In defense of the media buyer

Blaming media buyers used to be one of the more reliable coping mechanisms in commerce (as both a DTC and agency founder, I can testify).

When performance dipped, the explanation was familiar and socially acceptable. The agency didn’t get the brand. The account was overcomplicated. Someone chased the wrong audiences or scaled too aggressively. For a long time, these explanations weren’t just convenient, they were structurally accurate.

Accounts were complex, manual, and fragile. Small decisions compounded. If things weren’t working, it was entirely plausible that someone had set it up wrong. That’s what makes the current moment awkward. The complaints still circulate, but the conditions that made them true have largely disappeared.

The difference in agency management pre iOS14.5 and why expectations today are changing

Before iOS14.5, Facebook ad accounts rewarded a very particular kind of obsession. You could spend days carving audiences into neat little buckets, stacking exclusions, building 1%, 2%, and 5% lookalikes from different seed lists, and convincing yourself that the difference between profitability and disaster lived somewhere inside the ad set.

Often, it did.

Accounts were split by prospecting tier, funnel stage, creative angle, and geography, each with its own bid strategy and budget logic. Brands ran separate campaigns for interest targeting, lookalikes, and retargeting, all manually balanced against each other. Forget an exclusion, overfeed a warm audience, or misjudge a bid cap, and performance could quietly deteriorate while everyone stared at blended ROAS trying to work out what changed.

In that environment, blaming the media buyer wasn’t deflection, it was diagnosis. There were more than a few ways to misconfigure an account, and enough manual control that execution quality genuinely showed up in results. You could point to a specific decision and say, yes, that mattered.

Meta has spent the last few years dismantling that setup, one “helpful” feature at a time. Campaign budget optimization flattened spend decisions. Automated placements removed channel bias. Advantage+ audiences discouraged explicit targeting altogether. Even creative testing, once a carefully managed process, moved system-side.

What used to be a complex machine you could tinker with is now closer to a black box with a help centre article explaining why this is better. The complexity didn’t disappear, it just moved somewhere you can’t see. And as that happened, the ad account stopped being the most convincing place to look when things go wrong.

Where responsibility quietly moved

As delivery became automated, the role of the media buyer shifted from operator to overseer. The job is no longer about clever segmentation (sort-of) or bid manipulation. It’s about setting guardrails (e.g target cost per add to cart, target CPA etc.) monitoring volatility, and understanding how the system responds to different inputs over time.

This is where creative became central, not philosophically, but mechanically. When targeting is broad and optimization is automated, creative is the clearest signal the platform can act on. Hooks, pacing, format, angles, and iteration volume now do more work than account architecture ever could.

This shift is uncomfortable for founders because it removes a layer of insulation. Weak performance is harder to attribute to execution when the system is largely making the decisions. More often, Meta is simply reacting to what it’s being asked to distribute. The offer, the message, the creative judgment behind it (you’ve heard the spiel).

Automation didn’t remove responsibility. It moved it.

Agencies still matter, but their leverage has changed. They interpret noisy feedback, prevent obvious mistakes, and help guide iteration. They are no longer the primary drivers of performance in the way many founders still imagine, and that’s where the narrative starts to strain.

How teams are restructuring around the new reality of automation

As this shift becomes harder to ignore, the team model around paid media is changing with it.

The old setup was expensive and fragmented. An external agency running media. A separate design team producing assets. A strategist translating between the two. Multiple layers, slow feedback loops, and a lot of cost for what was often incremental improvement.

That model made sense when execution lived inside the account and scale required manpower. It makes less sense when the platform handles delivery and speed matters more than structure.

What’s emerging instead is a simpler core. An in-house, part-time director or senior media buyer who understands the system, owns performance accountability, and stays close to the business and your guardrails. Around them, a lean creative function that blends strategy and execution rather than separating the two.

Technology is doing the rest. AI tools now handle ideation (though quite poorly, so designers still somewhat ‘own’ this layer), editing, and production at a speed and cost that simply didn’t exist a few years ago. Creative volume is no longer constrained by headcount. Testing isn’t gated by agency timelines or production bottlenecks, now that you have software stacks helping design teams deploy bundles of high quality assets at speed, combined with Meta’s ever growing AI suggested creatives (my god these suck).

The irony is that this shift is making teams more creative, not less. When production friction drops, experimentation rises. When iteration is cheap, judgment matters more. The bottleneck moves away from execution and back to taste, which in my opinion adds to the excitement!

As automation continues to absorb mechanics, marketing is quietly returning to the thing it can’t automate. Narrative. Instinct. Creative leadership.

I’ll dig into that more next time, with a potential feature from Mr Don Draper himself.