Key Takeaways
- Publift's differentiator isn't a single feature — it's a combination of a publisher-first philosophy, genuine transparency (sometimes to their own detriment), and a world-class experimentation and data science function that lets publishers test and validate incrementality rather than taking a vendor's word for it.
- The single biggest mistake publishers make in monetisation is buying into shiny new solutions without a proper A/B testing framework to validate true incrementality — often robbing Peter to pay Paul while adding stack complexity and paying more in fees.
- On identity strategy: publishers need to start with what data they actually have, whether it's accessible and clean, whether they're using it compliantly across all jurisdictions, and what the addressability landscape looks like in their specific markets — before choosing any ID solution from the 200+ now available globally.
- AI is Publift's primary innovation lever — particularly inside their Fuse platform, where they're building AI into the experimentation framework to remove human bottlenecks and accelerate the speed at which publisher tests can scale from 1% to 100% of traffic. But Amy is clear: garbage in, garbage out — robust data infrastructure is the non-negotiable foundation.
- Amy's framework for future leaders is simple: empathy and accountability. Understand where people are coming from so you can walk beside them rather than drag or push them — and make sure accountability runs both ways, not just inward.
Has the core problem Publift was built to solve changed — or just got more complex?
David: Amy, Publift was built on a simple idea — that publishers don't need more tools, they need help making ecosystems work. From your perspective today, is that still the core problem, or has that challenge evolved as ecosystems have become more complex?
Amy: Well, since the early days of Publift, the entire market has been revolutionised. When Publift started, header bidding wasn't even a thing yet. So I'd say the challenge has definitely evolved and there's absolutely more complexity and more tools than ever. I think Publift's obligation to publishers is to understand which of those tools are actually delivering value, and then democratising the access to those solutions without there being a huge dependency on the publisher's internal teams.
Amy: Over the last 10 years or so, we've been through cycle after cycle of the buy-side, sell-side tug of war. I'm not really talking about the SSPs versus DSPs war and everyone eating each other's lunch — I'm talking about media owners and advertisers and the ambition to uncover true value despite that messy middle. I really believe, and I sincerely hope, that in the next 10 years that tug of war will dissolve. And I think that agentic trading is the key enabler for making that happen. We need behaviour change, we need adoption, we need a programmatic revolution, so to say. But this really excites me — there's a huge amount of value to be gained from both publishers and advertisers in this evolution. And I think we need to be cognisant of technology actually getting in the way of that progression. A little shameless plug here for my colleague David Bickle — he wrote a very clean, articulate article about just this a few months ago. I encourage anyone interested in this concept to look that up on LinkedIn and have a read.
Definition — Agentic trading: an emerging model in programmatic advertising where AI agents autonomously negotiate, buy, and sell advertising inventory on behalf of publishers and advertisers, potentially removing much of the friction and intermediation in the current ecosystem.
What makes Publift truly different — and what is its north star?
Paul: What is it that makes Publift truly unique? Obviously from where it started compared to today, there's a lot more competition globally. What sets you apart, and how does that influence the way you think about growth and innovation?
Amy: It's a handful of things — never one thing. And this might sound a little bit cliché, but first and foremost it's absolutely the Publift people. Everyone is just focused on the success of the publisher. This is absolutely the philosophy of the business. We work for the publisher, we do what is best for the publisher, and their needs are always our top priority. And that mentality is thriving across the organisation in every team. Our ability to uphold that core value is why publishers come to work with us and it's also why they stay.
Amy: And then what makes us different to competitors — we're transparent. Of course everyone says that they're transparent, but everyone knows there are different levels of transparency. Some of the most prominent companies in our industry can't give answers to some pretty basic questions, and that's a bit of a concern. But we're committed to transparency at its fullest — sometimes to our own detriment, often actually. But what this does is it builds real trust with the publishers that we work with. It means they're not just customers anymore — they become partners, because they know we're in it with them.
David: Sounds like you're saying there's a people-first approach to the way you're managing.
Amy: People first, customer first, yeah.
Paul: I heard this on a podcast the other day — they were talking about the cultures within large businesses like Amazon and SpaceX, and how the people running those businesses — Jeff Bezos and Elon Musk — each have these really strict north stars. For Jeff Bezos, it's something like "does this help our customers?" and for Elon it's "does this get us closer to Mars?" Every decision is framed from those questions. If the answer is no, deprioritise that thing. Would you say that's similar to you guys — does this help our clients?
Amy: Is this the right thing for the publisher? That's it. That's what we do. You might have just rewritten the Publift mission statement there for me.
