Strategy and positioning · Answered by Golden Section from more than 400 B2B software companies observed
Not by size alone. Stop serving the customers who cost more to win and to serve than they pay, and size is only one predictor of that. Score every customer on contract value, expansion, churn, acquisition cost and cycle length, support load and custom work, then see where the bottom of the ranking clusters. Small accounts that buy quickly, set themselves up and renew are good business. Small accounts that expect enterprise service sit in what the customer segmentation play calls death valley. In our experience about a fifth of any cohort is a poor fit; for those, raise the price to what serving them costs or help them leave in an orderly way. Before dropping a whole segment, check that the ones you keep still cover a quarter of your market. Next step: rank customers on a written quality score.
Keep small customers who are cheap to win and serve; reprice, redesign the service model for, or offboard those whose cost exceeds what they pay. Decide on a written quality score, not on contract size.
| Metric | Value | What it means | Source |
|---|---|---|---|
| Poor-fit share of a cohort | about 20% | customers whose cost to serve or expectations do not fit the productraise their price or let them go | Golden Section operating viewGolden Section operating view |
| Sales cycle guide | about 2 weeks per $20K of ACV | expected cycle length scaled to contract valueMistake 139; a small contract has to be won quickly or it costs more to sell than it earns | Golden Section, publishedThe B2B Software Mistakes List |
| Minimum market coverage | at least 25% of SAM | sum of the segments you keep, as a share of serviceable marketcheck it before dropping a segment; below 25%, reconsider | Golden Section playbookCustomer Segmentation |
Small customers are expensive in specific ways. Support tickets, onboarding hours and requests scale with the number of customers, not with what they pay, and a small contract still needs a sales cycle. Mistake 139 on our list puts the guide at about two weeks per $20K of contract value, so a $5K customer has to close in days to be worth winning. When small accounts also want custom work, they land in the customer segmentation play's death valley: high customization at low contract value, revenue not worth pursuing.
But size is a proxy, not the test. The same play scores customer quality on several factors, and plenty of small customers come out well because they need little and stay for years. The decision belongs to the score, worked through the unit economics of each group.
When the answer is to let some go, do it on purpose. The account management process includes offboarding as its fourth stage, because an orderly exit protects your reputation and a disorderly one becomes a public argument. A repriced plan with a lighter service model is often the first offer. Customers who will not pay what they cost are better served elsewhere, and so are you.
All numbers are invented. A company at $3M in annual revenue has 400 customers. The 180 with contracts under $3K bring in about $360K, 12% of revenue, and generate 45% of support tickets. A quality score puts 110 of them in the bottom quartile. The company creates a self-serve plan at $4,200 with email support and recorded onboarding, and offers it at each renewal. Over two renewal cycles 60 accept and 50 leave, taking about $90K of revenue. Support tickets fall by a quarter, and one support role moves to onboarding customers in the core multi-site segment.
If small customers are your core segment, the fix is a cheaper way to serve them, not a different customer. In some verticals today's single-site customer becomes a multi-site group; keep the ones with a credible path to that.
From the Golden Section mistakes list, each paired with the play that prevents it.
Holding every customer at any cost lets misaligned accounts shape the product and the support load.
A vocal minority of small customers can consume a team; focus belongs with the accounts that operate as partners.
A small contract with a long sales cycle loses money before the customer ever goes live.
In the order we would run them. Each is on its own page, most with a free Excel template.
Scores customer quality so the decision rests on evidence rather than contract size.
Works out what each customer group earns after the cost to win and serve it.
Builds a plan priced at what serving the smaller customer actually costs.
Runs offboarding as a defined stage, so customers who leave do so cleanly.
Customer plays Which customers to keep, reprice or offboard is decided and executed through the customer plays.
Decide on economics and behavior, not revenue alone. Candidates are customers who cost more to serve than they pay and will not accept a price that covers it, customers whose requests pull the product away from your core segment, and customers who will not work as partners. Offer a repriced plan first, then offboard in an orderly way so the exit does not cost you references.
It raises logo churn for a period and may dent net retention, so report the departures as a deliberate decision with their own line. Done well, gross margin and support cost improve, and the retention of the customers you keep is the number a buyer cares about.
Reviewed by Dougal Cameron, CEO & Co-Founder on 2026-10-22. Golden Section observations are labeled separately from external benchmarks and illustrative arithmetic.