Joe Fraser is the founder and leader of the media agency Half Dome. Starting the business at just 25 years old, Joe has grown the agency from a “back of house” implementation partner to a full-service independent powerhouse with over 60 staff and $150 million in billings.
In this episode, Joe and Ellie discuss the journey of building an agency from scratch, the importance of investing in people and culture early—hiring a Head of People and Culture as their eighth employee—and the definition of success beyond the balance sheet.
The conversation then shifts to the current “AI craze” and the battle between the Holdcos and Indies. Joe shares a refreshing perspective on the “AI Slop”—the danger of technology creating non-distinctive, cookie-cutter work—and his theory on the “reverse bell curve” of AI adoption in the workforce, where it empowers the top performers but potentially makes the bottom tier lazier.
They also tackle the thorny issue of media measurement, the risks of agencies “marking their own homework,” and why he believes independence between execution and measurement is critical for true client value.
You can listen to the podcast here:
Follow Managing Marketing on Soundcloud, Podbean, TuneIn, Stitcher, Spotify, Apple Podcast and Amazon Podcasts.
Actively non-sycophantic AI, like when that launches, I mean, I’m in!
Yes.
Transcription (Edited):
Ellie Angell:
Hello everyone. My name is Ellie Angell and welcome to Managing Marketing, a podcast where we discuss the issues and opportunities facing marketing, media and advertising with industry thought leaders and practitioners. Today I’m joined by Joe Frazer, who is founder and leader of the media agency Half Dome, based out of Melbourne. Welcome Joe, and thanks very much for joining me.
Joe Frazer:
Thanks, Ellie. Good to be here.
The Birth of Half Dome: A People-First Approach
Ellie Angell:
I love a good agency birth story. Tell us about how Half Dome came to be. What was your vision at that time and maybe a bit about how things have morphed over the last seven or eight years you’ve been in business?
Joe Frazer:
I wish that our genesis story had more guts, risk involved. I was 25 when we started Half Dome, which is a time in your life where you have very few obligations. We were really fortunate. We started Half Dome to meet a need that we saw in the market, which was that there were a lot of independent media agencies at the time, especially in Melbourne, that didn’t have a huge amount of digital capability. For the first 18 months or so, Half Dome was actually a partner to other indies executing all of their digital work.
The sliding doors moment was when we had an opportunity to pitch alongside one of those agencies for a company called Maurice Blackburn. That pitch kind of changed our direction. It meant that we went from being a back-of-house implementation team to really being front-of-house, leading a lot of the client relationship and developing our capability into that of a real agency.
What did we stand for? It’s funny, our eighth employee was our head of people and culture, which was probably an odd decision at the time—with five actual employees and three founders to then hire the most highly paid person in the business. But it was probably representative of how little we knew how to navigate that space. That was Lisa Lee, who’s gone on to start her own company, Learner.
Ellie Angell:
I worked with Lisa Lee as Lisa Vercoe back in the day. She’s an absolute weapon.
Joe Frazer:
I remember she started facilitating sessions between the three of us founders really early to have a vision, have a mission, what do we want to stand for? We used to say it wouldn’t matter if we were starting an Uber company or a fleet of taxi drivers, what we were trying to achieve was pretty consistent. And that was that we wanted to build a business that people who worked in it remembered as being the most important part of their career. We wanted it to be the place where they not only did the best work, but where they built the networks that they relied on into the future. That ambition on the people side of things has probably been the consistent thread that’s followed us the whole way through.
Ellie Angell:
I love it that you invested so early in something that even large agencies at that point in time were not wholly investing in, and that it’s paid off for you. You scaled up to 60-odd people and 150 million plus in billings. And now you’ve got an equally capable people and culture lead, Renee Murray, who came from a completely different industry.
Joe Frazer:
100%. Her previous role was in education for an outdoor education company running school camps. Prior to that, she was working in retail. A completely different background, a far more operational take on people and culture to what we were used to. She’s brought a pragmatism to the people and culture space that we probably lacked at the time. If I can pick out two people that have had the greatest impact on Half Dome beyond the founders, it would probably be the two heads of people and culture that we’ve had along the journey.
Ellie Angell:
That sends a fantastic message to the industry. People are becoming more and more important in very different ways, and to have that function and to commit to it is fantastic.
The AI Scale Frontier: Indies vs. Hold Cos
Ellie Angell:
The old trope of Hold Co versus Indy was founded quite heavily on the concept of trading at scale. That prevalence is fading, but it has been replaced by a new scale-driven frontier, which is AI at scale. I’m talking about the WPPs of the world with platforms bringing generative, predictive, agentic, or automated AI at scale to media trading, content creation, insight generation, and data. How can smaller agencies compete in this world? What are you finding that still makes a difference with your clients?