Paul: The conversation I heard was around how those two businesses, despite massive size and continued growth, maintain an innovative environment and culture. We see this all the time with enterprise businesses that get so large they start to crumble under their own weight — bureaucracy, politics. Amazon in particular seems to have maintained a certain level of innovation despite being one of the largest companies in the world. So how did they do that?
Amy: Oh my god, there are so many levers. And I think we definitely have — and I alluded to this before — tooling fatigue on the supply side. There are so many different versions of the same thing. Everyone is powering something with AI. More often than not, the sell is that the AI is what makes the thing special. And I'm like, that's an interesting paradox — the commoditisation of that is real. So in our view, something that actually moves the needle is something that delivers genuinely incremental value. A super simple example: ID solutions in markets that run 99% Chrome market share are probably not going to move the needle. There are better ways to win and grow when that kind of dynamic exists. The combination of levers varies hugely by individual publisher and across markets. And understanding which levers belong where — that's the true lever. That's where companies like Publift can bring the know-how to guide publishers through the tooling that's actually useful versus all of the noise.
What are the most common monetisation mistakes publishers make in 2026?
Paul: I picked up something the other day — I got sent to a website and I was on the page for less than 10 seconds before I realised it was the worst ad experience I'd ever seen in my life. From your perspective, where do you see publishers — and ad monetisation partners — getting it wrong? What are the common mistakes in monetisation in 2026?
Amy: I think I saw your LinkedIn post about that. I was like, we could help them.
Paul: Yeah, I was being very careful not to name any names. But from your perspective — what are publishers getting wrong?
Amy: I would say publishers make a big mistake when they take a vendor's word for it on incrementality. Someone comes in and sells something really shiny, and the publisher doesn't have an adequate experimentation or A/B testing framework to actually test it properly. So they buy into a concept, run a test, it looks shiny on the top line — probably because the vendor has done the uplift calculations — and they go, "that looks really nice," and then they keep paying for the thing. And then a huge amount of the time, they're just robbing Peter to pay Paul while also bringing more complexity to their stack. Their vendor relationships get disrupted and they're probably paying more fees along the way.
Amy: I'd say this is the single biggest mistake — not fully understanding the auction dynamics that exist in a supplier stack, and then thinking that everything that brings in a dollar creates more money in its totality. We're really good at solving for this at Publift. We have an incredible data science and yield function, and we've built this testing and experimentation framework that is world class. I couldn't believe it when I joined the company and looked at the things that had been built and the insights they're generating — it's phenomenal. And this tooling is at the fingertips of our entire customer success team. And our enterprise customers can configure their own experiments inside the platform. It's pretty amazing.
Definition — Incrementality testing: a method of measuring whether a new ad technology, partner, or tactic is generating genuinely additional revenue, rather than simply redistributing existing demand. Without it, publishers can't tell if a new solution is growing the pie or just moving money around.
What does a practical identity strategy look like for publishers today?
David: We're in a world of signal loss — privacy regulations, multiple identity frameworks competing for adoption. What does a practical, working identity strategy look like for a publisher today?
Amy: Well, I feel like this has been my answer for nearly every question so far today, but — it depends. I also want to give credit to my time at Adform for how much I learned about the identity landscape. The unbiased view that they have on identity frameworks and solutions there is pretty amazing — it's a really great place to learn.
Amy: But I think first and foremost, publishers need to understand what they have and how it's stored. What are the foundations of the data? Is it authenticated or not authenticated? Do you have cookies, device IDs, IP addresses? What do you have and where does that data actually live — is it accessible, is it clean? And then they need to understand the actual ability to use all of that data compliantly. What are the requirements across all of the jurisdictions where your audience is, and do you meet those requirements? I think these two questions help to create a gap analysis on where a publisher currently sits and what their options are.
Amy: And then the bit that sometimes gets missed is understanding what the existing addressability landscape looks like in the markets where the publisher operates. The way to drive better addressability in Asia looks very different to Europe versus Australia. Browser and device market share vary hugely across markets. The demand for different ID types is very different across different markets. Understanding that spread is really important because it affects what you actually do and the partners you choose to work with.
Amy: But once you've answered all of those questions, publishers should be able to identify the gaps they have and prioritise: what's the shortest path to value when I'm looking at — I think there are 200 ID solutions globally now — which ones are actually worth talking to? And then you can prioritise your effort from there.
Definition — Addressability: a publisher's ability to identify and target users with relevant ads in an environment where third-party cookies are limited or absent. Addressability strategies typically combine first-party data, identity solutions, and contextual signals.
Paul: I like what you were saying around incrementality on the publisher side. It's interesting that on the buy side right now, incrementality testing seems to be a huge focus — I'd love to see a lot more of that on the publisher side. How are you guys using AI right now — whether LLMs, generative AI, agents — what are you building, and how is that helping with the testing framework?
