Employment without the protections

What one big client is

One client providing most of your income looks like stability and is the opposite. You have the insecurity of self-employment (no notice period, no redundancy, no sick pay, no pension contribution, no holiday entitlement) combined with the dependency of employment. You have taken the worst half of each arrangement.

The threshold worth watching is roughly a third. A client above a third of your income has disproportionate influence over your business. Above a half, they effectively determine whether you continue.

Why it happens without a decision

Nobody chooses this. It assembles itself, and the mechanism is entirely reasonable at every step.

A good client gives you more work. The work is interesting and the rate is acceptable. Because you are busy, you do less outreach. Because you do less outreach, the other clients drift. Each individual month is a sensible response to the situation, and at the end of two years one client is 70% of income and you have not written a new-business email since the summer before last.

The feeling throughout is of doing well, which is the most dangerous part.

What it costs even while it is working

Three costs arrive before anything goes wrong.

Your rate stops moving. Raising prices on the client who is most of your income is a conversation you cannot afford to lose, so you do not have it. Meanwhile your market rate rises without you.

Your standards slip quietly. Scope creep, late payment, rude feedback, a project run badly: you absorb all of it, because the alternative is a fifth of a business. The client is usually not exploiting you deliberately; you stop pushing back, and they cannot see that you have stopped.

Your portfolio narrows. Two years of one client’s work is two years in one category, one style and one market. When you do need new clients, your evidence points at exactly one kind of buyer.

How it ends

Almost never because of anything you did. A new marketing director arrives with their own agency. A budget is reallocated. The company is acquired. A funding round does not close. Your contact, who was the real relationship, leaves. The category has a bad year.

And it ends fast. A client at 60% of income does not taper; they finish the current project and there is nothing after it. That is when you discover that rebuilding a pipeline takes three to six months from a standing start, and that your runway is shorter than that.

The defence, which has to be scheduled

The only effective response is to keep the controllable channels running while you do not need them, which requires it to be scheduled rather than motivated. Two hours a week, in the diary, during the busiest periods, spent on outreach and on contacting past clients.

This is precisely what nobody does, because the two hours feel unnecessary when the work is flowing. The whole point is that they are unnecessary in the week you spend them and load-bearing in the quarter after the big client leaves.

A useful rule is to keep at least three clients live at any time, even if two of them are small. Not for the revenue: for the relationships, the portfolio variety and the fact that a dormant client is far easier to restart than a stranger is to find.

When the concentrated client is good

Sometimes the right answer is to accept the risk deliberately. A client at 50% who is well-paying, pleasant and stable may be worth it for a period.

If so, make it an explicit decision with conditions attached:

• The runway buffer goes up, to cover the months it would take to rebuild.
• The two hours a week continue regardless.
• You review it every quarter rather than noticing it annually.
• You do not let the arrangement change your rate discipline with everyone else.

That is a managed risk. What is not managed is the same situation arrived at by drift and never examined.

The disguised version

One further form worth recognising: a single agency supplying work from many end clients. It feels diversified because the projects vary, and it is not. If the agency loses its account or restructures, all of it goes at once, and you have no relationship with any of the end clients whose work you did.

Treat an agency as one client for concentration purposes, because that is what it is.

Rebuilding after it happens

If it has already ended, the sequence is fixed and there is no clever version.

Start with the channels that can move fastest: direct outreach, and agencies or studios who might subcontract. Contact every past client, including ones from years ago, with an honest note that you have capacity. Take work slightly below your ideal for a short period to cover the gap, with a defined end to that concession. And rebuild the buffer before you rebuild the rate.

It is a hard three months and it is survivable, and the people who come through it best are the ones who start on day one rather than spending three weeks hoping the client comes back.

The limitation

Concentration is not always avoidable. Some markets are small enough that one or two buyers dominate the region. Some specialisms have few possible clients by nature. A person in a small city working in a narrow field may have no way to spread the risk.

Where that is true, the defence shifts from diversifying clients to diversifying income shape (a productised service, teaching, a product, work in an adjacent category) and to holding a substantially longer runway than somebody with twelve clients needs. The risk cannot be removed, so it gets absorbed instead, and absorbing it requires knowing it is there.