We’ve gotten very good at buying media people don’t care about

As AI takes on more of the mechanics of media buying, maybe agencies should spend less time defending the buttons we push and more time thinking about what is actually worth investing in.

By Chris Marine, Founder & CEO, Campfire

I was talking with my business partner, David Gogel, recently about platforms like Google. Not really about Google as a company, or even what AI search is going to do to advertising. We were talking about the actual Google search results page and how much of our collective lives we have spent staring at it. For two decades brands have poured billions of dollars into appearing there, agencies have built entire practices around helping them climb higher on it, and marketers have obsessed over every change to the ranking system. Now that familiar experience is rapidly changing as AI becomes part of how we search, and there is something interesting about watching it happen. Nobody really seems to miss the old page.

David made a comparison that has been stuck in my head since. Think about a publication you have read for years, or a local newspaper you grew up with, or a radio show, television program, podcast, creator or even a video game that has become part of your routine. If one of those disappears, you notice. “There’d be an ache in your day because you’d miss that,” he said. I immediately related as I continue and feel at loss when the 11:35pm hour hits and I don’t have The Late Show. More on that from me on a previous podcast episode The Night Late Night Went Quiet. David’s point is that is a very different relationship than the one most of us have with a search results page, and we both got into passionate conversation for a regular Thursday night around why the advertising industry should care a lot more about the difference.

We have become incredibly good at finding places where people are. We can identify what they search for, what they watch, which websites they visit, what stores they walk into and which demographic or behavioral characteristics make them more or less likely to buy something. Then a machine can find millions more people who behave similarly and bid for their attention in milliseconds. It is an extraordinary technical accomplishment. What I’m less convinced of is whether all that sophistication has made us any better at answering a much more basic question: Where do people actually care that a brand shows up?

I came into advertising from television broadcast journalism, so maybe I’m predisposed to think about media this way. Before I was selling it and then later buying it on the agency side, I was helping make it. Media was never an abstraction to me. It was the reporting somebody did, the show somebody produced, the audience that chose to come back tomorrow and the community that formed around all of it. Advertising helped pay for that. Somewhere during the industry’s pursuit of efficiency, though, we got very good at separating the audience from the thing that earned their attention in the first place. A reader became a user. A publication became inventory. A television show became impressions. A neighborhood became a geo-fence. Eventually the media itself became almost incidental to the transaction.

There are very good reasons why this happened. Standardization made advertising easier to buy, technology made it easier to scale and performance data gave marketers more accountability for where their money was going. I’m not interested in rolling any of that back. Campfire buys paid search, social and programmatic media every day, and those channels work. Those channels account for more than 80% of our total media portfolio across clients when aggregated. What I do think we need to get more comfortable admitting is that operating those systems is becoming less differentiated by the year. The platforms are increasingly doing the work themselves.

David had another analogy for this. He compared performance media to passive investing. There are parts of your financial portfolio where you probably don’t need somebody sitting there every day making trades just to prove they are actively managing your money. You put some of it in an index fund, establish what you are trying to accomplish and let the system compound. His argument was that parts of digital media are heading toward the same place. “The machines are actively managing it at a speed in which humans can’t compete,” he said. And where we came out of that is… maybe that’s okay.

If Meta can evaluate more signals than any buyer ever could and optimize toward an objective faster than a person can move a bid, we shouldn’t build an agency model around pretending otherwise. Give the platforms the right inputs. That’s key and often overlooked. Without context behind the strategy the inputs are garbage. Give these platforms good creative. Give them clear objectives and appropriate guardrails. Measure the hell out of the result. But maybe we can stop pretending that a person moving another lever inside somebody else’s algorithm represents the highest form of strategic media thinking. That doesn’t dismiss the buyers role, but the separation between buying and strategy is becoming even more apparent.

Sometimes Meta deserves more credit for the Meta campaign than the agency does. Hard truths.

Again, I don’t mean that dismissively. Knowing how to manage performance media remains critical, just as an accountant still needs to understand accounting software even though the software does a lot of the math. But increasingly I see those capabilities as the price of admission rather than the reason a marketer should hire one agency over another. If your agency can generate a strong return from paid search, good. It should. If it can optimize a social campaign toward the platform objective, great. So can plenty of other agencies, and increasingly so can the platform itself.