Joe Frazer:
I have a perspective on the AI push in our industry that might be different to the mainstream. We talk about AI continuing to deepen why it is we are distinctive from those holding groups.
What we’re offering is an ability to help businesses navigate marketing challenges, primarily through media. What makes a great media plan? What makes a great creative idea? One of the core tenets is distinctiveness. I think AI is definitionally bad at that. To try and outsource things like ideation and planning and coming up with the strategic platform, which some of these big multinationals are trying to do, is definitionally creating something that is non-distinctive. And that’s a really tough value prop in a world where you’re the growth partner of businesses.
If speed and reliability are what is deemed as success, which is probably what the holding group model has shifted towards, then the AI tools are probably going to do those things better than a massive group of people. But if real, deeply distinctive work that can help you gain an unfair share in a marketplace is what success looks like for you, that’s hard to replace. The clients predisposed to go down the route of using AI for a lot of these things are probably the clients that are still working with holding groups anyway.
Ellie Angell:
Has that changed who you’re aiming at as a result of the way things are going in the market?
Joe Frazer:
At a company level, we are clearer than ever on the types of pitches we are good at and the types of businesses we’re good at working with. That’s made our life a bit easier in terms of figuring out where we should place our own efforts.
The New Challenge: Measurement and Oversimplification
Ellie Angell:
We are dealing with so much uncertainty on the side of the clients at the moment—AI, data, job insecurity, economic uncertainty. Are you seeing that uncertainty bleed into the way that people are behaving in pitches on the client side?
Joe Frazer:
The biggest thing that we see is that paid media is experiencing such a widely shifting landscape at this point in time, and the biggest shift happening is measurement.
We had a time where we were measuring all digital channels and saying they were all working because we could measure them, and we weren’t measuring traditional channels. Now we have a quick proliferation of research that shows what effective media does, and we have all these new fancy tools helping us measure across-channel in a way that strips out a lot of the noise. We are probably seeing the weaponization of shallow-level understanding of a lot of these different concepts, resulting in really precise pitches that almost make things seem overly simple. It’s like, “We’ve got a perfect measurement solution, so we just want you to maximize the ROI as per our perfect marketing mix model.”
I actually think that it’s kind of in this really dangerous trough of understanding that happens before true widespread understanding of how to measure and optimize media better. There’s this brutal oversimplification happening, which is almost making things more transactional on the media side.
Ellie Angell:
What’s contributed to that? Is it the sales pitch from MMM? The willingness to put things in neat little boxes for the C-suite? The desperation to attribute and build a bridge between marketing as a sunk cost and marketing as an investment?
Joe Frazer:
I just think that we’re on our way to truly understanding good measurement. The pure measurement players in the market tend to have large teams of data scientists and research teams that are helping build rigor into how they do media measurement. The temptation for agencies to sort of espouse silver bullet solutions is there. These things are complex and hard, and you can’t just plug your data in and say, “Here’s a silver bullet answer.”
If anybody’s walking in with this brutal confidence in any of the measurement solutions out there, saying it solves all of your problems, they probably are either commercially benefiting from that rollout or just don’t understand that they are complex things that need to be fine-tuned over time.
The Ethics of Agency-Run MMM
Ellie Angell:
What are your views on the ethics behind a media agency running its own market mix model for its own client?
Joe Frazer:
Let’s presume, as you say, that agencies are not putting their foot on the scales or trying to over-report the impact of certain things to reinforce benefit for themselves. I just don’t think that’s happening. I do think it’s dangerous just putting one person in the back corner of the office and saying he’s running great high-end market mix models.
Even the people who are investing in big teams, doing it, they have research teams, they have workflows to challenge outputs. They have operational rigor to ensure these models are as accurate as possible and challenged in an ongoing fashion. If you’re talking about $10 million plus worth of media investment, just having one person in the back corner making calls that can shift millions of dollars across different channels, different markets, into different messaging—I just think that is a dangerous model conceptually.
Ellie Angell:
I’ve worked on pitches where things have become incredibly sticky and difficult to unpick because of these kinds of services being built in.
Joe Frazer:
I think the concept of independence across your different parts of your partner ecosystem feels like a more defendable midterm option. Just being able to pull one part out and have everything else continue is critical for businesses these days.
The Human Impact of AI: Optimism and “AI Slop”
Ellie Angell:
You’ve got loads of talented people. What are they telling you? Do they feel these uncertainties we’ve been talking about?
