---
title: "The Agency Owner's Dilemma: When Success Makes Change Harder"
description: AI is changing how agencies create and capture value. For established owners, the hard part is building what comes next without risking what pays for it.
image: https://www.thinkinginfields.com/hubfs/The%20future%20of%20agencies.jpg
---

[Skip to content](https://www.thinkinginfields.com/blog/the-agency-owners-dilemma-when-success-makes-change-harder#main-content)

![Thinking in Fields petrol](https://www.thinkinginfields.com/hs-fs/hubfs/Thinking%20in%20Fields%20petrol.png?width=651&height=286&name=Thinking%20in%20Fields%20petrol.png)Homepage

- [Home](https://www.thinkinginfields.com)
- [Decision Architecture](https://www.thinkinginfields.com/decision-architecture)
  
    - [Decision Architecture Review](https://www.thinkinginfields.com/decision-architecture-review)
    - [AI & Value Readiness Diagnostic](https://www.thinkinginfields.com/ai-value-readiness-diagnostic)
    - [The Value Creation Formula](https://www.thinkinginfields.com/the-value-formula)
    - [Decombobulated](https://www.thinkinginfields.com/decombobulated)
- [About](https://www.thinkinginfields.com/about)
- [Thinking](https://www.thinkinginfields.com/blog)
- [Contact](https://www.thinkinginfields.com/contact)

- [Home](https://www.thinkinginfields.com)
- [Decision Architecture](https://www.thinkinginfields.com/decision-architecture)
  
    - [Decision Architecture Review](https://www.thinkinginfields.com/decision-architecture-review)
    - [AI & Value Readiness Diagnostic](https://www.thinkinginfields.com/ai-value-readiness-diagnostic)
    - [The Value Creation Formula](https://www.thinkinginfields.com/the-value-formula)
    - [Decombobulated](https://www.thinkinginfields.com/decombobulated)
- [About](https://www.thinkinginfields.com/about)
- [Thinking](https://www.thinkinginfields.com/blog)
- [Contact](https://www.thinkinginfields.com/contact)

![When the client owns the creative ](https://www.thinkinginfields.com/hs-fs/hubfs/The%20future%20of%20agencies.jpg?width=1602&height=762&name=The%20future%20of%20agencies.jpg)

Agencies Essays

# The Agency Owner's Dilemma: When Success Makes Change Harder

![David Finch](https://7528315.fs1.hubspotusercontent-na1.net/hub/7528315/hubfs/raw_assets/public/mV0_d-cms-elevate-theme_hubspot/elevate/images/avatar-placeholder.jpg?width=48&height=48&name=avatar-placeholder.jpg)

 David Finch

October 24, 2026

*AI is changing who creates, controls and captures the value of marketing. For independent agency owners, the greatest challenge may not be recognising the need to change, but finding a way to build tomorrow's business without putting today's success at unacceptable risk.*

#### **September 2028**

Alice stared at her screen.

She was Head of Growth at an established manufacturer of professional power tools and had just launched a campaign for its latest product, an AI-enabled drill.

Yes, a drill.

A drill that could assess a wall, spot hazards and recommend fixings before making a hole.

At 9.20 that morning, she'd briefed Bob, her AI marketing assistant, on the commercial objectives, target audiences, positioning and budget. Bob already had access to the company's brand guidelines, creative assets, customer intelligence and the performance of previous campaigns.

By 9.45, he'd presented several concepts, each supported by recommended channels, audience estimates, investment requirements and projected returns.

One caught Alice's attention immediately. It set a 1950s tradesperson, equipped with a hand drill, ruler and spirit level, alongside a modern tradesperson using the new drill. The old and the new, side by side, for a brand that generations of tradespeople had trusted.

Alice challenged a few of Bob's assumptions and chose it. He modelled the investment, with the recommended £50,000 budget projecting a 185% return over two months, subject to assumptions about conversion, sales and attribution.

She approved it.

At 11.30, the first advertisements appeared across selected social media channels. Bob began testing creative variations and adjusting delivery within the parameters Alice had approved. It would take weeks to know whether the campaign would meet its commercial objectives, but the early signs were encouraging.

