Bob Lord has an opinion on this. And unlike most people who've spent 30-plus years in advertising, he's not being diplomatic about it.

"Traditional agency models slow brands down," Lord told me on the ADOTAT Show, with the calm directness of a man who left advertising entirely, spent eight years at IBM learning quantum computing and neural networks, and came back to find the industry somehow worse than when he left. "They don't actually make them go faster."

Let that sit. The entire agency-industrial complex. The pitches, the retainers, the staffing plans, the quarterly business reviews with 47 slides nobody reads. And the guy running one of the largest independent media agencies in America is saying, out loud, on the record, that the model slows brands down.

He's not wrong. And the reason he's not wrong is the reason nobody wants to talk about it: the traditional agency model was never designed to grow the client's business. It was designed to grow the agency's business. Those are very different things. The industry has spent three decades pretending they're the same.

Let's be honest about what the standard agency relationship looks like in 2026, because the polite fiction has gone on long enough.

An agency partner sells bodies. That's it. That's the model. You pitch for the business, you agree on a headcount. Eight planners, four buyers, a strategist, two analysts, an account director who mostly forwards emails. The client pays a retainer based on those FTEs. The Full-Time Equivalent model. The beating heart of every holding company P&L since the Clinton administration.

And what do those bodies optimize for? Marketing metrics. Click-through rates. Impressions. Marketing Qualified Leads. Cost per acquisition. A whole constellation of numbers that look great in a dashboard and tell you absolutely nothing about whether the company actually sold more of anything.

Lord nailed this distinction cold: "You could have click through rates and you could have a marketing criteria. You could set up a marketing qualified lead. You could have click throughs off your website. But that doesn't necessarily mean that you're actually driving the business of the company."

Read that again. Your agency can hit every single KPI in the deck. Green arrows everywhere, champagne at the QBR. And the client's business can still be flat or declining. The MQLs were up! Great. The sales were down. Does that make any sense?

Lord's answer: "No."

Here's the structural problem nobody in the agency world wants to cop to: the FTE model creates zero incentive to fix this. The agency gets paid whether the client's business grows or shrinks. The retainer hits the bank on the first of the month regardless. You could replace half the team with an AI platform tomorrow and the client's outcomes might actually improve. But the agency's revenue would crater. So nobody does it. The model protects the model. The client's growth is someone else's department.

The agency reports to the CMO. The CFO is across the hall, asking uncomfortable questions about marketing ROI that nobody can answer with precision. And the agency has no structural reason to care, because the CFO doesn't sign the retainer.

Strategy gets thrown over the wall. The agency builds a deck, presents it in a room with too many people, sends the PDF, waits for redlines. The client's team marks it up. Another round. Another round. Three weeks later something gets approved that barely resembles the original idea. Nobody owns the outcome because nobody's comp is tied to one.

That's an agency partner. That is, with minor variations, how this industry has operated for 30 years.

Now here's what Lord is building instead. And it's not a tweak. It's a different species.

"We're actually looking at the legacy models that are FTE based. They're people-based models," Lord said. "Those don't necessarily generate growth for the company. So we're moving our models to much more of a performance-based model, which is driven by a platform and a technology that drives the customer's business."

The key phrase is "drives the customer's business." Not manages the campaign. Not optimizes the media plan. Drives the business.

What does that mean in practice? Lord made it painfully specific: "For every dollar that you spend, you should see a widget being performed at the other end, whether it's a new car, whether it's a new bed being sold. But your dollar as a marketer, as a shareholder, you should show that there's a return on that dollar."

A widget being performed. Not a click. Not an impression. Not a lead score. A car. A bed. A subscription. An actual transaction in the actual economy that shows up on the actual income statement that the actual CFO reads.

And this is where Lord's argument gets genuinely interesting, because he's not just talking about measurement. He's talking about the most dysfunctional relationship in corporate America: the one between the CMO and the CFO.

"This is an opportunity for us to use technology and data in a way where the CMO and the CFO become friends for once in their life," Lord said. "And you take an agency and you move it from being an agency partner to a growth partner. We're actually moving their business forward."

Friends for once in their life. That's funny because it's devastating. The CMO and CFO have been adversaries for decades. One spends money on brand awareness and consideration funnels. The other squints at the P&L and asks what, exactly, all that spend actually produced. They've been having different conversations about the same budget using different metrics in different meetings. And the agency, comfortably collecting its FTE retainer, has had no incentive to reconcile them.

Lord's argument is that the technology finally exists to give them the same scorecard. And when you give the CMO and CFO the same scorecard, you can tie the agency's compensation to it. And when the agency's compensation is tied to actual business outcomes, suddenly the agency starts behaving very differently. It stops selling bodies and starts selling growth.