Joe Frazer:
The AI craze is undeniable. The youngest people in our organization are often the ones that are innovating the most in that space. The biggest thing that I would say about that is I think the people who are most engaged in that space are the ones that also have the most hope about the future.
It’s getting harder for 20-year-olds today to have great optimism, but the ones that are the most optimistic are those who are excited by these technology shifts that are reshaping what opportunity looks like. They’re the digital natives who are going to be the most comfortable building a work style around this new technology. I’ve seen this great shift in quite a pessimistic view of the world from younger people in our organization to probably being a more optimistic view of the world for the first time in a while.
Ellie Angell:
That’s really interesting. We want younger people to be engaged in this industry and motivated by it. I think it’s also interesting with young people and this term AI slop, which I’ve heard—this concept of just shitty, AI-driven stuff. There’s a lot of cynicism and discernment about that now. The technology may become the enabler and not the actual creative end game, and that will be driven by young people who know how to discern what is true human creativity versus what is AI slop.
Joe Frazer:
I’ll tell you the part that I like in the AI slop category, but I feel like people don’t talk about it enough. I almost see it as being like a reverse bell curve at the moment. There are outliers in every business who are the people right at the top end who are using AI really effectively. They’ve probably improved their productivity by 20, 30% and improved the quality of work as well.
It’s interesting at the bottom end of the bell curve, the people who AI has made them worse at their job. They’re not reading emails that well anymore. They’re just sending back boring, cookie-cutter crap. There’s a whole bunch of people where AI is making them worse, and I think that the risks of bad AI don’t get talked about enough.
Ellie Angell:
I couldn’t agree more. How do you see that manifesting?
Joe Frazer:
The people using AI to write emails for them or just generate content in the purest sense of the word—not improve their ideas or challenge their thinking, but just do stuff for them that they would otherwise do—they’re the ones probably right at the bottom of that page.
The people that are using AI to help shape ideas, as a partner in crime who can challenge them so that they can come really prepared to meetings or having thought deeply about different concepts—they’re the ones at the top end.
Ellie Angell:
That reminds me of Hillary Badger, ECD at Publicis, who was on the podcast recently. She uses AI as an antagonist, not as something to replace. She fights with the AI to get to a better solution.
Joe Frazer:
Actively non-sycophantic AI. When that launches…
There’s one other interesting use case I’ve seen: subject matter experts who have been able to lean into AI to then move across disciplines while retaining a high level of output. I can think of one or two people who have been able to go from being a great media buyer to leveraging AI effectively within their specialism, and then almost overnight they’ve gone, “I can actually also do really good strategic outputs. I can be a strategist now as well.” They’re the people that I’m bottling up and studying.
Ellie Angell:
I think those people would be weapons in pitches and as client leads, because that is the forward-facing articulation of a successful client lead. If you use AI to empower that and to make those T-shaped skills more rock solid and more real, that’s really interesting. We might be going into an era where the long tail will get longer in terms of the unicorns at the top and the rest.
Horizons: A Brutal Focus on the Client
Ellie Angell:
What’s next for you and Half Dome? What’s big on your radar?
Joe Frazer:
We have just come out of probably our best year yet. And that year was built off a brutal focus on our clients and the work that they’re doing.
If you line up every agency at the moment, they’re talking about what’s next—the new product, the new market, the new hire. We’re just really comfortable that our client mix is strong. We’re growing organically. There’s a lot of demand in the market for a really lent-in media agency who’s got good capability across both performance marketing as well as traditional above-the-line media.
Our strategy for the next 12 to 18 months is going to be more of the same. We will continue to not be distracted by all the shiny new stuff, continue to retain an independent position in the market, and continue to work with clients who are looking for good people, doing good work with teams that have a lot of continuity. That’s really where we’ve built all of our success over the last two years.
Ellie Angell:
I love the candour of that answer. You have to be quite brutal with your current client list and the ways in which you’re working. Getting lost in the shiny stuff can be detrimental to that.
Joe Frazer:
When I talk to other indie owners, some of them speak only about their clients: “Client X had a great year. Client Y has got a really exciting launch on the horizon.” They’re the ones that are consistently growing well. Then on the other side, you go, “How are you going?” And they go, “We’re launching into this or we’re building this or we’re doing this now.” They’re doing a whole lot of stuff, but they are not the ones who are growing quickly. I think it’s probably harnessing that mentality: if our clients are going well, then we’re going to be a good agency.
Ellie Angell:
That makes perfect sense. Hey Joe, it was lovely to have you, always lovely to talk to you. I wish you all the best with Half Dome and your mortgage and your children and everything else going on.
Joe Frazer:
Thanks, Ellie.
