The problem is not the big customer. It is what the big customer does to everything else.
Most founders understand, abstractly, that relying on one customer for a large share of revenue is risky. What they underestimate is how the dependency corrupts decisions long before the customer ever threatens to leave. The product roadmap bends toward their requests. Pricing discipline erodes because you cannot afford to lose them. Contract terms get negotiated in their favor because the power asymmetry is real and both sides know it. By the time concentration shows up as a line item in a due diligence report, the damage to the business is already structural.
This article is about those mechanisms, not the generic advice to go get more customers.
What the thresholds actually mean by business model
The standard benchmarks are a useful starting point. In B2B SaaS, analysts flag a single customer at more than 10% of ARR as a warning sign, with risk becoming elevated when one customer exceeds 20-30% or when the top five customers together exceed 40-50% of total recurring revenue. A $10 million ARR company with one $2 million customer faces 20% concentration: that customer's churn immediately erases one-fifth of revenue and triggers downward pressure on valuation.
But the thresholds shift meaningfully depending on your model:
- B2B SaaS with annual contracts: The 10% single-customer threshold is the right starting point. The danger is compounded by renewal cliffs. If two or three large contracts co-terminate in the same quarter, a single bad renewal cycle can look like a business collapse even if churn is structurally normal.
- Professional services and agencies: Concentration tolerates less here because revenue is not recurring. A small business earning 60% of its revenue from one client faces existential risk if that client leaves. The lack of a contractual floor means exposure is immediate. A 20% single-client threshold is aggressive for this model; 15% is more defensible.
- Ecommerce and consumer goods: The risk compounds through channel concentration as well as customer concentration. If any single buyer represents more than 10% of total sales, or if the top five customers generate over 25% of revenue, the business is exposed. A consumer-goods company where two buyers account for 46% of revenue may look fine during 37% growth but becomes critically exposed when growth slows and those buyers gain renegotiation leverage.
- Marketplaces and platforms: Concentration on the supply side (a handful of top sellers or providers driving most GMV) is structurally identical to customer concentration on the demand side and deserves the same measurement discipline.
How founders optimize themselves into the trap
The psychology is rational at each individual step, which is what makes it dangerous.
A large enterprise contract arrives early. It funds the next 18 months of runway. The founder celebrates, reasonably. Then the customer asks for a feature that is not on the roadmap. The founder builds it, also reasonably, because the contract justifies the investment. The customer renews and expands. The customer asks for another feature. The founder builds it again. Two years later, the product has drifted significantly toward the needs of one customer segment, the sales team has been trained to sell to that segment, and the company's entire go-to-market motion assumes a buyer profile that only that customer reliably represents.
This is not a failure of discipline. It is the predictable output of incentive structures. The large customer provides immediate, concrete revenue. The diversification benefit of a second or third large customer is diffuse and probabilistic. Founders are not irrational to prioritize the certain over the uncertain. They are, however, systematically underweighting the option value of independence.
The compounding factor is that the big customer usually knows their leverage. They negotiate payment terms that stretch your DSO. They push for most-favored-nation pricing clauses that cap your ability to raise prices for anyone. They request exclusivity provisions disguised as partnership language. Each concession feels minor in isolation. Collectively, they formalize the dependency in the contract itself.
Contract terms that disguise dependency
Several contract structures function as concentration risk amplifiers that do not appear in revenue dashboards:
- Co-termination clauses: When a large customer bundles multiple products or subsidiaries into a single renewal date, you face a binary event. Either the whole relationship renews or none of it does. This converts what looks like a diversified book of business into a single concentrated risk event.
- Most-favored-nation (MFN) pricing: MFN clauses require you to offer the large customer any better price you give to anyone else. They effectively prevent you from moving upmarket or adjusting pricing strategy without triggering a retroactive discount for your most important account.
- Custom SLA commitments: Bespoke service-level agreements create operational concentration. Your engineering and support capacity gets reserved for one customer's uptime requirements, which means your ability to serve new customers is constrained by an existing obligation you cannot easily renegotiate.