Alice had used AI extensively in her previous role, but nothing quite like this. It was as though the organisation's collective knowledge had become an active participant in its marketing.

#### **The agency that wasn't there**

Of course, Alice and Bob are fictional, and September 2028 hasn't happened yet. Bob could misread the market, optimise the wrong measures or produce creative that initially performs well but damages the brand later. Nevertheless, none of it is difficult to imagine as businesses connect more of their intelligence to increasingly autonomous systems.

But what interests me isn't the speed. It's that Alice didn't brief an advertising agency, a creative agency, a media agency or a marketing automation specialist.

She briefed Bob.

That's not to say agencies played no role. External specialists may have helped build the brand, develop the intelligence systems or establish the commercial principles guiding Bob's decisions. They just weren't needed for this campaign, and may not be for the next.

So, the challenge for the thousands of independent agencies built around providing exactly those capabilities is: If clients can increasingly do for themselves what they once bought from you, what will they continue to pay you for?

#### **Where the intelligence now lives**

Let’s look at who holds what in Alice's story. The manufacturer owns the customer relationships, the commercial data and the brand. Bob works within its systems, drawing on its accumulated intelligence. Alice owns the campaign objectives and the authority to approve the investment. The organisation has decided how performance will be measured and how far Bob can act on his own.

The intelligence, and the decisions made from it, now reside inside the client.

And that's a much more significant change than just cheaper campaign assets. For years, agencies created value by knowing things their clients didn't: how media channels worked, how audiences behaved and how the technology fitted together. Clients paid for that expertise as well as for the people who delivered it.

When that intelligence reaches clients directly, not through an occasional AI prompt but through systems connected to their own customers, products and commercial objectives, the relationship changes. The agency no longer necessarily holds the intelligence from which decisions are made, or the authority to interpret it.

This is important because the value of marketing isn't just confined to the activity. It also depends on the judgement behind it, the decisions it informs and the responsibility for the outcome. I've argued before that [when intelligence becomes abundant, judgement becomes scarce](https://thinkinginfields.com/blog/when-intelligence-becomes-abundant-judgement-becomes-scarce). Scarce, though, isn't the same as protected.

Mike Lander's [Q3 2026 Agency Reset](https://piscari.com/agency-reset-q3-2026/) report makes the point plainly: planning, strategy, creative and orchestration are not automatically protected from automation.

Clients still value an outside perspective. Up to the Light's [*What Clients Think 2026*,](https://www.uptothelight.co.uk/what-clients-think-report/) based on 700 client interviews, found that 64% felt their agency could be more assertive or tougher with them. But wanting an agency to challenge your thinking isn't the same as wanting it to run your marketing operation.

#### **Value creation and value capture are different things**

The economics are already shifting, though not in a straight line. After a year tracking how AI, in-housing and more sophisticated procurement are affecting agencies, Lander's conclusion is that AI may make an agency more productive without making it more profitable.

The [Wow Company's BenchPress 2026](https://www.thewowcompany.com/benchpress) findings show why. Among agencies with fee income above £1 million, 68% said AI had helped them deliver faster and 54% that they were delivering more value to clients. Only 16% said it had made them more profitable, and 4% that they were charging more. Meanwhile, 38% said clients had taken work in-house and 32% that clients expected to pay less. BenchPress's own reading is that clients have been the biggest winners so far, even as agency gross profit rose for the first time in four years.

These are survey responses rather than proof of causation, and I explored some of them in [*Running Faster on Melting Ice*](https://thinkinginfields.com/blog/running-faster-on-melting-ice). But they illustrate the distinction that matters here. An agency can help its clients create considerably more value while retaining less of that value for itself.

If clients understand that similar outcomes need less effort, why wouldn't they question the price? Procurement has the commercial responsibility to ask. So do business owners. And as their own AI capabilities mature, more clients will do some of the work themselves.

This doesn't mean marketing becomes less valuable. Businesses still need to understand customers, build brands and generate demand. The uncertainty is around who provides those capabilities and who captures the economic benefit.

