The Second Client Problem: Why Your Agency Stalls at Two Big Accounts

Almost every agency we've met hits the same wall. Not at zero clients, not at ten — at two. Two accounts big enough to pay the team, demanding enough to eat the founders' calendars, and fragile enough that one bad quarter from either kills the whole business.

This is the second client problem. It looks like success from the outside and feels like a hostage situation from the inside. If you're nodding, this piece is for you.

What the second client trap actually looks like

The pattern is boringly consistent. You land Client A — maybe a Series B startup or a mid-market retailer — and you staff them with roughly half the team. Margins are decent because you're billing senior rates and the work is interesting enough to keep engineers happy.

Then Client B arrives, often through a referral from A. Now you're at ~90% utilisation. The founders stop selling because they're the two most senior engineers on the two most important accounts. Hiring slows because onboarding a new engineer requires one of those founders to peel off and mentor them, which neither can afford.

From the outside: "They're killing it, fully booked, probably turning work away." From the inside:

If either of these two clients pauses for a quarter, we fire three people and I remortgage the house.

The trap isn't the clients. It's the shape of the P&L they produce.

The revenue concentration math

A quick back-of-envelope. Say you're a 12-person agency doing roughly $2.4M a year. Client A is $1.1M, Client B is $900K, and the rest is odds and ends. That's 83% of revenue in two logos.

In SaaS, nobody would fund you at that concentration. In services, it's considered normal — until it isn't. The honest risk-adjusted value of that revenue, if you discount for churn probability over 18 months, is probably closer to $1.6M than $2.4M. You're running the business as if the headline number is real. It's not.

Why founders get stuck here

There are three forces holding you in place, and they all feel rational in isolation.

The margin feels great right now. Big clients mean fewer sales cycles, less context switching, longer engagements. Your effective margin on a well-run $900K account is dramatically better than on nine $100K accounts, because overhead doesn't scale linearly with headcount — it scales with number of relationships.

The clients want more of you, not less. Good work generates expansion. The temptation to say yes is enormous because the alternative — telling a happy client you can't take the next phase — feels insane. So you take it. Now they're 55% of revenue instead of 45%.

Hiring to break out requires money you don't have yet. To diversify, you need bench capacity to pitch and onboard a third and fourth client. Bench capacity means hiring ahead of revenue. Hiring ahead of revenue when 83% of your revenue is two phone calls away from leaving feels irresponsible. So you don't.

The result: a stable, profitable, utterly stuck business.

The structural moves that actually work

We've watched agencies escape this, and the ones that do almost always use some combination of four moves. None of them are a single clever tactic — they're structural.

1. Cap any single client at a hard percentage

Pick a number — 30% is aggressive, 40% is realistic for most agencies under 20 people — and treat it as a policy, not a preference. When a client's expansion would push them over the cap, you either raise rates to slow the growth, subcontract portions out, or politely sequence their roadmap over a longer window.

This sounds like leaving money on the table. It is. You're paying an insurance premium in the form of foregone revenue, and in exchange you buy the ability to lose any single client without existential consequences. Founders who've been through a surprise offboarding never argue about this twice.

2. Build a productised tier under your custom work

Most agencies have one offering: bespoke engineering, billed by time or by fixed scope. That's a single revenue shape. If you add a second shape — a fixed-price, fixed-scope offering that smaller clients can actually afford — you create a funnel of clients 3 through 15 who collectively dilute concentration without requiring you to land another whale.

Examples we've seen work:

• A 4-week discovery and architecture package at a fixed price
• A "first production deploy" package for teams moving off no-code
• A security/performance audit with a defined deliverable
• A managed Shopify or Next.js maintenance tier at a flat monthly rate

The point isn't the specific product. It's that productised work can be sold by someone who isn't a founder, delivered by someone who isn't your most senior engineer, and closed in a sales cycle measured in days instead of months. See how we structure this on our services page for one version of the shape.

3. Separate the delivery lead from the technical lead

On big accounts, the founder is usually both the account owner and the lead engineer. That's the single biggest reason you can't sell to a third client — the person who closes deals is pinned to delivery.

Split the roles. The delivery lead owns the client relationship, scope, invoices, and roadmap conversations. The technical lead owns architecture and code review. These can be two people, or one person who is explicitly not a founder. The founder's job becomes sales, hiring, and unblocking — not shipping tickets.

This is painful because for years the founder's identity has been "the person the client trusts." Transferring that trust is a six-month project, not a Slack announcement. Start now.

4. Instrument the pipeline you've been ignoring

Most stuck agencies have no pipeline because they haven't needed one. The symptom is simple: ask the founders what deals they expect to close in the next 90 days and you get vague gestures.

Fix it with the lightest possible CRM discipline. Something like:

stages:
  - inbound: form fill, referral, or warm intro
  - qualified: budget confirmed, timeline within 6 months
  - scoping: paid discovery or proposal in flight
  - verbal: client has said yes, contract pending
  - closed_won: SOW signed

review_cadence: weekly, 20 minutes, founders only
target_coverage: 3x of next-quarter revenue gap

That's it. The point isn't the tool — pick anything, even a spreadsheet. The point is that you cannot escape the two-client trap without a visible, countable pipeline, because the moment one of your two big clients wobbles, you need to know what's behind them.

What this costs you in the short term

Let's be honest about the trade-off. In year one of actually fixing this, your revenue growth will probably slow. You'll turn down expansion from Client A, you'll hire a delivery lead who isn't billable at full rates for six months, and you'll spend founder time on sales calls that don't close.

In our experience, the hit is in the range of a quiet year or two — not a catastrophe, but noticeable. The payoff is that by the end of that period, no single client is a gun to your head, you've got a hiring plan that doesn't depend on one invoice, and the founders have their calendars back.

The agencies that skip this phase and keep accepting expansion from their two whales usually make more money for 18 months and then have a very bad quarter. We've seen it enough times to stop being surprised.

Where we'd start

If you're reading this and recognising your own P&L, do three things this week before anything else:

1. Calculate the real concentration numbers. Not vibes — percentages of trailing twelve-month revenue by client. Write them down.
2. Pick a concentration cap and tell your co-founder. Not the team yet, not the clients. Just agree on the number.
3. Block two hours on Friday to list every warm intro, lapsed lead, and "we should talk again" email from the last 18 months. That's your pipeline starter kit.

The second client problem is boring, structural, and almost entirely solvable. It just requires admitting that the thing making you comfortable right now is also the thing that will hurt the most when it breaks. For more on the economics side of this, our blog has pieces on retainers and paid discovery that pair well with this one.