Five years ago, a competent media buyer could outperform a less competent one with the same creative by being smarter about audience structure, bid strategy, and placement selection. That advantage has mostly disappeared. The major platforms (Google, Meta, TikTok) have invested heavily in machine-learning systems that find the right buyers given enough creative variety to test against. They do it faster, at lower cost, and with less variance than nearly any human media buyer can. This isn’t a controversial claim inside the platforms themselves; their own documentation tells advertisers to give the algorithms broad audiences and lots of creative. It’s the agency category that hasn’t caught up.
Most agencies still sell what they sold in 2018: clever audience builds, refined bid strategies, technical expertise in the platforms’ settings panels. That work used to be where the advantage was. It mostly isn’t anymore. What wins now sits one layer up — in the ad itself. The hook in the first three seconds. The offer. The landing page the click goes to. The conversion experience after the click. Agencies that haven’t reorganized around that shift are spending their hours optimizing the layer that’s been automated, and producing flat numbers as a result.
We built the firm around the new advantage rather than the old one. Creative production is a core service, not an add-on. Conversion design is a core service, not an add-on. Media management still matters — the platforms aren’t fully self-driving, and someone has to set them up, watch them, and intervene when they go sideways — but it’s the smallest of the three lanes by hours, not the largest. The numbers we produce reflect that ordering.
The thesis isn’t a slogan. It produces specific decisions about how the firm runs engagements — which clients we take, what we charge for, and where we say no.
Most accounts we audit are spread across too many channels, too many campaigns, too many audiences. Concentration beats diversification at most budget levels. We’d rather run two campaigns that work than ten campaigns that fill the funnel.
Telling clients to cut spend, not add to it. Refusing to launch on channels that won’t move the number. Pushing back on “we should also try TikTok” requests that aren’t backed by anything.
The platforms have gotten better at finding buyers than any media buyer has. The leverage in 2026 sits in the hook, the offer, the first three seconds, the landing page. Most agencies still spend their hours in audience settings and bid strategies. That’s optimizing the wrong layer.
Spending real time on creative briefs, hooks, and landing pages. We don’t typically sell creative as a standalone service — we solve for it through constant testing. Test the creatives until we have a winner, then test as many variables inside the winner as we can to keep extracting performance.
Most agencies treat brand work and performance work as separate practices, run by separate teams, on separate timelines. That’s how you get a beautifully designed brand book that contradicts what the Meta ads actually say. The two should be built together. The brand is what makes the ad work twice — once when someone sees it, again when they remember it next week.
Selling brand and performance as integrated, not à la carte. Doing real brand work even on performance engagements, or saying no to engagements that won’t allow it.
Across the accounts we manage, we average a 0.0× ROAS in 90 days, with clients seeing roughly a 0% reduction in cost per lead and a 0% lift in conversions in that same window. Fast-cycle verticals run substantially higher.