The Agency Model Is Dying. What Comes Next?

The holding company model rested on three pillars: scale buying power, creative production capacity, and information asymmetry. Scale buying power meant agencies could negotiate media rates that clients couldn't access independently. Creative production capacity meant agencies had the talent bench to produce campaigns across channels. Information asymmetry meant agencies knew more about the media market than their clients did.

AI is dismantling all three. Automated buying reduces the value of scale negotiation. Generative AI compresses creative production timelines and costs. And data transparency — when clients actually demand it — eliminates the information gap that agencies monetised.

This doesn't mean agencies disappear overnight. But it means the economic logic that sustained holding companies at their current scale is eroding, and the rate of erosion is accelerating.

Here's what most clients don't understand about holding companies: the incentive structure is fundamentally conflicted. Agencies earn money by spending client money. The more media they buy, the more they earn — through disclosed commissions, undisclosed rebates, principal trading positions, and inventory arbitrage.

This isn't a conspiracy theory. It's a business model. And it means your agency's financial incentive is to spend your budget, not to optimise it. When an agency recommends a media plan, the question you should always ask is: would they recommend the same plan if they were paid a flat fee regardless of spend?

I've sat in the rooms where these decisions are made. The conflict isn't theoretical — it shapes every recommendation, every plan, every budget allocation. And until the compensation model changes, it always will.

The best agencies are transparent about this and work hard to mitigate the conflict. But the structure itself is the problem, and no amount of good intention fully overcomes structural misalignment.

The in-housing trend is real, but it's not a panacea. Brands that bring capabilities in-house gain transparency and control, but they also take on recruitment, technology, and management complexity that agencies used to absorb.

The smart move isn't binary. It's disaggregation — keeping some capabilities in-house (data, analytics, first-party audience management), partnering with specialists for others (creative strategy, technical implementation), and eliminating the generalist holding company layer that adds cost without adding proportional value.

This is exactly the model that's growing fastest. Specialist consultancies that go deep on specific problems. Technology platforms that automate what agencies used to do manually. Client-side teams that own strategy and data. The holding company's role as orchestrator-of-everything is being replaced by a more modular, more transparent, more aligned supply chain.

The agency of the future — if we're still calling it an agency — looks nothing like a holding company. It's smaller, more technical, more specialised, and compensated for outcomes rather than activity.

At Galahad, we built the model we wished existed when we were clients. Advisory that's grounded in building, not just recommending. Technology products that solve specific problems — Grail for GEO, Lancelot for media quality, Enable for AI capability building. And a compensation structure that aligns our incentives with client outcomes, not with media spend.

The holding companies will adapt or they'll consolidate into irrelevance. The talent is already leaving — the best people in media, data, and technology are moving to specialist firms, to client-side roles, or to building their own companies. What remains in the holding companies is increasingly the middle: the account managers, the generalists, the process people. That's not where the value is moving.

The transition won't be clean or fast. But the direction is clear, and the organisations that recognise it early will be better positioned than those that cling to a model whose economic logic has already expired.