And moving to outcome-based fees won't necessarily be the answer. If the agency has limited influence over the client's wider commercial decisions, how confidently can it promise an outcome? And if the client controls the data, intelligence, customer relationships and systems, what gives the agency enough bargaining power to retain its share?

The principle at the centre of Thinking in Fields is:

> *The defining capability of modern organisations is no longer access to intelligence, but their ability to convert intelligence into value.*

For agencies, I think it needs extending. Converting intelligence into value is one challenge, but retaining an economically sustainable share of that value is another. Simply declaring that agencies will become consultancies doesn't answer it. Clients won't pay a premium for strategic advice just because it comes from someone with a different job title, and if they hold the intelligence, they may reasonably ask what the advice is based on.

#### **Why success makes change harder**

Most agency owners can see some version of this. The difficulty is deciding what to do about it while the business is still performing well, and no one knows for how long.

Despite the uncertainty, many agencies are still reporting growth. The Alliance of Independent Agencies' July 2026 [\*Agency Performance Pulse\*](https://www.synergist.co.uk/blog/agency-performance-pulse-2026), based on 87 independents and presented as directional rather than representative, found that 54% had grown revenue over the previous six months and 48% had improved their EBITDA margin. Several respondents described the difficulty of investing in AI while clients expect work to become cheaper.

I've owned and grown businesses for much of my working life, and one thing experience teaches you is that risking what you've built feels very different from risking what you hope to build.

Think about *Who Wants to Be a Millionaire?* A contestant who has just won £250,000 might risk much of it on the next question. Ask the same person to withdraw £250,000 from their savings and stake it on a single answer, and the decision looks very different. The amount is identical. What has changed is whether the money still feels like winnings or has become part of what they own.

Success turns one into the other. In the early years, profits feel like winnings, and reinvesting them is part of the opportunity a founder has chosen. Fifteen or twenty years later, the same profits have become the owner's wealth: a profitable business, a strong team, loyal clients and a comfortable income. Every decision is measured against what they already have.

That isn't necessarily an unwillingness to change. It may be an entirely rational assessment of what's at stake, and a version of what Clayton Christensen called the innovator's dilemma, in which successful businesses make sensible decisions that leave them exposed. For most independent agencies there's an important difference. The capital at risk usually belongs to the person making the decision, and the consequences extend to employees and their families.

I've experienced this myself. In [You Can't Optimise Your Way Out of the Wrong Identity](https://thinkinginfields.com/blog/you-cant-optimise-your-way-out-of-the-wrong-identity), I described running two economic models side by side in an agency I founded, one built around judgement and the other around delivery. Clients told us they preferred the new approach, but whenever the business came under strain, the delivery model pulled us back. It was simply easier to measure, manage and defend internally.

AI doesn't make the decision easier. Growth has always required people and systems to be funded before revenue arrives. The difference is that AI changes what has to be funded: technical capability, specialist hires, integration, data governance and experimentation. It may also reduce employment costs, but usage-based costs rise with activity, and successful automation can become expensive before the revenue model catches up.

So, the established owner must consider investing real money in an alternative whose economics aren't yet proven, and do so while the existing business is still producing profits. The danger isn't that the profits stop. It's that they continue for several years while the conditions supporting them weaken. Revenue can grow and margins can stay healthy while clients acquire capabilities that make the agency progressively less important to their future decisions.

So while the financial statements may look reassuring, the long-term economic position may be changing considerably.

#### **The advantage of being able to fail**

Independent agencies have long argued that their size is an advantage. They're closer to clients and quicker to decide, and many in the Pulse survey expect to win share from the networks for exactly those reasons. Those advantages are real. But AI may be changing what makes scale valuable.

For international groups such as Publicis, WPP and Omnicom, the opportunity isn't simply to employ more people or produce more work. It's the ability to experiment across many clients, industries and technologies at once.

Perhaps most importantly, they can afford to fail.

Most agencies are small. [The Agency by Agency Atlas 2026](https://agencybyagency.com/atlas/agency-size/) estimates that 62% of the UK's 24,747 active agencies employ one or two people, while the 0.6% employing 251 or more generate nearly half the sector's turnover. An experiment costing a few hundred thousand pounds could seriously affect a smaller agency's annual profits and long-term viability. For a global group, it may be one of dozens running at the same time.