- Evergreen auto-renewal with long notice windows: A 90-day or 180-day cancellation notice window feels protective. It is, until the customer sends the notice and you have six months to replace 25% of your revenue. Long or evergreen terms can defer visibility and create renewal cliffs, especially if multiple agreements co-terminate without diversification.
The seven warning signs worth tracking
Concentration risk is not a static number. It is a trend. A customer at 18% of revenue who was at 12% twelve months ago is a different problem than a customer who has been stable at 18% for three years. The trajectory matters as much as the level.
Track these signals on a monthly basis for your top five accounts:
- Revenue share trend: Is the percentage rising? A customer moving from 18% to 21% to 26% to 29% to 31% over five quarters is a compounding exposure, not a stable one. Visualize the trendline, not just the current number.
- DSO creep: Persistent late payments or rising days sales outstanding from a top account signal working-capital strain and possible over-dependency. When you cannot enforce your own payment terms with a customer, you have already lost negotiating leverage.
- Feature request volume: If one customer is generating a disproportionate share of your product roadmap requests and those requests are getting prioritized, your product is drifting toward a segment of one.
- Sales cycle distortion: If your sales team is spending more than 30% of its time managing, expanding, or renewing one account, your growth capacity for new customers is being consumed by retention of existing concentration.
- Pricing exceptions: How many custom pricing arrangements exist for this customer? Each one is a data point about your actual negotiating position.
- Geographic or sector anchoring: If your largest customer operates in one region or sector, a downturn, regulatory shift, or geopolitical event in that region becomes your problem too, even if your other customers are unaffected.
- Contract term asymmetry: Are the termination rights balanced? Many enterprise contracts allow the customer to exit with 30 days notice while requiring the vendor to continue service for 90 or 180 days. That asymmetry is a concentration amplifier.
Two tools are worth knowing because investors and acquirers will use them during diligence.
The concentration ratio (CR5 or CR10) is straightforward: sum the revenue share of your top five or ten customers. A SaaS company with $2 million in annual revenue where the top five customers generate $800,000 has a CR5 of 40%, which indicates high concentration risk. Comparing to a prior year CR5 of 25% reveals increasing concentration and signals a need for diversification.
The Herfindahl-Hirschman Index (HHI) provides more precision. You calculate it by squaring each customer's revenue share (as a decimal) and summing the results across all customers. An HHI approaching 1.0 indicates near-total concentration in one customer. An HHI below 0.15 is generally considered low concentration. The advantage of HHI over simple concentration ratios is that it captures the distribution across all customers, not just the top tier.
The contrarian case: some concentration is a feature, not a bug
Here is the nuanced take that most concentration risk articles omit: early-stage companies should often accept high concentration deliberately, because the alternative is spreading limited resources across too many small accounts and building nothing well.
The problem is not concentration at $1 million ARR. The problem is concentration at $10 million ARR with no credible diversification trajectory. A company that lands one $2 million enterprise customer at $3 million ARR and uses that contract to build the product, the case study, and the sales playbook is making a rational bet. A company that is still 40% dependent on the same customer at $15 million ARR has failed to execute the diversification that the early concentration was supposed to fund.
The question to ask is not whether you have concentration risk today. It is whether your current concentration is buying you the time and resources to eliminate it, or whether it is consuming those resources to maintain itself.
Practical thresholds for different stages
As a working framework, apply these thresholds relative to your stage:
- Pre-product-market fit: Concentration is expected and acceptable. One or two design partners at 50-80% of revenue is normal. The goal is learning, not diversification.
- $1-5 million ARR: Begin tracking. A single customer above 30% warrants a documented plan for how you will reduce that share over the next 12-18 months.
- $5-15 million ARR: No single customer above 20%. Top five customers below 50%. If you are outside these bounds, it should be a board-level conversation with a concrete mitigation roadmap.
- Series B and beyond: Institutional investors and acquirers will apply the 10% single-customer threshold aggressively. Exceeding it does not kill a deal, but it will affect valuation multiples and introduce reps and warranties around customer continuity.
