Latest / The CMO Podcast with Fexingo: Marketing Leadership, Budgets, and Executive Strategy / Why CMOs Are Building In-House Media Buying Desks
Transcript
- Lucas: Luna, I want to start with a number: two million dollars. That's roughly what a mid-market DTC brand spends annually on programmatic display — not including social, not including search. Just display. And a growing number of CMOs are asking whether that money should stay with their agency or come in-house. Luna: Two million in display alone? That feels high for a mid-market brand. I'd have guessed more like one to one point five. Lucas: It depends on the category. For a brand in, say, home goods or apparel, where customer acquisition cost is compressed and you're competing for the same pool of retargeting audiences, two million is pretty standard. The point is, at that spend level, the agency's management fee — typically ten to fifteen percent — starts looking like a meaningful line item. That's two to three hundred thousand dollars a year, just for the service of placing the buys. And CMOs are looking at that and thinking: I could hire two people for that. Luna: And by 'two people' you mean actually building a mini in-house media desk. Lucas: Exactly. We're not talking about building a full trading floor. We're talking about a desk of maybe three people: a media buyer who knows The Trade Desk or Amazon DSP, a programmatic analyst who handles the data layer, and a senior manager who oversees the strategy and relationships with the supply-side platforms. That team costs around three hundred and fifty thousand dollars fully loaded. So the math pencils out — in year one, you're roughly cash-neutral versus the agency fee. By year two, you're saving money, plus you own the data, the relationships, and the speed of decision-making. Luna: But isn't there a capability gap? A good agency brings years of negotiation leverage, access to preferred deals, brand safety tools that an in-house team might not be able to afford on their own. Lucas: That's the pushback I hear from every agency exec I talk to. And it's valid to a point. A standalone in-house desk won't get the same floor rates from an exchange as a holding company that moves a billion dollars a year. But the trade-off is transparency. When the agency runs the buy, the client often doesn't see the full auction dynamics — what they paid per thousand impressions, what the bid density looked like, whether they're buying adjacent to content they'd rather avoid. In-house, you see everything. And in the current environment, where every dollar is being scrutinized, a lot of CMOs are choosing transparency over absolute lowest cost. Luna: There's also the data ownership piece. When an agency runs the programmatic buy, the data — the audience segments, the lookalike models, the conversion signals — typically sits within the agency's ad tech stack. If you switch agencies, you might not get a clean transfer. Lucas: That's exactly the nightmare scenario that prompts the move in-house. I spoke with the CMO of a mid-market beauty brand who did this switch eighteen months ago. She told me that when they brought programmatic in-house, they discovered their agency had been running their retargeting campaigns against a broad 'shoppers' segment that included people who had visited the site once, three years prior. Not because the agency was malicious — they just weren't paying that close attention. The in-house team cleaned that up in two weeks and saw a thirty percent improvement in return on ad spend. Luna: And that's the kind of story that gets other CMOs thinking. But I want to push on the cost structure a bit. That three-person desk you mentioned — three hundred and fifty thousand dollars — doesn't include the technology costs. The Trade Desk or Amazon DSP, the data management platform, the verification tools. Those can add another hundred thousand or more annually. Lucas: Fair point. So let's update the math. Fully loaded team plus tech stack, you're probably at four hundred and fifty thousand in year one. Against a two-million-dollar spend with a fifteen percent agency fee — that's three hundred thousand. So year one, you're actually one hundred and fifty thousand over the agency cost. The break-even comes in year two, when the agency fee savings accumulate. Forrester published a report last quarter that analyzed a dozen brands that made this shift and found that the median break-even was at month thirteen. So it's not an instant win, but within a reasonable horizon. Luna: And what about the cultural side? You're asking a marketing operations person to suddenly become a media trader. That's a different skill set. Lucas: It's a huge shift. The best in-house desks I've seen hire from agency-side programmatic teams, not from the brand side. They bring people who are used to the speed of real-time bidding, who know how to read a bid landscape report, who understand frequency capping and viewability thresholds. The CMO of that beauty brand told me they hired their programmatic lead from a holding company agency — a person who had managed eight figures in spend across multiple clients. That person didn't come cheap, but they knew the ecosystem inside out. Luna: So the CMO has to be willing to pay market rate for that talent, which can be higher than typical marketing salaries. That's a tough conversation with the CFO if the whole pitch was cost savings. Lucas: It's a tough conversation, but the better pitch isn't cost savings alone. It's about capability, speed, and data ownership. The CFO should hear: we will spend the same or slightly more in year one, but we will own our audience data, we will be able to test creative variations in hours instead of weeks, and we will have full visibility into where every impression dollar goes. That's a value proposition that goes beyond the P&L. And frankly, in a world where third-party cookies are functionally dead and everyone is scrambling for first-party data, owning your programmatic data is an asset. Luna: And this is exactly the kind of conversation that makes the CMO role more strategic. It's not just about brand campaigns