That confers a greater capacity to learn, though not a guarantee of using the learning well. Large groups carry legacy structures and competing priorities of their own, and WPP's restructuring shows how much organisational difficulty size can create. Lander concludes that ownership structure alone doesn't settle which kind of agency is better placed.

But the capacity to absorb failure shouldn't be underestimated. If smaller agencies can't afford to experiment on the same scale, is there another way for them to learn?

#### **What if clients helped pay for the learning?**

Many agency owners understandably hesitate to invest in capabilities with no immediate return. But what if that investment was made together with clients rather than in isolation from them?

An agency could work with a valued client on a commercially important problem, explore how AI might help solve it and agree to experiment together, with the client contributing towards the costs. The agency builds capability while reducing its exposure. The client develops something directly relevant to its business. Both learn, and that learning becomes part of the value the relationship creates.

The practice itself isn't new. Paid pilots and test-and-learn budgets are familiar, and Lander recommends starting with a funded client problem and validating the economics before expanding. What interests me is what it does to the owner's dilemma.

First, it changes whose money is at risk. The owner is no longer staking their own savings on a single answer, because part of the stake comes from the relationship. The exposure is reduced rather than removed, but reduced exposure is often what makes an established owner willing to play on.

Second, it gives a smaller agency its own version of what the global groups have. They learn across many clients at their own expense. An independent running funded experiments with three or four clients can build a portfolio of learning it could never afford alone, and the understanding gained across all of them stays with the agency.

The timing may be better than it appears, because most clients are still a long way from Alice. [\*What Clients Think 2026\*](https://www.uptothelight.co.uk/what-clients-think-report/) found that 92% of clients would welcome more AI expertise and guidance from their agencies. On average, clients rated their own understanding of AI at four out of ten, and their agencies' at three. That's an opportunity for agencies that can demonstrate genuine capability, rather than for the sector as a whole. The report notes that many clients doubted their agencies had that knowledge, or would be willing to share it if they did.

Which raises the question most owners will ask. If a client helps pay for the work, who owns what is learned, and can the agency take it to the next client, possibly a competitor?

The instinct to protect intellectual property is legitimate. But the line worth drawing isn't between sharing and withholding. It's between reusable methodology and experience, which the agency can carry forward, and client-specific information and intellectual property, which it can't. That line is rarely clean, particularly if the agency later works for a competitor, so it needs agreeing commercially and contractually before the work starts. What the agency carries forward is where its share of the value is held.

None of this guarantees future revenue. An agency could help a client become capable enough to bring more work in-house. But withholding knowledge that clients can increasingly obtain elsewhere may be the greater risk.

#### **The decision facing the agency owner**

I've seen technological change reshape the economics of an industry before. During my years in printing, digital printing took over many types of work that lithography had dominated. Printing didn't disappear, but there are considerably fewer production businesses than there once were. Some built broader propositions around fulfilment, distribution and installation. Others remained specialists in much more concentrated markets. The lesson wasn't that everybody needed to abandon printing. It was that technology altered the economic choices available to customers and to the businesses serving them.

I suspect agencies will experience something similar, at a much faster pace. Some will prosper under familiar models for years, others will change substantially, and some will consolidate or disappear. Some owners will make entirely rational decisions to preserve profitable businesses. For someone approaching retirement, that may be the right choice. For someone hoping to build a business that thrives for another twenty years, the calculation is very different.

Mike Lander asks agency leaders to keep three questions in mind. How will AI change clients' alternatives and costs to using agencies? What will they still pay your agency to deliver, profitably? And why will they choose you?

These are questions about clients and markets that every owner should answer. I'd add a fourth, about the owner. How much of what you've built are you prepared to put at risk to find out?

There isn't an easy answer, and I don't think there should be.

Think back to Alice. No agency was in the room at 9.20. But someone had shaped how Bob made decisions: the principles that guided him, the limits of what he could do without asking and the way his success was judged. If that someone was an agency, the most important contribution was made long before the campaign, and probably before anyone could be sure it would pay.

