“If you don’t make predictions, you’ll never know what to be surprised by.” — Dan Davies

The last few years have been a turbulent time for communications agencies. It feels like every week there’s a new article about agencies’ talent exodus, their inability to retain older talent, or the fundamental wrongness of their business model. Many agencies are already calling 2023 an annus horribilis, and the question is whether it’s a blip – a post-COVID correction, perhaps – or whether it represents something rotten in the foundations of the industry.

The agency model that exists today is fundamentally the one that was created in the 1980s, following the radical reorganisation and consolidation started by Martin Sorrell – first at Saatchi & Saatchi, and then at WPP. But it feels like that model is creaking. There are threats from independents and from management consultancies, and from AI, but more than that it feels like there are no more pips left to squeeze; the headroom for growth is simply not there in the model any more.

Taking an example at random, if you read the 2022 IPG annual report and compare it to ten years ago, what’s changed? (IPG are one of the “big four” agency groups, and have agencies like McCann, FutureBrand, MullenLowe, R/GA and Huge within their stable.) Revenues have grown from $7bn to $9.5bn. That’s an increase of 35%, or a meagre 3.1% CAGR. But salaries have increased at a slightly higher rate (3.6% CAGR), growing from $4.4bn to $6.3bn, and headcount grew in lockstep, growing 3% a year from 43,300 to 58,400.

In other words, across an entire decade, in order to generate 35% more revenue IPG had to hire 34% more people and spend 42% more on their salaries. That’s ten years in which we’ve seen the widespread adoption of mobile technology, hybrid working, the creation of new media channels, and the emergence of big data and AI. In all that time the productivity of IPG has remained entirely static.

The same is true of Omnicom (home to BBDO, DDB, TBWA, and more). $14.2bn of revenue in 2012, $14.3bn of revenue in 2022; $10.4bn of salaries in 2012, $10.3bn of salaries in 2022. It’s the same depressing productivity picture, only without any revenue growth either.

WPP (Wunderman Thompson, Ogilvy, Wavemaker, et al.) is the same: $10bn of revenue delivered with $6bn of staff costs in 2012; $14bn delivered with $8bn of staff costs in 2022. 40% revenue growth, 33% salary growth.

Publicis – home to BBH, Leo Burnett and Saatchi & Saatchi, and rounding out the big four – is identical to its brethren: €6.6bn revenue with €4bn of salaries in 2012; $12.6bn revenue with $8.2bn of salaries in 2022. That’s a 90% growth in revenue that required a 105% growth in salaries.

There are seemingly no economies of scale in this model. Holding companies acquire, consolidate, reorganise and restructure. They grow top-line revenue year after year. But still the same 60% salary-to-revenue ratio stubbornly remains. There are no synergies, no value added by being part of a bigger group; they are exactly the sum of their parts, and no more.

That’s because the fundamental shape of this model is hours-based. Agencies sell their employees’ time. Some do that directly, by submitting timesheets to their clients and being paid a variable amount according to the time spent. Others do so slightly less directly, by pricing things according to output, but they tend to still set their prices according to the time involved, and they’re pricing versus competitive agencies who are selling hours, which acts as an anchor in negotiations.

There’s nothing wrong in the long run with an hours-based model. Look at lawyers and consultants. But if you adopt that model, there are only two routes to grow: increase what each person bills, or increase the number of people. Lawyers have been fabulously successful at doing both, with models that provide for both an enormous number of junior associates doing grunt work and a smaller number of incredibly experienced partners billing upwards of $1,000 per hour.

Within the agency world, though, billing-per-head has proved stubbornly difficult to grow. That’s partly because barriers to entry for the industry are low: the overhead to setting up a new agency is nonexistent, and there’s a significant push factor that encourages people to leave larger agencies. Pubs throughout London (and I assume New York, Paris and Tokyo too) are filled on a nightly basis with conspiratorial chat between creatives and accounts people thinking about going it alone, and many of them take the leap. That increases the supply of agencies in a way that can only have a negative effect on pricing.

It’s also because clients’ bargaining power is huge. In what other industry would clients be able to demand that prospective suppliers spend tens or even hundreds of thousands of pounds producing work for nothing, or next-to-nothing, in competitive pitches? Every agency is hungry, and knows that if they fail, if they demand too much, or if the client simply gets bored or wants a better deal, then they’ll lose out to another even hungrier competitor.