Amy: I would say AI and data are the primary tools that we use when we're innovating. But ultimately, the true innovation just relies on really smart people. Of course, we're looking at our business operations and seeing how we can get more efficient by implementing AI workflows. But the really interesting stuff is in leveraging AI capability inside the way that our platform — Fuse — operates. The experimentation framework I was talking about earlier — we're bringing AI into the framework to essentially remove any human bottlenecks that aren't necessary. We want to be able to predefine what an experiment should look like and then take that experiment from 1% of traffic to 100% of traffic based on a set of predetermined variables or logic, without needing too many human thresholds in the process. This is really exciting — it just means the speed to value can multiply for all of our publishers and we can learn more, faster.
Amy: But the foundation is really important. You can't do any of that without a really robust data infrastructure. Every AI post on LinkedIn is basically "garbage in, garbage out" — and it's absolutely true. Having infrastructure that's clean and can be easily queried with the AI tooling you're choosing to use just opens so many opportunities. My mind is absolutely exploding with the possible applications of the foundations that we're in the process of building. Anyone who says they know exactly what they're going to be doing with AI for the next 12 months — that's nonsense. If you think you know that, you're going to be behind in 12 months, because things are going to fundamentally change every week, every month in the meantime.
What are the real risks of AI in adtech — and what's the lost craft problem?
Paul: I was talking to a friend over in New York who's doing something around AI literacy, and I commented on one of his posts — I said I think I'm using Cursor wrong, because I've been trying to use it and I've been disappointed every time. But when I see what he's using it for, I'm like, this is a specialist skill right now, the deep applications for it. Does it become a broad-based piece of knowledge applied across the whole organisation, or do organisations need AI engineers who are the experts? What approach have you taken at Publift on AI education or training?
Amy: We are just in the thick of it. We have this huge push across the organisation to surface use cases for efficiency — problems that teams have been trying to solve for a long time and haven't been able to because of resource constraints. Bring them to the table, because we're really well placed to address them at speed now. But something that doesn't get a huge amount of airtime from the AI influencer types is the risk.
David: I'm so glad you said that — this is an area I'm really acutely listening to at the moment. Jump on in.
Amy: Integrating AI tooling into an organisation is really dangerous if it's not governed appropriately, and if the security measures surrounding it aren't robust enough. Moving fast is obviously great, but we have an obligation to our customers to keep their data really secure. And we're very focused on ensuring we meet that obligation. So we put this push out across the organisation to use AI — and then there have been a few occasions where we've had to say: "hold on, come back, don't do that bit." Just making sure that everyone is safe and following protocols designed to protect the integrity of our infrastructure.
Amy: The other thing that plays on my mind — and I don't have a firm answer to this — is around foundational knowledge. I always say the best people in adtech have done some kind of work in ad ops. They understand how things fit together. That experience of how an ad actually gets to a page is really valuable. And I fear it's going to be a lost art in the age of AI. All of the grunt work won't be done by humans anymore. So what happens in 10 years when we have people who are a decade into their career and have never actually solved a problem by clicking the right buttons? How do we leverage AI to make things better and faster, while still holding on to the foundational knowledge that's required to build those AI solutions in the first place? This is something I spend too much time thinking about.
David: You're onto something there, Amy. There's a process I think about — aircraft run by Qantas and the big brands across the globe have two people in the flight deck, a copilot and a captain. They have an autopilot connected to so many aircraft systems so the plane can take off, stay in the air, and land. My view is that's an agentic solution with humans in the loop, standing by to make sure that if the automated system doesn't perform, they can go manual and take over. And to come back to your point about risk — those aircraft have been tested through so many rigorous processes. We are moving so fast into the world of AI right now that there are no checks and measures in place. If something goes wrong upstream in one technology platform and is passed down through others — in a family tree concept spread out to all the others — nobody's there to check that what we're getting is accurate. Where's your view on that?
Paul: Do you mean like one of the biggest tech companies in the world using a coding agent that went and deleted a huge portion of the codebase without any oversight?
David: Yeah, exactly. That's a really good case of where things can go wrong really quickly.
Amy: I mean, like I said, I don't have an answer — because the economics of maintaining that foundational skill base just don't stack up. And you see it even with young people — or people new to the industry — coming in. If you take header bidding as an example, or just the general unified auction, it makes absolutely no sense if you don't understand why it was built in the first place. People come in and see GAM and header bidding and Open Bidding and TAM and all these different things running concurrently and they're just like: "this is ridiculous, who would build this?" And it's like — no one built it. It was an evolution, it was iterative. It's almost like — I don't know if you're a religious person — but there's before and after Christ. I think about before header bidding people and after header bidding people. They have a fundamentally different depth of understanding of how the market works and why it is the way it is. And now we're going to have people that were before AI and people that were after AI. And I'm a bit of a purist, and it makes me a little bit sad that a lot of that art and skill base is going to be lost over time.
