The conversation was recorded in a small meeting room at the fintech's east London office in early May, over the course of an unusually cold spring morning. The head of paid, who agreed to be identified as Halvor for the purposes of this piece, has been in the role for six years across two companies. His current employer is a UK challenger bank with a substantial paid-media budget and a competitive category. He runs a team of three.
The observation
Ruaridh Kastanis: Let me start with the phrase you used when we spoke on the phone last week. You said, and I wrote it down: "I don't touch the bid anymore." Explain what you mean by that.
Halvor: What I mean is, in the strictest sense, that I have not manually adjusted a bid on any live campaign in this account in the last ten months. Not once. And I have not, in that period, particularly felt the loss. That is a substantial change from what I was doing in this role three years ago, and it is worth some reflection.
Three years ago, in a similar role at a different company, I was spending perhaps a third of my time on bid work. Reviewing keyword-level CPCs, ad-group-level bid adjustments, dayparting adjustments, geographic modifiers, device modifiers. There was, at the time, a live and honest debate in the industry about whether a competent human bidder could out-perform the platform's automated bidding. In 2022, on that account, in that category, the honest answer was: sometimes, marginally, at the cost of substantial time.
By 2024, on the account I inherited when I started this role, the honest answer had shifted. The platform's automated bidding had improved substantially. My marginal ability to out-perform it, on the accounts I audit and the accounts I run, had shrunk to something not worth chasing. I have not touched a bid in a productive way in ten months because the platform is now doing that work better than I would.
What he does instead
Ruaridh: So what fills the time?
Halvor: Roughly four things, in decreasing order of how much time they consume.
The first is creative. Not producing the creative — I have a team member for that, and a small agency retainer — but reviewing, commissioning, and editing. The single largest leverage I have on the performance of this account, this year, has been on the quality and volume of the creative we ship. The platforms are, increasingly, treating the creative pool as their primary optimisation surface, and the returns to a better creative pool are, on our data, higher than the returns to almost any other intervention I could make. Perhaps 30-35% of my week goes on this.
The second is measurement. Which is to say, the internal work of maintaining a picture of what the account is actually doing that is not just the platform-reported picture. I run a small monthly incrementality routine, I maintain a weekly reconciliation between platform-reported and finance-reported numbers, I keep an eye on the deliverability of the conversion signals we feed back to the platforms. Perhaps 20% of my week.
The third is landing pages and site-side conversion. I do not build the pages — that's a product team — but I sit in the meetings where the pages are prioritised and I bring the data on what would produce the largest lift. This is perhaps 15% of my week.
The fourth is the internal politics of the paid function. Explaining to the CMO why the reported ROAS number is not the metric we should be running the function on. Explaining to the CFO why the finance-reported number is not either. Explaining to the head of growth what the reallocation of budget between paid social and paid search over the last quarter has actually done. Perhaps 15-20% of my week, depending on the quarter.
Ruaridh: That's 80%, roughly. What's the remaining 20%?
Halvor: New channel experimentation, mostly. We ran a small connected-TV test in Q1. We're currently in the second round of testing retail media placements. Those experiments take some of my attention. And, honestly, some of it is just — the job has expanded to include things it didn't include three years ago. Answering questions from an analytics team that is more mature than the one I had before. Reviewing MMM outputs from a partner. That kind of work.
The team shape
Ruaridh: What does your team look like?
Halvor: Three people, all of whom I hired. Two of them are what I would call practitioner-analysts — they can run a paid-search account, they can also read a spreadsheet and produce useful analysis. The third is a creative coordinator, whose full-time job is to shepherd creative production and testing across all of our channels.
The shape is deliberate. I hired for people who could think about the paid function as a system, rather than people who could operate the platform UIs. The platform UIs are, by design, becoming less rewarding to operate. The systems-thinking skills are becoming more rewarding to bring to the function. My team looks less like the paid-media team of five years ago and more like a small in-house consultancy.
Ruaridh: How do you hire for that?
Halvor: With difficulty. The traditional paid-media pipeline produces people who are trained to operate the UIs. The universities do not, generally, produce people trained in what the modern paid function actually needs. I have, in the last two years, hired two people who came from analytics backgrounds with no prior paid-media experience, and one person who came from a marketing consulting firm with no in-house experience. All three have been substantially more productive in the role than the candidates I saw who came with five years of paid-media agency experience.
"The industry is going to have a hiring problem for the next three to five years. The people it needs are not the people its pipeline produces. The people its pipeline produces are increasingly hard to place. The gap will close, but not before it's uncomfortable."
The CFO conversation
Ruaridh: Last question. If a head of paid reading this has never sat down with their CFO to discuss the paid function's economics honestly, what should they do first?
Halvor: Pull one number, run one meeting. The number is: contribution margin per pound of paid media spend, over the trailing 90 days. Contribution margin means net revenue after returns and refunds, minus gross COGS, minus fulfilment. Divide by paid media spend in the same window. That's the number.
The meeting is: sit down with the CFO. Show them the number. Ask them what number they had assumed, in their internal model, the paid function was producing. Almost always, their assumption is materially higher than the actual number. The conversation about why that is — the reconciliation between the finance model and the operational reality — is the conversation that unlocks a productive relationship between the paid function and the finance function for the next several years.
I have run this meeting three times in my career. On all three occasions, the CFO's initial reaction has been unpleasant. On all three occasions, the relationship six months later has been the most productive I have ever had with anyone in the finance function. It's a conversation worth having. Nobody has it because it is not, in the short term, comfortable. But the long-term productivity of the paid function, in my experience, depends on it more than on almost anything else.