The short version
- Concentration is counted in decision-makers, not logos. Eleven brands across eight accounts is eight decisions, and three of those brands sit behind one of them.
- Verticals concentrate too. Eight accounts across five verticals means at least two share a market, and a freeze there reaches both.
- Delivery knowledge concentrates faster than revenue. A rebuild of 42 routes with 387 entries and 545 assets teaches a stack that transfers nowhere.
- Weekly delivery is the cheap defence: roughly 50 chances a year to notice an account going quiet, against 12 on a monthly cadence.
Client concentration is the share of revenue and senior attention that one decision can remove. Logo counts hide it. We run 11 brands across 8 accounts, and three of those brands answer to one owner, so the brand count overstates how many separate decisions the work rests on.
What is client concentration risk?
The share of revenue, and of the senior attention behind it, that a single decision can remove. That decision usually has nothing to do with the work. A marketing lead changes or a budget moves into product, and neither is a verdict on anything you shipped.
Our own counts: 8 client accounts, 11 brands, five verticals, weekly delivery on every account. Eleven brands look like eleven relationships. They are eight, because three of those brands sit with a single owner.
The counts we keep, and what each one hides:
| What we count | Our number | What the count hides |
|---|---|---|
| Brands managed | 11 | Three of them answer to one owner |
| Client accounts | 8 | How unevenly revenue and senior hours split |
| Verticals served | 5 | Eight accounts in five markets means some share a market |
| Delivery cadence | Weekly, every account | Which account absorbs the most attention |
| Stack knowledge | Counted nowhere | It leaves with the account that taught it to you |
Why does the brand count overstate diversity?
Three of the brands we run belong to one owner. Each needed its own voice guide and its own ICP, because shared ownership does not produce shared positioning, and a buyer looking at one of those products has usually never heard of the other two.
That is three accounts of work resting on one relationship. Three content calendars and three sets of review notes resolve to one person deciding whether any of it continues, and if that decision goes the wrong way, three brands leave the book on the same day.
Verticals stack the same way. Eight accounts across five verticals means at least two sit in the same market, and a downturn there reaches both in the same quarter. It will not feel like concentration, because the two accounts have different names on them.
What concentrates besides revenue?
Delivery knowledge, and it concentrates faster than money does. One rebuild we ran came to 42 routes, with 387 content entries and 545 assets moved out of a hosted CMS. Roughly one record in six carried something irregular, a legacy field or a reference pointing at nothing in the new model.
Publishing on that site turned out to be a push to a branch rather than a deploy command. Learning that costs a day, and none of it transfers to the next account. When the relationship ends, the process built around its stack ends with it, in the quarter with the least room to build another.
Attention concentrates on the same account. A rebuild takes senior hours that would otherwise spread across the book, and it takes them for months. That never appears on an invoice, which is why nobody counts it until the account is gone.
Rank the book by the number of separate decision-makers behind it. That list is always shorter than the client list.
How do you reduce it without turning down work?
Standardise delivery so that no account’s process is shaped around its own stack. A workflow built for one client’s tools has to be rebuilt when they leave, in the exact month there is no capacity for it.
Ship weekly on every account. A weekly cadence gives roughly 50 chances a year to notice a relationship going quiet, against 12 on a monthly one. A month of silence is information you can still act on. By the end of a quarter, the decision has been made in a room you were not in.
Keep the relationship structural rather than personal. Each client gets a portal showing their own work, so the account does not live inside one person’s inbox. That does not stop a client leaving. It stops the leaving from taking the record of the work with it.
The two pieces this sits between are the day our largest account left and why the cadence is weekly.
Frequently asked questions
What is client concentration risk?
The share of revenue and senior attention a single decision can remove. Count it in decision-makers rather than logos: three brands owned by one company are one decision, and all three leave on the same day.
How do you measure client concentration properly?
Rank the book by decision-maker rather than account name, then add the senior hours each one absorbs. Two accounts in the same vertical belong together as well, because a budget freeze in that market reaches both at once.
Does managing several brands for one client reduce concentration risk?
No. It multiplies the delivery load without dividing the exposure. Three brands under one owner each need their own voice guide and ICP, so the work is three accounts wide while the decision to continue sits with one person.
Is a concentrated account safe if the client is happy?
Satisfaction is not the variable. The decisions that end engagements are structural: a new marketing lead, or a budget moved into product. Neither is a judgement on the work delivered.
What concentrates besides revenue?
Delivery knowledge and senior attention. One rebuild covered 42 routes and 387 content entries, with a publish path that had to be learned before anything shipped. That knowledge is worth nothing on the next account.
How do you reduce concentration without refusing revenue?
Standardise delivery so no process is built around one client's stack, and ship weekly so an account going quiet shows up in a month rather than a quarter. Keeping the record of the work in a client portal does the rest.
Sources
- Chua Network delivery data across 8 client accounts (internal fact bank)
- Chua Network engagement records, anonymized (internal experience bank)
