The Standing Gallop

This month, two holding companies raised their full-year organic growth guidance to the identical range: 4.5% to 5%.

Impressive, but in the last quarter, one of them grew the whole company by around five per cent. The other grew the whole company by 0.15%. So how can this full year goal be possible?

Eight months ago, the day the Omnicom-IPG merger closed, I wrote that it had built a bigger horse, not a faster one (and still a horse, and not a car). Yesterday's second-quarter results appeared at a glance to prove me wrong: 6.1% organic growth, ahead of every expectation, guidance raised, John Wren declaring the momentum of the new Omnicom, built for an era where speed, integration, and scale matter most.

The horse, he wants us to believe, is galloping.

Look closer. It hasn't moved.

And in November, when I wrote about Omnicom's gross-versus-net reporting and warned that market leadership demands market transparency, I described exactly the machinery that makes a standing horse appear to run. Well, yesterday that machinery ran at full speed.

So today's piece is not really about Omnicom's quarter. It's about how to read a holding-company growth number and know what it's actually made of. Consider this a field guide, published a few weeks before WPP hands us the biggest test case of the year.

More than a decade ago, inside Publicis, I worked on a project that nobody outside the building would ever have noticed. Across our production businesses (then called Prodigious, now part of a very different Publicis Production) we moved the reporting from gross to net. It was unglamorous, technical, and slow. It transformed how a substantial business measured itself, and it made our numbers look smaller, not bigger, on the day we finished. It was the right thing to meet both the rule and spirit of GAAP.

The leadership view was simple: accuracy builds trust, and trust is the only currency a holding company actually banks. That was ten years ago, in the last corner of Publicis where gross and net could still blur. That project taught me the thing this piece turns on: a reported growth number is not a fact of nature. It is the product of definitional choices. Every one of those choices is legal. Every one is defensible in isolation. And every one is a choice.

Which brings us to yesterday.

Let me be precise before I am critical. Every number Omnicom published yesterday is accurate, audited, and footnoted in its own release. Nobody is hiding anything; it is all there in the tables. What follows is not an allegation. It is an assembly manual for the three mechanisms that turn a standing horse into a galloping one.

Mechanism one: draw the perimeter.

Omnicom's headline growth applies to "Core Operations" or the businesses management has chosen to keep. Its own table shows what happens when you widen the lens. Combined Omnicom-plus-IPG revenue in Q2 2025: $6,552m. Total revenue in Q2 2026: $6,563m. The entire enterprise, everything the merger assembled, grew by 0.15% in a year. The 6.1% lives exclusively inside a boundary and the boundary keeps moving. On yesterday's call, CFO Phil Angelastro confirmed the annualised revenue earmarked for disposal has risen from $3.2 billion to $3.5–3.6 billion, and that around 60% of the newly added businesses are advertising businesses. Asked how sustainable the growth rate is, John Wren answered candidly: they have divested the low-growth businesses that previously dragged the growth rate down. That is the perimeter mechanism described by its own operator. Remove the slowest parts of the horse, and the average speed of what remains rises while the animal stands still.

Mechanism two: report the gross number.

Omnicom remains the only major holding company reporting organic growth on a gross basis, pass-through costs included, and Q2 shows why it matters. Growth was led by Integrated Media and Experiential, both up around 10%, precisely the two disciplines that carry the highest share of pass-through costs. Brian Wieser's Madison and Wall, the same analyst I cited in November, ran the arithmetic: third-party service costs of roughly $1.75 billion in the quarter, against a like-for-like base nearer $1.25 billion, implying around $500 million of the revenue growth came from pass-throughs, against a total organic gain of $339 million. On Wieser's maths, comparable net revenue declined roughly 2% organically. Read that again. On the basis every other major player reports, the industry's largest company may have shrunk last quarter. Meanwhile the discipline called Advertising fell high single digits and now represents 15.7% of core revenue. At the world's largest advertising company, advertising declined again, for the second consecutive quarter, and is heading towards a seventh of the business.

Mechanism three: let cost play the part of growth.

Core margin rose almost 200 basis points to 17.8%, and the release attributes it, in its own words, primarily to cost reduction synergies. The $900 million synergy programme is running ahead of plan. But adjusted EBITA still missed analyst expectations, the implied full-year EPS guide sits below consensus, and the drop-through of incremental revenue to profit ran below what the synergy arithmetic promised, which is what you would expect if a meaningful slice of that incremental revenue is pass-through cost wearing a growth costume. Margin built on subtraction is real money. It is not the same thing as a business getting better at what clients pay it for.

Three mechanisms. One headline. A standing gallop.

And to be fair to the horse: this will likely be the best print any traditional holding company reports this season. The new business is real. Pruning weak revenue is what a serious acquirer should do. Two hundred basis points of margin expansion in the first year of the largest integration in advertising history is execution by any standard. If the future of this industry was a horse race, Omnicom may win it.

That last sentence is the problem.

Because the future of this industry is not a horse race, and here is where I part company with every analyst note that scores yesterday as a beat.

Scale was the entire logic of this merger. But scale relative to what, and in which part of the value chain? Measured against other agencies, the new Omnicom is unarguably the biggest: the biggest at media planning, which platforms are automating; at creative production, which AI is compressing; and at traditional services, which clients are in-housing. Measured against the pool that actually determines who captures marketing value (the platforms, the media owners, the adtech and martech companies, the consultancies) Omnicom's combined revenue is a rounding error on Google's advertising business alone.

This is the trap I described in A Bigger Horse, now visible in the results rather than the deal model. The relevant question was never "are we bigger than WPP?" It was "are we becoming something the wider ecosystem cannot commoditise?" Yesterday's release answers that question in its own discipline table: the growth is in pass-through-heavy media volume, the decline is in advertising, the margin is coming from severance, and the businesses being sold are fetching modest proceeds because the market is not paying much for what the middle no longer needs. The perimeter can keep shrinking, quarter after quarter, and the reported number can keep looking healthy that is precisely what makes the mechanism so seductive.

I moved a business from gross to net a decade ago and watched the numbers get smaller and the trust get bigger. Omnicom, at four times the size and with the whole market watching, is running the trade in the opposite direction.

Which sentence will be true in December?

Come December, both Publicis and Omnicom will very probably report that they delivered organic growth of 4.5 to 5 per cent. Both sentences will be accurate. Only one will describe a company that grew.

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 29 July 2026 in the Brandflow newsletter on LinkedIn.