Concentration: when one customer is half the revenue.

A single design-partner customer can mask everything wrong with the GTM motion. Here's how we test for healthy vs. unhealthy concentration — and the three remediation paths IC tolerates.

01The disguise

A single customer producing fifty percent of revenue is the most ambiguous signal in early-stage SaaS. The same data point describes two opposite companies. One has captured a flagship reference account that will pull the next ten customers into the funnel; the other has built a custom integration that masks the absence of a repeatable product. The IC partner reading the deck cannot tell them apart from the revenue mix alone. The disambiguation is the diligence.

Across 1,142 dossiers we scored 34% of Series A candidates as exhibiting unhealthy concentration — defined precisely below. Of those, roughly a third recovered through one of the three remediation paths IC tolerates. The remainder either pivoted, raised at materially worse terms, or ran out of cash before the design-partner relationship could be diversified.

The flag matters because concentration deforms every other metric the deck reports. Revenue growth: high, if the design partner is expanding. ARR: high. Logo retention: a perfect 100%. NPS: stellar — the one customer is, by definition, the most engaged one. Burn multiple: excellent. Every metric the founder shows you reads like a Series A worth doing. And the company is one renewal cycle away from a zero, which the metrics will not warn you about because there is only one customer to be wrong about.

A founder with one customer at fifty percent of revenue is not necessarily building a bad business. They are building a company that has not yet been forced to prove the GTM motion repeats. Until it has, every other diligence claim is conditional. The work of this note is to make the conditionality explicit, provide the four tests that distinguish the captured-flagship case from the customised-integration case, and enumerate the three remediation paths that turn the latter into the former.

02Healthy vs unhealthy — the four-test screen

We score concentration against four binary tests. Each is structurally answerable from the deck plus a thirty-minute founder call. Pass three of four and the concentration is healthy; pass two and the diligence continues; pass one or zero and the flag escalates.

• Test one — pipeline composition. Of the founder's top-ten pipeline opportunities by ACV, how many resemble the design-partner relationship in size and integration depth, and how many resemble the modal customer profile the founder is targeting at scale. Healthy: the pipeline is heavy on customers shaped like the eventual GTM motion, not customers shaped like the design partner. Unhealthy: the pipeline is six more versions of the design partner, and the founder is rationalising this as “leaning into what's working.”
• Test two — contract structure. Is the design-partner contract a standard SaaS subscription or a hybrid services contract with a partial subscription component. Healthy: a standard, renewable, multi-year contract with normal SaaS terms. Unhealthy: a services-heavy contract where the subscription line is a small fraction of total revenue and the bulk of the cash is consulting fees recognised as recurring revenue.
• Test three — product shape. Of the features built in the prior twelve months, how many serve only the design partner and how many serve the broader market. Healthy: roughly 70% of the roadmap is broad-market and the design partner is using the same product the next ten customers will. Unhealthy: roughly 70% of the roadmap is design-partner-specific and there is a fork between “the platform” and “what we built for our biggest customer.”
• Test four — renewal economics. What does the design partner pay at the next contract date, and does the unit economics of the renewal match the unit economics the founder is proposing for the next ten customers. Healthy: the renewal is at market rate, the gross margin is in the cohort range, and the founder can defend the renewal price against the comparable. Unhealthy: the renewal includes ongoing services fees, custom integration maintenance, or a discount tied to the original design-partner status that will not extend to the next ten customers.

Founders pass test one most often (pipeline framing is easy to articulate), fail test three most often (the roadmap reveals the customisation), and the combination of failing tests two and four is the most common pattern in dossiers we eventually flag as unrecoverable.

03The 50% rule and what it measures

The fifty-percent line is convention, not science. We adopted it because it is the threshold at which the loss of a single customer produces a financial event the company cannot absorb inside one quarter. Below fifty percent, the company can lose the largest customer and survive long enough to find a replacement. Above fifty percent, the loss compresses runway by more than the time required to close the next account.

Two refinements matter for the early-stage context.

The first is that the threshold scales down at earlier stages. At seed, a single customer producing 30% of revenue triggers the same diligence concern that 50% triggers at Series A. The cushion of replacement opportunity is smaller, the sales cycle to a new customer is longer, and the founder has not yet demonstrated the GTM motion the larger company has nominally established.

The second is that the threshold should be measured against contracted revenue, not booked revenue. A company with three customers each paying twenty percent and a fourth in a six-month pilot at forty percent has a structural concentration risk the booking accounting hides. The pilot is the company. The rest is decoration. We measure against contracted, ratably-recognised revenue under multi-year commitment. The pilot is excluded from the denominator. The flag triggers earlier than the booked-revenue number suggests.