What interests me is what happens to all of that human brainpower as the machines get better at the machinery. I think we should put it back into understanding media.

Take the extra time to figure out what your audience actually loves, not simply where it can be found. Find the publication they would be genuinely disappointed to lose. Understand why somebody has listened to the same podcast every week for six years. Know which creator has earned enough trust that the audience will actually listen when they recommend something. Know why a video game has become a gathering place, why a local sports broadcast is still a shared ritual or why a regional news organization matters to the community it covers. Then do the harder work of figuring out what your brand can add to that environment.

I want to be careful here because this can quickly turn into the familiar “old media was better” argument. It isn’t. David actually stopped our conversation at one point to make sure we weren’t heading there. “This isn’t two guys being nostalgic about a bygone era,” he said. He brought up gaming. If the smartest place for a footwear company to show up is inside Fortnite (referencing our own client work), fantastic. That may be a much richer media idea than buying another conventional ad unit around the same audience. The point isn’t print versus digital, linear versus streaming or legacy media versus whatever we decide to call new media this year. It’s the difference between media people value and media people merely tolerate.

Advertising has spent an extraordinary amount of money perfecting the latter. We have become better at interrupting people, retargeting them and finding them again somewhere else. We have optimized the distance between exposure and conversion down to an absurd degree of precision. What we haven’t solved nearly as neatly is how a brand becomes familiar, trusted, culturally relevant or simply worth remembering. Those outcomes are harder to squeeze into a dashboard, which has a funny way of making them seem less valuable even when they may be doing more to determine whether someone eventually buys.

That bias toward what is easiest to measure has consequences beyond any one campaign. Advertising dollars don’t merely buy access to media. They help fund what media gets to exist. Where billions of dollars move influences which publishers hire another reporter, which creators can keep producing, which local stations maintain a newsroom and which forms of entertainment get another season. I don’t think brands should invest in ineffective media simply to keep it alive. That turns the argument into charity, and charity isn’t what I’m talking about. Brands have a responsibility to grow their businesses. What I find strange is the assumption that investing in media people genuinely care about is somehow at odds with doing that.

If a customer has chosen to subscribe to a publication for 10 years, there is information in that relationship. If they spend hours inside a gaming community or trust a particular creator enough to return every week, there is value there too. A brand showing up intelligently in those environments is entering a relationship the audience has already chosen. That feels fundamentally different to me than locating the same person anonymously somewhere else on the internet simply because the impression was cheaper.

It may also require more work. Someone has to call the publisher. Someone has to understand the community. Someone has to develop an idea that doesn’t fit neatly inside the standard ad unit. Someone has to make a judgment about whether the brand actually belongs there. Maybe that’s precisely the point. As AI and automation absorb more of the predictable work in media, the future value of an agency, for ours at least, will be found in the parts that remain inconvenient, subjective and deeply human.

For years the media business has been moving in the direction of making everything easier to transact. I’m increasingly interested in the opportunities that didn’t survive that filter. The ones that take another phone call, another conversation or another month of thinking. The partnerships where the publisher knows something about the audience that isn’t available in a targeting interface. The ideas where the media company is part of creating the experience rather than merely supplying the impressions afterward.

The industry talks constantly about AI taking jobs, but perhaps the more useful conversation is about what work we should be relieved to give it. I don’t particularly want talented strategists spending the next decade proving they can optimize against machines designed to optimize themselves. I want them understanding people, culture, communities and the media that connects all three. I want them figuring out where a brand should take a chance, what it should help create and where its money can produce something more valuable than another efficiently delivered impression.

Maybe that’s where media agencies need to head next. Not away from performance or technology, but toward a clearer division of labor. Let the machines become better at finding people wherever they happen to be. Let humans spend more of their time figuring out which places are actually worth showing up in.

And every once in a while, before adding another channel to the plan, maybe ask the question David asked me.

If this medium disappeared tomorrow, would anybody miss it?

A note on how this was made. These thoughts and words are my own. I did use AI as an editing tool to help structure the narrative, tighten the writing and keep the ideas focused. The experiences, opinions and point of view are mine.