In 1961, Marion Harper Jr. performed a small piece of corporate magic. He put several advertising agencies inside one company and declined to make them look alike. They could keep their names, their people, their temperaments and even their competing clients. The parent would provide capital and control. The agencies would provide the electricity. He called the structure Interpublic.
The arrangement solved a peculiar Madison Avenue problem. A single agency could not comfortably sell soap for two rivals. A holding company could own separate shops for both, provided the walls held. IPG did not merely become another advertising group. It supplied the diagram that WPP, Omnicom, Publicis and others would enlarge for decades.
The federation was the product
To a chief marketing officer, IPG was less a company than a cabinet of specialists. McCann and FCB made global advertising. UM, Initiative and Mediahub planned and bought media. Weber Shandwick and Golin handled reputation. IPG Health spoke the regulated dialect of medicine. Jack Morton built experiences; Octagon worked in sports and entertainment; Craft produced content; MRM handled customer relationships.
The customer did not buy a software seat. The customer bought teams, hours, ideas, negotiations and execution. Revenue arrived through retainers, project fees, commissions, media and production markups, and occasional performance incentives. The holding company earned its place by financing the expensive parts, settling conflicts and persuading independent-minded agencies to collaborate when a global client wanted one answer.
That last number matters. IPG served famous, enormous marketers, but no single one dominated the ledger. Its largest client supplied roughly 4 percent of 2024 revenue before billable expenses. The biggest sectors were financial services, healthcare, and technology and telecom. Diversification softened any one departure. It could not soften several large media decisions at once.
The $2.3 billion answer
By the 2010s, the client-conflict problem was no longer the only organizing problem in advertising. Google, Meta, Amazon and their peers owned audiences, ad inventory, interfaces and rivers of behavioral information. Marketers wanted their agencies to know not just what to say, but who should see it, where, how often and whether the person bought anything afterward.
IPG's response was unusually concrete. In 2018 it paid $2.3 billion in cash for Acxiom Marketing Solutions. Acxiom could organize customer records, resolve identities and turn first-party information into addressable audiences. The purchase gave the creative federation a data spine. KINESSO followed in 2019 to connect identity, media activation and measurement. Interact arrived in 2023, combining Acxiom data, Adobe technology and agency workflows across content, customer experience and commerce.
Brand, campaign and craft
Data, activation and measurement
It was a logical bet, and a copyable one: buy the missing capability instead of pretending every creative agency can grow it from scratch. Keep the expert brand. Connect it through shared tools. Then make the combination easy for client teams to use. This works when the acquired asset truly becomes infrastructure. It works less well when it remains another name on the agency chart or when clients cannot see one accountable operating system.
What failed first
The creative reputation did not suddenly evaporate. McCann, FCB and Weber Shandwick continued collecting industry honors. The first visible failure was more prosaic: principal media buying. IPG said competitor scale or offerings cost it existing and prospective assignments. Its full-year 2024 revenue before billable expenses declined 2.3 percent to $9.19 billion. Operating income fell. A $232.1 million goodwill impairment landed on digital specialist agencies. Huge and R/GA were marked for sale.
The 2024 ledger / USD
A large company, with the pressure hiding in the smaller lines
The problem was not that IPG lacked data or talent. It was that media had become a scale market while technology demanded continuing, centralized investment. A client comparing holding companies was comparing buying leverage, identity coverage, commerce signals, automation and the ease of assembling all of them. A federation built to protect difference now had to prove sameness where sameness saved money.
That changed management's mind about independence. Another round of internal simplification could remove cost, but it could not instantly create a larger media pool or spread technology investment across twice as much revenue. In December 2024, Omnicom agreed to acquire IPG in stock for roughly $13.3 billion. The promised annual cost synergies were $750 million. That number was not a side note. It was a diagnosis.
The sale price and the price of staying separate
The deal closed on November 26, 2025. Former Omnicom holders owned 60.6 percent of the combined company; former IPG holders owned 39.4 percent. The new organization kept the Omnicom name and OMC ticker. Krakowsky became Co-President and Co-Chief Operating Officer. Acxiom's identity capabilities became an anchor for a wider platform, alongside Omnicom's Omni and Flywheel assets.
Then the estimate moved. In February 2026, Omnicom doubled its total cost-synergy target to $1.5 billion, including $900 million for 2026. Savings at that scale do not come from switching the office coffee. They suggest overlapping management, real estate, systems, production, procurement and agency structures. They also explain the conditions under which the playbook hurts: when consolidation removes the people clients trust, creates account conflicts, slows decisions or turns distinct creative cultures into interchangeable departments.
What a smaller company can copy
Four rules from IPG, minus the 53,000-person org chart
- Keep customer-facing specialists distinct when their reputation and craft are the reason clients hire them.
- Centralize expensive, repeatable infrastructure: identity, measurement, procurement, security and production technology.
- Measure whether acquired capability is used across the company, not merely whether it adds revenue on its own.
- Treat scale as useful only when it improves the client's result or lowers the cost of delivering it.
IPG's story is tempting to read as a defeat. It is more useful as a migration. The company began with agencies founded in 1902 and 1911, incorporated as McCann-Erickson in 1930, and became Interpublic in 1961. It traveled from print persuasion to television, global networks, media markets, customer databases and AI-assisted production. It kept asking the same commercial question in new forms: what must be shared, and what must remain special?
Harper's original answer was to share the owner and protect the agency. Omnicom's answer is to share far more - data, identity, platforms, purchasing and operations - while hoping the agencies still feel worth naming. That is the experiment now. A holding company built the category by preserving difference. Its successor has to prove that enormous scale can do the same.
Keep going