About The Investigation. Every Wednesday, ADOTAT publishes one long piece built from primary sources rather than press releases: written on-record answers, contracts, filings, transcripts, and the questions the trades tend not to ask. The format is deliberately narrow.

One subject, named companies, named executives, and every claim put to the people it concerns before publication, with their answer printed as they gave it or their silence noted. Nothing here is sponsored, nobody reviews it in advance, and the reporting runs whether or not the subject is a friend of the publication.

WPP is doing the right things.

Late, under a chief executive who inherited the mess rather than built it, but doing them, nevertheless.

The H1 2026 results, reported August 6, were the first encouraging set in two years. Like-for-like net sales fell 4.7 percent across the half. But Q2 came in at minus 2.8 percent against Q1's minus 6.7 percent, which is the difference between a company falling and a company landing. Margin up 20 basis points to 8.4 percent. Net debt down £326 million. And the shares had their biggest single-day gain since the 1995 IPO.

And the number that ought to matter most in a story about a pitch: WPP ranked first in JPMorgan's net new business rankings for Q4 2025 and again for H1 2026. Estée Lauder. Jaguar Land Rover. Henkel. Airbnb. Heineken. Honda.

That is not a company nobody wants to hire.

Cindy Rose, on the call: "When we show up as one WPP, media-led, data and technology-enabled, structurally integrated with creativity at our heart, we are the growth partner of choice."

Stabilize in 2026, grow in 2027, accelerate in 2028. I believe in her.

Two quarters of sequential improvement is not proof a three-year plan is working. It is the first evidence in two years pointing that way, and after the decade WPP has had, somebody should tell the woman she is not imagining it.

Our view is that none of it will be enough to keep Coca-Cola intact.

That is a forecast, not a reported outcome, and we are labelling it as one before we go any further. It rests on base rates, on WPP's own record in this specific kind of review, and on what three people who do this for a living told ADOTAT. Not one of them argued that WPP is badly run today. That is what makes the answer interesting rather than obvious.

Because there is a second number from the same results, conceded on the same call, which received roughly one-tenth the coverage.

Gross client losses remain elevated, at the top of the 500 to 600 basis point range.

The whole story in two lines. WPP is winning new business faster than it has in years and losing existing business faster than it can afford to. Everything Rose has fixed lives in the first column. Coca-Cola lives in the second.

When Coca-Cola opened this review, WPP welcomed the chance to "showcase how our integrated media, data science, and agentic technology platform and solutions will continue to drive future growth."

Now read what Coca-Cola says about the same relationship. In a statement provided to ADOTAT by Scott Leith, Vice President of Financial and Corporate Communications, the company says its partnership with WPP Open X "has helped to modernize our marketing approach and deliver significant business value."

Sit with that for a second.

The client says the incumbent delivered.

The incumbent says its technology will keep delivering.

And the account is in review anyway.

Nobody puts $1.44 billion out to tender because they are mildly annoyed. This is not a company firing an underperformer.

This is a company that has decided the arrangement itself is the thing it wants to reprice.

Which is why the scope is the entire story and almost nobody has said so out loud. Coca-Cola is reviewing media, data and technology. WPP's answer is that it is very good at media, data and technology.

That is not a defense. That is the thing being audited.

There is a sentence in the middle of Michael Farmer's answer to ADOTAT that explains this review better than anything either holding company has said about it, and it is not about agencies at all.

"Coca-Cola has not been a strong performer for a very long time. Its sales in 2025 were barely above its sales in 2012."

Farmer has spent his career measuring what agencies are asked to do against what they are paid.

His point is structural rather than gossipy: "Companies with weak sales performance become 'cost reducers' across every element of cost in their portfolios."

Fourteen years of essentially flat sales at one of the most valuable brands on earth. That is the context in which a client announces it wants to move "from traditional media planning to the emerging ways we need to reach consumers through technology, including agentic tools."

Farmer's read on that language: "I suspect that Coca-Cola's focus on agentic tools is a signal that they want lower costs in a marketing environment in which their performance has been below their hopes."

Everybody covering this review is scoring the agencies. The client has been the underperformer for fourteen years.

Almost every story about this review has described a $4 billion account, which is a wonderful number, performs beautifully on LinkedIn, and is wrong.

The $4 billion was the 2021 global consolidation figure, the assignment WPP won five years ago and built Open X to service.

Since then it has been carved like a Thanksgiving bird. North America went to Publicis in 2025, roughly $700 to $800 million. Japan and South Korea sit with Dentsu. Global creative and PR stay with WPP Open X no matter who wins.

None of that is in this review.

So ADOTAT did the radical thing and asked. COMvergence produces the estimates most of the trades quote secondhand without ever picking up a phone. Olivier Gauthier answered in writing: "When excluding North America and Japan, the total currently reviewed amounts to $1.47B."

Then, the same afternoon, before we printed it, he revised his own number upward: "Actually, please use $1.5 billion. I've added the Brazil spend, even though media agencies don't really operate in that market per se."

A source correcting himself against his own interest, unprompted, before publication. In this industry that is close to a religious experience.

Then it moved a third time, and this is how a figure earns the right to be printed.

Coca-Cola's statement excludes North America, Japan and Korea. Three markets. COMvergence had calculated on two. Korea has not appeared in a single piece of trade coverage of this review. We took it back to Gauthier. He answered the same day:

"Coca-Cola's media spend for South Korea is $65M, bringing the total net media spend in review to $1.44B."

Three revisions. Up when he added Brazil, down when we found Korea. The number is worth printing not because it is large but because somebody actually checked it.

So: not $4 billion. $1.44 billion, on COMvergence's estimate, more than 60 percent smaller than the figure the trades have repeated for three months.

It is still an enormous amount of money, and WPP would very much like to keep it.

COMvergence's published 2025 results covered $37.4 billion in media spend reviewed across more than 4,400 account moves. Market-wide, incumbent retention was 21 percent, the lowest in eight years. Inside that: Publicis Media first with roughly $10 billion in net new business. WPP Media at negative $6.9 billion.

A March 2026 AdTech Today report, in its own framing, put WPP Media's retention at 16 percent of the $8.3 billion it defended, citing COMvergence.

We could not obtain the original page, so we attribute that figure to AdTech Today rather than presenting it as verified COMvergence data.

WPP pushed back on the reading, and we are printing the objection here in the free section rather than behind a paywall.

Sam Weston of WPP Media, August 28: "I think your core thesis and the math behind it is off base. We've seen significant improvement in both retention and new business since the launch of WPP Media, evidenced by our retentions and expansions on major accounts like the UK Government, Henkel, Reckitt."

And on the number, citing COMvergence's own published report: WPP Media's retention rate "was essentially the same as Omnicom's (16% vs their 17%) and from a retained billings standpoint we placed 3rd of 6, just $340M out of second place."

That is a fair challenge and it lands. A rate and a dollar total measure different things. An incumbent defending a larger book can keep a smaller share of a bigger pile and still finish third in dollars kept. Sixteen against seventeen is not an outlier, and a characterization in an earlier draft of this piece that called WPP's record an outlier did not survive Weston's table, so it is not in this one.

It is also not the number this piece turns on.

Because there is a second figure, given to ADOTAT directly, which is not in the report Weston sent us.

"As for WPP Media, they were the incumbent in 68 pitches and successfully defended only 10 of them."