital-efficient vertical SaaS. It also concentrates growth in existing accounts, which eventually saturate.
Not until you know which funnel stage broke. More leads help only if the stall is at the top; if deals are dying after qualification, more leads just produce more lost deals.
110% NRR is exactly the kind of durable revenue our lending is underwritten on. Once a new-logo channel is converting again, non-dilutive capital can fund more of it; before that, the fix is operational.
Growth capital lending →Reviewed by Dougal Cameron, CEO & Co-Founder on 2026-09-23. Golden Section observations are labeled separately from external benchmarks and illustrative arithmetic.