Fifty percent of AI is overhyped. We just don't yet know which 50 percent.
Andrew McCormick Former Chief Growth Officer, dentsu X
Interviewed by Justin Cooke
Published
Andrew McCormick is the former Chief Growth Officer of dentsu X. He began his career in journalism (more than ten years, including a first job at New Media Age where he convinced his editor to make him social-networking correspondent and reported on Facebook launching its first office outside the US). From journalism he moved into the agency world, including a startup-period role at Essence. His remit at dentsu X covered media, data, platform ecosystems, marketing effectiveness, and the global agency model. dentsu was the only major media network born in Asia and taken into Europe and the Americas, which gave him a distinctive perspective on translating practice across cultures. In this conversation he sets out the agency exists in service of the client principle; the 2.5 million people employed in the UK creative industries, around 7% of the workforce scale; the outcomes, not outputs remuneration model; the mission, ambition, culture framework for a successful agency; the best client you have is the one you already have discipline; the amazing creative in the right place, plus a one-, three-, five-, ten-year learning agenda effectiveness principle; the AI affects everything, unlike digital media or retail media which affected one part of the marketing stack observation; and the 50% of AI is rubbish caution.
Agency exists in service of the client
On the principle.
Agencies should be doing everything in service of the client. They paid the bill, it's their money. If you're not doing something in service of the client, you're doing something wrong.
The agency also has a perspective on innovation and change. As the environment changes, the agency is essentially the research-and-development arm of the client's team, constantly on top of innovation and how to use it to grow brands.
YouTube as the worked example.
Twenty years since Google acquired YouTube for $1.65bn. At the time everyone thought it was a huge sum that would never work because YouTube made no money. YouTube now makes around $40bn a year in ad spend. My son doesn't come down on a Saturday and turn on BBC One; he turns on YouTube.
The job for the agency: understand the platforms, the media behaviour, what creative cuts through, influencers and partners. The agencies that succeed make sure they're at the forefront of that change.
Outcomes, not outputs: the remuneration model
On the model.
The most important focus for agencies is remuneration. It's no longer good enough to be paid for outputs; it needs to be outcomes. How am I driving growth, how can we measure it, how can we be paid for it?
Many agencies are producing stuff without clarity on how effective it's being. When that's reported back, it becomes a cost centre rather than an investment. All media and agency fees should be investments where you know what you're getting back. A good agency today puts forward an outcomes-based model in the proposal: we don't get paid unless this happens.
On the best client.
The best client you can have is one you already have. Doing the best work for current clients leads to retention, expansion, and reference cases. A couple of times in my career I've shut down the agency for new business because we won a couple of new clients and wanted to focus on onboarding them properly. That should always be the first priority.
Pretty good story to tell prospects too: we can't take any more new business right now, we're taking care of our clients. Saying no builds demand.
Mission, ambition, culture: the three pillars
On the model.
A successful agency has a mission. At Essence the mission was to make advertising more valuable to the world. That informed everything. For people receiving ads: don't interrupt them; make whatever they encounter a good experience. For clients: no wastage, as much working media as possible.
An ambition that's achievable but takes hard work to reach (top-three fashion agency in the UK, top-five digital agency in Europe). So when you reach it you can celebrate.
A good culture. Beyond platforms and marketing, people is the most important thing. Create a culture of innovation. People come to work with a smile and want to do good work. Half-way to winning.
What still drives effectiveness: amazing creative in the right places
On the fundamentals.
Amazing creative in the right places. The creative could come from the in-house team or a creative agency. Where it lives might come from the media agency or in-house media team.
Where it lives can be anywhere: product placement in a film, point of sale, retail media, delivery vans, gaming. The creative spark really understanding the brand and how it can gain attention, plus being in the right places.
On simplification as the agency service.
The job of the agency is to say what matters for the client. There are many metrics but they aren't all key metrics. The short-term measurement debate misses the point. If you're serious about growing the client, you need a one-, three-, five-, ten-year measurement plan. What's the learning agenda? What do we want to learn over one year, over three, over five, how does it inform where we put marketing spend? A long-term learning agenda makes the ambition real.
AI is the biggest change in our lifetime to marketing
On the question.
I'm not a hype guy, but AI is the biggest change in our lifetime to marketing and to how we work. Embrace it and make sure AI works for you rather than falling behind while competitors gain advantage.
AI is different from previous iterations. Digital media first affected performance marketing. Retail media affects point-of-sale strategy. AI affects everything you do as a brand. Productivity, automation, in-platform optimisation are all changing. We haven't worked out how to apply AI tools to make creative better; still working on that.
On the line.
A lot of AI technology is good and put to work in the right places matters. There's also around 50% that's rubbish and won't be relevant in five years. Working out which 50% is a job for all of us. The old Wanamaker line applied to current advertising: I know half my advertising is working, I just don't know which half.
The question for the board
If only half of AI hype will prove real, what share of our agency relationships ties payment to outcomes versus to outputs we cannot defend?