That's the shift. From agency partner to growth partner. Same words, completely different economics.

Lord even has a historical proof point. "When I was at Razorfish, Mercedes-Benz was my client and we were tied to the sales that were being driven in North America by the number of cars being sold. We had a bonus tied to that."

Cars. Not clicks. Not impressions. Cars rolling off lots. That was the metric. And somehow, between then and now, the industry decided that was too hard and went back to counting eyeballs.

If the growth partner model is so obviously better, why isn't everyone doing it already? Because the current system is extraordinarily good at protecting itself.

Start with the technology. Lord pointed out that agencies have spent billions building proprietary tech stacks over the last decade. But those systems weren't built to make the client's business run better. They were built to make the agency run more profitably. Faster trafficking. Automated reporting. Internal dashboards that justify the retainer. The technology serves the model. The model serves the agency. The client is a secondary beneficiary at best.

Then there's compensation. The FTE model is predictable revenue. Wall Street loves predictable revenue. Holding companies love predictable revenue. Partners love predictable revenue. Performance-based comp means variable revenue. Variable revenue means uncertainty. Uncertainty means your stock price takes a hit. So publicly traded holding companies have a fiduciary incentive to keep selling headcount even when they know it's suboptimal for clients.

And then there's what Lord calls "agency inertia." The phrase came up multiple times in our conversation, and he means it structurally, not as an insult. Agencies are large organizations with established processes, hierarchies, and business models. Changing any one of those is hard. Changing all of them simultaneously, while still serving clients and making payroll, is close to impossible.

"In my mind," Lord said, "the agency inertia model held everyone back from actually innovating."

But here's the part that should scare every agency CEO reading this: the consumer didn't wait. Lord was emphatic on this point. "There is a fundamental shift that's happening around the consumer and where the consumer is doing." Discovery has moved. Buying behavior has moved. The way people find and evaluate products has moved. And while the agencies were perfecting their inertia, "every marketing pillar that we've actually believed in and have thought about in the last 30 years is thrown up in the air."

Every pillar. Not some. Not the digital ones. All of them. And the agencies optimized to sell headcount against the old pillars are now selling a service nobody fully needs anymore.

Look. I like Bob Lord. I think he's one of the sharper people I've talked to on this show. But ADOTAT doesn't do hagiography, so let's pressure-test this.

Is "growth partner" just a rebrand? Agencies have been calling themselves "strategic partners" and "business partners" and "transformation partners" for years. The PowerPoints change. The retainers don't. Lord's counter is that he's actually tying comp to outcomes, which is different. But he also admitted Horizon is "in a kind of hybrid mode" with some clients still on pure FTE retainers. The vision is clear. The execution is in progress.

Can mid-size agencies afford this? Performance comp means you eat what you kill. If the client's product tanks because of a supply chain issue or a bad quarter or a pandemic, the agency's revenue takes a hit for something completely outside its control. Lord can absorb that volatility at Horizon's scale. A 50-person shop in Chicago probably can't. Does the growth partner model only work if you're already big enough to survive the variance?

What happens when the product is mediocre? Lord's model assumes that better marketing and distribution, powered by better technology, will move more widgets. But what if the widgets are bad? What if no amount of AI-optimized media planning can make people buy a car they don't want? The growth partner model ties your revenue to your client's product quality, which is the one thing you have zero control over.

These are real questions. Lord would probably tell you that's exactly the point. That an agency willing to bet on outcomes is an agency that will fight harder to influence product strategy, distribution strategy, and customer experience. That the bet itself changes the behavior.

Maybe. Let's see.

Here's the thing about Bob Lord: he's not saying agencies are useless. He spent his career in them. He came back to run one. He's saying the current model makes agencies less useful than they could be. Less useful than they should be. Less useful than the technology now allows them to be.

The growth partner model is a bet. It's a bet that you can measure business outcomes well enough to tie compensation to them. It's a bet that clients will choose accountability over predictability. It's a bet that agencies can swap technology for headcount and pass the savings on as better performance rather than pocketing the margin.

Lord is making that bet. Out loud. In pitches. On stages at CES. On this show.

"We are out now pitching performance-based models as Horizon Media," he told me. He's putting the FTE proposal and the performance proposal in the same deck and letting the client choose.

The question isn't whether Lord is right. The question is whether anyone else has the stomach to follow him. Because the FTE model is comfortable. It's predictable. It's safe. And as Lord's old boss Ginni Rometty used to say: "Growth and comfort don't coexist."

The agencies that figure that out will survive. The ones that don't will keep polishing their dashboards while the world moves on without them.