Founders dislike this measurement because it makes their largest customer look larger. The discomfort is the signal. A founder who routinely measures concentration against booked revenue is producing a number designed to clear the diligence screen, not a number that describes the business.

04The three remediation paths IC tolerates

When concentration is real and the company is otherwise fundable, IC will price the risk into the round. The three remediation paths below are the ones we have seen work in our corpus. Each requires a specific commitment from the founder and a specific milestone schedule.

Path one: customer diversification with a hard ratio target. The founder commits to a specific concentration ratio by a specific date — for example, “no single customer above 25% of ARR by end of Q4.” The path requires the founder to demonstrate the pipeline composition test, then to deliver against the ratio on a board-reportable cadence. This is the path most companies attempt. It works when the pipeline is real and the GTM motion is repeatable. It fails when the pipeline is theoretical and the founder is buying time.

Path two: design-partner contract conversion to standard terms. The founder restructures the design-partner contract at the next renewal to remove services components, eliminate custom integration maintenance, and align pricing with the standard rate card. The path requires the founder to be in good enough standing with the design partner to ask for an unfavourable renegotiation, and to have built enough product value into the relationship that the design partner accepts. This is the path that produces the cleanest outcome — the design partner becomes a regular customer at scale and the concentration ratio falls because the denominator changes shape, not because the numerator falls.

Path three: structural separation of the design-partner business. The founder treats the design-partner contract as a one-time services engagement that funds development of the platform, but does not count it in ARR going forward. The financial reporting reflects this. The IC prices the round against the platform-only revenue, treats the design-partner contract as a non-recurring asset, and underwrites the company on the platform-only growth rate. This is the rarest path and only works when the founder is candid about the framing and the design-partner contract is genuinely terminal — the company is not betting on the renewal.

A fourth path — replace the design partner with a larger design partner, then claim to have addressed concentration — appears in our corpus and is not one we underwrite. It is the same risk shape with a different name and the IC will discount the round accordingly.

05What unrecoverable concentration looks like

Three patterns make concentration structurally unrecoverable, regardless of remediation path. We mark each as a hard flag.

• The vendor-of-record pattern. The design partner has signed a contract that explicitly names the company as the sole-source vendor for a category, and the contract value reflects the sole-source pricing premium. The next ten customers will not pay sole-source rates. The unit economics of the design-partner contract are not transferable. The company is profitable today and unprofitable at scale.
• The captive-customer pattern. The design partner has an equity stake in the company, a board seat, or a strategic relationship that makes the customer relationship dual-purpose. The customer is also an investor, partner, or acquirer. The economics of the customer relationship cannot be separated from the equity and governance relationship, and the company cannot diversify without negotiating a structural unwind with its largest shareholder.
• The platform-feature pattern. The design partner is a platform the company runs on, and the design-partner contract is partially payment for distribution. The platform can build the company's product at any time. The contract is, in effect, a delay. See Note 03 for the broader frame. The diligence here is whether the company is investing the contract revenue in building the moat that will outlast the platform's interest.

A company exhibiting any of these three patterns is not a Series A. It is either a services business, a captive subsidiary, or an option the platform has on a feature. None of these are uninvestable categories — they are simply different rounds at different valuations, and the IC has to be honest about which one is on the table.

06How the dossier flags it

Concentration is the third-most-cited flag in our corpus and one of the most consequential for round pricing. We measure it on every dossier and surface it in section four — Financial Forensics — alongside the burn reconciliation. The presentation is deliberately structural.

We compute four numbers. The customer concentration ratio (largest customer ÷ ARR). The top-three concentration ratio. The contracted-only ratio (the same numbers measured against multi-year ratably-recognised revenue). And the renewal-adjusted ratio, which projects each customer's revenue at the next contract date using the renewal posture the customer themselves articulated in their reference call.

The fourth number is the one that matters most. It is also the one no founder would have produced on their own, because it requires triangulating the contract data with the reference-call data. When the renewal-adjusted ratio is materially worse than the contracted-only ratio, the company is more concentrated than the deck implies, and the IC should ask the founder to walk through each customer's renewal posture explicitly.

Where the four numbers diverge, the dossier explains the divergence in plain prose. A company where booked concentration is 35%, contracted concentration is 52%, and renewal-adjusted concentration is 68% is a company in three different stages of concentration risk. The deck has the first number, the founder probably knows the second, and only the diligence has the third.

The recommendation on the dossier is binary against the four-test screen above. Three or four passes: noted, not blocking. Two passes: IC condition, recommend the founder commit to one of the three remediation paths with a specific milestone schedule. One or zero passes: structural concern, recommend the round price reflect the customer-loss scenario directly rather than the founder's blended projection.