UK advertising: the turnover league table lies — and the biggest 'agency' is Facebook

About 250 UK advertising companies publish a full profit-and-loss, booking £16.7bn of combined turnover — and that league table is a hall of mirrors. The single biggest entry isn’t an agency at all: it’s Meta’s UK arm, the platform the agencies buy from. Two more of the top twelve own the billboards rather than plan the campaigns. And among the actual agencies, “turnover” means two different things — client media money passing through one shop, fee income at the next — so the same line item can differ twenty-fold per employee between two firms doing comparable work. Rank this market by turnover and you learn almost nothing; read the accounting model first and the whole map snaps into focus. Figures are approximate — verify against a company’s own accounts before relying on any single number.

First, evict the impostors

Read the names before the numbers. Facebook UK (£3.11bn turnover, £362M pre-tax profit) is Meta’s British sales-and-engineering operation — it sells the ad inventory this market buys, and its cost-plus service economics have nothing to do with agency life. JCDecaux UK (£456M) and Global Outdoor Media (£234M) are out-of-home media owners — they rent out poster sites and screens; an agency is their customer, not their competitor. And an Alibaba cloud-computing subsidiary surfaces in the margin tables for good measure. All four sit in the advertising category in the raw data; all four are excluded from every competitive read below.

The caveat that changes every number: billings are not fees

An agency that buys media as principal books the client’s whole media budget as turnover and passes most of it straight to the platforms and media owners, keeping a thin sliver. An agency working as agent books only its fee. Both are normal; neither is better; but they make turnover — and margin — incomparable across the two models. The per-head figures make the split unmissable:

*Dentsu UK’s 13.6% is not a trading margin: £35.6M of its £44.3M pre-tax profit is dividends from subsidiaries, and at the operating line it made roughly 1% (the year before was an operating loss). Its turnover-per-head is still the honest fee-model anchor — its revenue line books fees and commissions, not billings.

No advertising person is £2M-a-year more productive than another. The top half of that table is booking client media money as its own revenue; the bottom half is booking something closer to fees. The sub-2% margins up top are the pass-through model working as designed — not distress — and the 5–10% margins below sit on a much smaller, fee-like revenue base (Dentsu’s 13.6% headline sits higher only because of dividend income — see the footnote). The two Omnicom entities are the proof that this is accounting, not economics: OMD EMEA and Omnicom Media Group UK belong to the same group and sit at opposite ends of the table. Never compare a margin across that divide.

The giants

With the impostors evicted, the top of the market is the holding-group UK arms plus a handful of scaled independents:

*Inside Ideas Group’s accounts consolidate Oliver’s worldwide operations for its parent, The Brandtech Group — only £96.6M of the £370.2M turnover is British (North America alone is £130.6M), so most of that £48.4M profit is earned abroad. On a UK basis it’s a ~£97M player; it stays in the table for scale, with that health warning.

…and the other ~240, once the impostors are set aside.

The reads worth keeping: The7stars — the largest independent in the market — held revenue +7% while trimming headcount −6%, the classic margin-defence posture. Dentsu UK is quietly contracting on both lines. Havas Media’s £110k of profit on £310M of turnover looks alarming until you remember the billings caveat — though even by pass-through standards it is running close to the bone. And Inside Ideas Group — the in-housing model, embedding agency teams inside client organisations — posts £48M of profit, but that’s Oliver’s worldwide business consolidated for The Brandtech Group; on its British turnover alone (~£97M) it belongs with the scaled independents, not the giants.

The shape of the market

The visible market starts at about £5M. Below that, most agencies publish abridged numbers, so the sub-£5M rows here are sparse and skewed — the under-£1M tier (35% profitable) is mostly shells and wind-downs, not a readable small-agency economy. From £5M up the picture is healthy and gets healthier with scale: 76% profitable at £5–25M, 84% at £25–100M, and 97% in the £100M–1bn band. Advertising’s mid-market, unusually, has no distressed tier — the struggling agencies simply shrink below the reporting threshold and disappear from view.

Where the money actually is: the fee tier

Strip out the group treasury artifacts (more on those in a moment) and the £18–60M fee-model tier is where advertising quietly makes real margins — 15–35% is common, which is what the billings-inflated giants’ sub-2% obscures:

*Freuds’ £9.6M pre-tax profit includes a £3.1M dividend from a subsidiary; its trading operating margin is ~22.5% — still comfortably inside the fee-tier band. Turnover also slipped ~6% year on year.