anymore; it's about infrastructure. Lucas: That's the thread I want to pull on. If these marketing conversations have sparked something you've actually used in your own planning or budgeting, that's the whole point of the show. We deliberately don't run ads on these episodes — no sponsors, no pitches. If you find value in that approach and want to support the show staying independent, you can find us at buy me a coffee dot com slash fexingo. It's a small gesture that keeps this exactly what it is: a conversation between two people who care about the craft. Luna: Absolutely. No pressure, just a way to keep the lights on without selling your attention. Lucas: Exactly. So back to the in-house desk. Once you've made the move, the next question is: what does the relationship with the agency look like now? Because you're not firing the agency entirely — you're just pulling programmatic in-house. Luna: Right, the agency might still handle creative, social, search, maybe offline media. So you're creating a split, which can strain the relationship. Lucas: It absolutely can. I've seen agencies react in one of two ways. Some see it as a loss of trust and become less collaborative. Others see it as an opportunity to specialize — they double down on the areas where they still add value, like high-touch creative or complex measurement. The CMO I spoke with said her agency actually became more responsive after the split, because they knew they were being evaluated on a narrower scope. Luna: That makes sense. If the agency is no longer the 'media partner,' they have to prove their worth on the remaining pieces. It probably also affects the compensation model. Lucas: Yes, typically the agency retainer shrinks, but the performance incentives might increase. Some CMOs are moving to a model where the agency gets a higher bonus if the overall marketing mix — including the in-house programmatic buy — hits the revenue target. That aligns incentives. But it requires a level of data sharing and trust that not every client-agency relationship has. Luna: Is there a scale threshold where this doesn't make sense? If you're only spending five hundred thousand on programmatic, building a desk seems harder to justify. Lucas: I would set the floor at around one million in annual programmatic spend. Below that, the math really doesn't work — you're better off with an agency or even a managed service from the DSP itself. But above that, it becomes a viable conversation. And it's not just about cost. Some CMOs are doing it for speed: they want to be able to turn campaigns on and off in minutes, not wait for an account manager to respond. Luna: That speed advantage is especially relevant in categories with volatile demand — travel, events, retail during holidays. If you're relying on an agency to adjust bids during Black Friday week, you're losing opportunities. Lucas: Exactly. The other argument I hear from CMOs who have made the shift is around talent development. They say that running a media desk in-house creates a new career path within the marketing organization. Instead of losing their best digital talent to agencies, they can offer them a role that involves real trading and data analysis. It makes marketing more attractive to quantitative people who might otherwise go into finance or tech. Luna: That's an interesting point. Marketing has long struggled to attract analytical talent because the work was seen as soft. Giving people a real trading desk changes that perception. Lucas: It does. And it also changes how the rest of the organization views marketing. When the CFO sees a team that is managing a P&L in real time, with daily bid adjustments and weekly attribution reports, marketing starts to look more like a revenue center than a cost center. Luna: So what's the one piece of advice you'd give a CMO who is considering this move? Lucas: Do a pilot. Don't try to bring everything in-house at once. Pick one channel — say, programmatic display — and run it in-house for six months while the agency handles the rest. Use that period to prove the model, build the data infrastructure, and train your team. If it works, expand. If it doesn't, you haven't blown up the entire media operation. Luna: Piloting also gives you a clean before and after comparison. You can measure the same KPIs — cost per acquisition, view-through rate, brand lift — and have a concrete case for the board. Lucas: Right. And I'd add one more thing: invest in the measurement layer early. If you bring media in-house but you're still relying on last-click attribution from your web analytics tool, you're flying blind. You need a proper attribution model — or at least a media mix model — to understand whether your in-house buying is actually delivering incrementality. Otherwise, you might be saving on fees but wasting on ineffective impressions. Luna: So it's not just a procurement decision. It's a data infrastructure decision. Lucas: Exactly. And that's why the trend is still early. The brands that are doing it well — like the beauty brand I mentioned, or a few direct to consumer apparel companies — they had strong analytics teams in place before they made the move. The brands that try to shortcut that step tend to struggle. Luna: I think that's a realistic take. It's not for everyone, but for the right brand with the right talent, it can be transformative. Lucas: And it's a trend that's only going to accelerate. As ad tech platforms become more user-friendly and as the talent pool of programmatic specialists continues to grow, the barriers to entry will keep dropping. I wouldn't be surprised if, in five years, in-house programmatic is the norm for any brand spending over a million dollars digitally. Luna: That would be a big shift from where we are today. But listening to the numbers, it's hard to argue against the logic. Lucas: The logic is there. The execution is the hard part. But for the CMOs who get it right, the payoff goes beyond the P&L — it's about building a marketing organization that truly owns its outcomes. Luna: And that's a powerful position to be in. Lucas: It is. And maybe that's the real reason to consider the move.