David: The craft.
Amy: The craft.
Paul: We've already seen this happen, actually — and the before and after is offshoring. Before offshoring and after offshoring. The knowledge levels and experience in ad ops teams where all trafficking has been done offshore is just completely different from what it was before. Going from Telstra Advertising Network — where we didn't do offshoring — our ad ops guys knew so much about how to get tags live, how to troubleshoot things. But then coming to News Corp, where they'd been offshoring for a number of years, it was quite jarring to see the differences. We never really closed that gap. Adtech just had to step in more and provide more support for troubleshooting. And I think in the Australian market in particular, publishers are siphoning out adtech knowledge at a faster rate than they're replacing it. And AI is going to speed that up. You've got people in the Australian market with 15-plus years of experience in publisher adtech, and then people with a couple of years — and the people in the middle are quite thin, in part because there aren't enough promotion opportunities for them to keep developing, so they move into other areas like product management.
Amy: Yeah. Missing the depth. The why.
How does Publift maintain culture and performance as it scales globally?
David: Publift has scaled to over 900 publishers and is expanding internationally. What changes when you take a monetisation business global, and how do you maintain performance, consistency, and culture at scale?
Amy: The main thing — and as obvious as it is — is that not everyone's online at the same time, and that makes things hard. This is the single biggest challenge of globally distributed organisations. And it's not my first rodeo — when I was at Adform, my team was APAC, completely removed from the core of the company in Europe. You just have to be super organised, communicate really well, track and document fastidiously. There's nothing worse than waking up in the morning to a platform bug and there's no one available to help you for eight hours. Clear escalation pathways, and unfortunately for some — a lot of late-night calls. But that operational efficiency and cleanliness is what makes sure the business continues to run and customers get what they need.
Amy: Culture in a globally distributed team is the most nuanced thing. We have the Publift culture that I've talked about, but we also have subcultures across the different offices and teams. And it's very important to me that those subcultures are respected and cherished as much as the broader Publift culture is. Whether our people are in Australia or the Philippines or London or the US or Dublin — we want them to really feel like they're part of something, but that they also have their own thing going on. Our senior leadership team is globally distributed — two members in EMEA, one in Dublin and one in London — and I lean on them a lot to foster culture, ensure inclusivity is felt across the org, and that information is disseminated effectively. That collaboration across the broader leadership team is really important.
Paul: At least you picked three places where the cultures do kind of align — London, Dublin, and Australia.
David: Getting the beer-drinking culture celebrated.
What legacy does Amy want to leave — and what does she want future leaders to take away?
Paul: As we look ahead — what's the impact you'd like Publift to have on the industry? And what's the one lesson you'd like to pass on to the next generation of media and tech leaders?
Amy: I just want to be known as someone who gets things done. Like, if someone at my funeral said "Amy was really good at getting things done," I'd be like — yes, mission accomplished. But I also care really deeply about the industry, and I hope that comes through in the way that I operate. And as I said at the beginning — I just want to be useful. It's my worst nightmare not to be useful. From a legacy perspective, that's who I want to be.
Amy: For future leaders — I would say empathy and accountability are two of the most important qualities in a good leader. Understanding where people are coming from so that you can walk beside them instead of dragging them behind you or kicking them from behind. That deep understanding of people is super important. And then accountability — for yourself, but also two-way. I think a lot of leaders, and I've definitely done this in my career, put so much accountability on themselves and not enough on the people in their team. And if you do that, you're really doing your team a disservice. Yeah — meet people where they are, be honest with them, and check yourself every day: are you being useful?
David: That might be useful. There's a KPI against that somewhere.
Amy: Are you being useful?
David: I don't know — is there a commercial return on that? I'm sure there is. Amy Jansen-Flynn, Publift Managing Director — thank you so much for joining us on the show today. Your journey in the media space is really inspirational.
Amy: Thank you. Thanks for having me. I had a great time.
David: Yeah, it's been awesome. Thank you so much. And to everyone watching or listening — both here in Australia and across the globe — thank you so much for your company. We look forward to you joining us next time on Media Tech Talk.
Up next on Media Tech Talk: Karen Halligan, CEO of OzTAM, Australia's official provider of television audience ratings, on what a typical day in the life of a CEO looks like in the rapidly evolving media industry — and how the organisation has evolved as the industry shifts from broadcast to streaming.
This is an edited transcript of Media Tech Talk, Part 2. The words are David, Paul, and Amy's own — lightly edited for readability (filler words, false starts, typos, punctuation). No claims have been rewritten or generated by AI.