Somebody had to fund that work while the existing business was still paying the bills.

> An agency's future may depend on helping clients need less of the services that built the business. The dilemma is paying for that future with the income those services still provide.

 

## Share this post

<https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fwww.thinkinginfields.com%2Fblog%2Fthe-agency-owners-dilemma-when-success-makes-change-harder><https://twitter.com/intent/tweet?url=https%3A%2F%2Fwww.thinkinginfields.com%2Fblog%2Fthe-agency-owners-dilemma-when-success-makes-change-harder><https://www.linkedin.com/shareArticle?mini=true&url=https%3A%2F%2Fwww.thinkinginfields.com%2Fblog%2Fthe-agency-owners-dilemma-when-success-makes-change-harder><https://pinterest.com/pin/create/button/?url=https%3A%2F%2Fwww.thinkinginfields.com%2Fblog%2Fthe-agency-owners-dilemma-when-success-makes-change-harder>[mailto:https%3A%2F%2Fwww.thinkinginfields.com%2Fblog%2Fthe-agency-owners-dilemma-when-success-makes-change-harder](mailto:https%3A%2F%2Fwww.thinkinginfields.com%2Fblog%2Fthe-agency-owners-dilemma-when-success-makes-change-harder)

## Keep reading

### [![](https://www.thinkinginfields.com/hs-fs/hubfs/abundance%20surrounds%20you%20-%20value%20comes%20from%20knowing%20exactly%20where%20to%20strike.jpg?width=1712&height=919&name=abundance%20surrounds%20you%20-%20value%20comes%20from%20knowing%20exactly%20where%20to%20strike.jpg) Strategy Essays Intelligence → Value The AI Scale Race: Are We Competing for Volume Instead of Value?](https://www.thinkinginfields.com/blog/the-ai-scale-race-are-we-competing-for-volume-instead-of-value)

### [![Solid Business Architecture dean't need perfect elements](https://www.thinkinginfields.com/hs-fs/hubfs/Irregular%20Shapes%20working%20together.jpg?width=1536&height=768&name=Irregular%20Shapes%20working%20together.jpg) Strategy Ownership of Intelligence Decision Architecture The Wrong Name for the Right Ambition](https://www.thinkinginfields.com/blog/the-wrong-name-for-the-right-ambition)

[![Thinking in Fields white](https://www.thinkinginfields.com/hs-fs/hubfs/Thinking%20in%20Fields%20white.png?width=228&height=100&name=Thinking%20in%20Fields%20white.png "Thinking in Fields white")](https://www.thinkinginfields.com/)

- [Home](https://www.thinkinginfields.com)
- [Decision Architecture](https://www.thinkinginfields.com/decision-architecture)
- [About](https://www.thinkinginfields.com/about)
- [Thinking](https://www.thinkinginfields.com/blog)
- [Contact](https://www.thinkinginfields.com/contact)

<https://www.linkedin.com/in/drbfinch/>

---

[Privacy Policy](https://www.thinkinginfields.com/privacy-policy) · [Cookie Policy](https://www.thinkinginfields.com/cookie-policy)  · [Terms of Use](https://www.thinkinginfields.com/terms-of-use)

Thinking in Fields Ltd. Registered in England and Wales. Company No. 03694025.

All rights reserved.

```json
{
  "@context" : "https://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "David Finch",
    "url" : "https://www.thinkinginfields.com/blog/author/david-finch"
  },
  "dateModified" : "2026-10-24T16:15:12.456Z",
  "datePublished" : "2026-10-24T16:15:12.000Z",
  "headline" : "The Agency Owner's Dilemma: When Success Makes Change Harder",
  "image" : [ "https://www.thinkinginfields.com/hubfs/The%20future%20of%20agencies.jpg" ],
  "mainEntityOfPage" : {
    "@id" : "https://www.thinkinginfields.com/blog/the-agency-owners-dilemma-when-success-makes-change-harder",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject"
    },
    "name" : "Thinking in Fields"
  }
}
```