The final pricing pressure is from technology. As technology improves, allowing creatives to do more with less, the benefits of those productivity increases accrue to the clients who are paying by the hour, not to the agencies employing people. (If the customer pays per unit, and productivity increases from one unit per hour to five units per hour, then the seller can sell five times as many for the same cost; if the customer pays per hour, then the customer can acquire five times as many for the same cost.) In the coming world of generative AI, that price pressure will become unbearable.

With pricing having been resistant to growth, what growth the industry has achieved in the past three or four decades has mostly been the result of increasing headcount: by finding new agencies to acquire, and by hiring more people into the agencies that you already own. But can that last forever? The industry is extremely consolidated; there are only a handful of obvious, independent agencies of any size that remain unacquired. (Wieden+Kennedy, Mother, a few others.)

At some point, the number of people in the industry butts up against limitations on the demand for its services. The story of the last few decades has been of ever-increasing growth in demand for creative and communications services: there are more brands than ever before, communicating in more channels than ever before, and relying on agencies to do it.

I don’t think this story will continue, for two reasons. First, the proliferation of brands in the last 15 years was largely a result of that era’s low interest rates and easy credit, and the once-in-a-lifetime emergence of D2C brands. And second, brands of all sizes are showing more willingness to take communications work in-house. It’s unlikely that the next 15 years will feature the same frenzy of new brand creations.

So, the model is rotten. But why not simply adapt it and switch to something more sustainable? That’s much easier said than done. Inertia is real, and it’s hard to change direction in a large organisation. And while the going is good, nobody wants to be the person who kills the golden goose. Huge, a digital agency that’s part of IPG, are a good example. In 2021 they began a process of shifting towards a product-based rather than an hours-based model. Two years on, results are unclear, and this summer they made another round of layoffs and restructured the executive team amid significant commercial pressure from the holding company.

No economies of scale. Limited demand-side growth prospects. A model that delivers the benefit of productivity increases to clients, not agencies. Limited opportunities for further M&A. This all paints a bleak picture of the last decade for the big four, and a bleak picture of the prospects for the industry in general. So, what do I think will happen in the next decade? In keeping with that Dan Davies saying I quoted at the start of the article – that if you don’t make predictions, you won’t know what to be surprised by – I think the next ten years look something like this:

  • Top-line revenue growth for the big four holding companies will be anaemic. Perhaps an average of ~2% CAGR across all of them, with at least one of them static over the decade.

  • Any increases in revenue will come with a proportionate increase in headcount. The ~60% ratio will continue to hold, or may even worsen; perhaps one group will tip towards a ratio in which salaries are more like 75% of revenue.

  • M&A activity from holding companies will certainly decline in value, and will probably also decline in volume.

  • Without easy growth options in the form of acquisitions, and in search of continued bottom-line growth, the holding companies will become increasingly desperate. There will be big headcount reductions, more will be demanded of the staff that remain, and agencies will attempt to increase billing-per-head by working people harder, rather than by finding a way to charge more.

  • As a result of these desperate measures, quality of client service and creative output will decline. This will create a vicious cycle in which it’s harder to maintain fee levels for inferior work, so prices slip.

  • Large clients, prompted in part by this decline in service and in part by longer-term secular trends, will accelerate their in-housing of capabilities that were once the preserve of agencies; the “in-house agency” will become more common.

  • One of the top four will go out of business, or at least stumble so badly that it’s opportunistically acquired by another. Either way, I think the big four will no longer be a four.

  • These effects will trickle down to smaller, independent agencies, who will feel an existential pressure on their models but with less ability to cut-and-squeeze. It will become very difficult (as if it was ever easy!) to run a ~200-person agency.

  • Lots of smaller agencies will, along the way, find new and interesting business models that move away from charging for time and find ways to increase billing-per-head. But this will happen in smaller, more boutique agencies, and the end result might not look recognisably like an “agency”.

These things might not come to pass. My intention is not to literally attempt to predict the future, because that’s a fool’s errand. The main thing is to think: if I understand this industry correctly, how might things come to pass? If they don’t, then I’ve missed something – and I should learn something. Of course, the job is also to make a bet on this, and to be positioned to take advantage of these shifts. To do that, I think I’d bet on those foresighted smaller agencies who are already shifting into different models, moving beyond the selling hours to something more thoughtful and sustainable.