The pattern: reputation-led fee businesses. Freud Communications — communications and PR — keeps about 22.5% of £28M at the trading line (the 34% headline includes a subsidiary dividend). Drummond Central, a Newcastle independent, keeps 27% and is growing. Pablo London and Hybrid News are growing at healthy mid-teens-plus margins. Onetag’s 36% on 11 heads is the UK entity of a wider ad-tech group — group-structure economics as much as software economics — and group creative shops like Adam&Eve\TBWA and PHD show that within the networks, the creative and planning entities carry the margin the media entities don’t.

A caution on the margin tier. Several entities posted margins no people-business can earn — 91% at one WPP out-of-home unit, 93% at a telephone-audio firm’s vehicle, 60% at a Bauer outdoor entity, 40% at an investments holdco. Those are intra-group fee and holding structures, not agency economics, and they are excluded above — as is the Alibaba cloud subsidiary. In a market this full of group entities, treat any margin above ~40% as a corporate-structure artifact until proven otherwise.

Growth, read with care — reorganisation wears growth’s clothes

In most markets, revenue growth backed by hiring is the honest organic signal. Advertising breaks that heuristic, because the holding companies constantly reshuffle their legal entities: when a group folds two agencies into one company, the surviving entity “grows” spectacularly in both revenue and staff. VML (UK) (+69% turnover, +65% staff) and WPP Media UK (+57%, +44%) are consolidations of that kind, not market share gains. At the wilder end, +12,526% at Blume Group and +1,293% at In Focus are low-base artifacts, and Making Science UK’s +296% came with a loss and a fifth of the staff gone — revenue acquired, not yet earned.

The credible organic growers are quieter: Bicycle London (+63% turnover, +48% staff, profitable), Anything Is Possible Media (+56%, +9% staff) and Fuse International (+105%, +11% staff) — all thin-margin, all consistent with young media shops winning billings-heavy accounts.

Market structure: less concentrated than it looks

On paper the top 5 hold 36.8% of the £16.7bn — but the top 5 includes Meta’s UK arm and a holding-group consolidation, and the billings model inflates every media agency’s share. Strip the impostors and translate billings to fees, and actual agency-land is a genuinely fragmented professional-services market: dozens of £20–100M firms competing on reputation, with no operator anywhere near control. Concentration measured on turnover is doubly distorted here — treat the curve as an upper bound.

Ownership and vintage

Of the 250 companies, 127 are corporate-owned against 111 individually owned — an unusually even split for a market this size, reflecting how much of it is founder-led independents. About 19% carry Holdings/Group/Topco-style names, the structural fingerprint of a buyout or a planned exit; agencies remain a steady private-equity and network acquisition target. The biggest founding cohort is the 2000s (83 of 250) — the digital-era independents, now the market’s £20–100M substance — with a healthy 2010–15 wave (41) behind it. Vintage is no moat: the pre-1990 cohort (44) is mostly network entities, while the strongest independents here are under twenty-five years old.

What the map shows

1. Turnover is not a scoreboard here. The same line means client media billings at one agency and fee income at the next — a twenty-fold per-head spread (Talon ~£2.2M vs Dentsu UK ~£115k). Read the accounting model before any ranking.
2. The biggest entries aren’t agencies. Meta’s UK arm, two billboard owners and a cloud subsidiary sit in the category; every competitive read has to start by removing them.
3. Sub-2% margins at the media giants are the model, not distress — and the 15–35% margins live in the £18–60M fee tier: Freuds, Drummond Central, Pablo London, Hybrid News.
4. Group entities’ growth mostly measures reorganisation. VML (UK) +69% and WPP Media +57% are mergers of legal entities; the honest organic growers are young media shops like Bicycle London.
5. It’s a fragmented, founder-heavy market under a distorted top. Corporate and individual ownership split almost evenly, the 2000s independents are its substance, and real concentration is far below what the turnover curve suggests.

Methodology and caveats

This covers only the UK advertising companies that publish a full profit-and-loss — most smaller agencies report abridged numbers and don’t appear, so the sub-£5M tiers understate the market’s long tail. A platform’s UK arm, out-of-home media owners, a cloud subsidiary and several intra-group fee structures are excluded from the competitive reads, and margins are never compared across billings-based and fee-based revenue models. Group entities’ pre-tax profit can include dividends from subsidiaries; where that materially flatters a margin it is noted inline. Figures are approximate and business-type labels are directional — verify any specific figure against the company’s own accounts before relying on it. This is analysis, not financial advice.