At the holding companies, my job was mostly proving our ads worked. At my own agency, the job is making them work. Those are very different jobs.
Here is what the old world looked like. Every client had a full funnel strategy in place: awareness up top, consideration in the middle, conversion tactics at the bottom. The upper funnel was always on, always feeding. So by the time a dollar hit PMax or Advantage+ Shopping, those campaigns were harvesting demand someone else had already built. Conversion volume was healthy, the algorithms had everything they needed, and “optimization” meant nudging a ROAS or CPA target and keeping targeting and creative fresh. The genuinely hard work was incrementality: rigorous matched-market and conversion lift testing to prove that Brand Search, ASC, and PMax were actually adding sales and not just taking credit for them.
Then I started working with challenger brands, and almost none of that transferred.
The dollars for full funnel activation just are not there. Nobody is warming the audience before your conversion campaign meets them. The exact same Google or Meta ad works measurably less hard when there is no awareness or consideration underneath it. And here is the part that took me longest to accept: the standard playbook of ROAS and CPA target adjustments is not just weaker on lean budgets. It is actively harmful, because every tightening starves an already signal-poor algorithm of the little data it had.
The funny part? The thing I used to spend months proving is free now. For a challenger brand, incrementality is not a research question. Every sale is incremental. There is no branded safety net, no pocket of guaranteed demand where you can park budget and report a pretty number. Which sounds terrifying, and is actually clarifying.
What this means for bidding
Target ROAS almost never works in low-volume, volatile accounts, and I have lived the failure mode with real money. A campaign hummed along at a 2.00 ROAS target for weeks, even overachieving it. Then it caught a few bad days and fell into a rut. Spend ramped down as the algorithm lost its nerve, then it would randomly blow through double the daily budget hunting for converters, and after a bad week and change the average ROAS had sunk under 1.50. Here is the trap: once you are in that rut, there is no way to dig out except lowering the target. So you lower it, the campaign rebuilds, and you wait for the next few bad days to knock it back down. It is an endless cycle of death. The real fix was getting off aspirational targets entirely.
The better play for these accounts, on both platforms, is Max Conversions or Max Conversion Value. But you cannot just set it and walk away, because left alone Max Conversions has a design flaw: it throttles itself on the great days and has no idea how to pull back and conserve on the bad ones. Run it hands-off and you get a flatlined ROAS or CPA over any longer window. The strategy only works with active management layered on top, and that layer looks completely different on each platform.
On Google, the manual levers still compound
Performance genuinely moves in the right direction with targeting changes, disciplined negative keyword maintenance, and ad text optimization. None of these is a silver bullet. Together they are a multi-pronged approach that ratchets efficiency up incrementally, month over month, while Max Conversions handles the bidding underneath. The algorithm sets the bids; you keep improving what it is allowed to bid on and what the searcher sees. That division of labor works.
On Meta, the levers that remain are defensive
You can still stop Meta from doing dumb things with your money: turn off the auto-enhancements it keeps switching on, fix where the click actually lands, exclude the customers you already own. What you cannot do anymore is drive performance with your hands. Edits to live ads break them or reset their learning, demographic and inventory targeting changes are so sensitive that touching them usually makes things worse, and Advantage+ treats your inputs as suggestions it is free to ignore. Meta’s message is “let it figure it out.” At holding-company spend levels, fine. In a conversion-starved account, the machine has almost nothing to figure it out with, and watching it flail is soul-crushing for anyone trying to do right by their client.
So on Meta there is exactly one offensive lever left: creative. Extremely diverse concepts, produced and shipped fast, so the algorithm always has something new to find an audience for. I will go deeper on this in future posts, because it deserves its own. But the short version is blunt: if you are not churning out genuinely diverse creative at speed, challenger brands are dead in the water on Meta.
The lever both machines feed on: the landing page
At a holding company, the landing page was somebody else’s department. Brand owned the site, another agency owned the CRO backlog, and my team optimized media to whatever page we were handed. At a challenger brand, the page is yours. And it might be the biggest performance lever in the entire account.
Real example. One of my clients had a product that was the worst converter in their lineup: 0.84 percent conversion rate on paid traffic and a 0.74 ROAS. The textbook media response is to cut bids, tighten targeting, restrict the product. We had tried versions of all of it, because you cannot bid your way out of a page that does not convert. So we stopped optimizing the media and rebuilt the product page instead: honest specs, real photography, proof moved up the page, cleaner and faster on mobile.
Same campaigns. Near-identical clicks and cost in the windows before and after. Conversion rate went from 0.84 to 2.05 percent. ROAS went from 0.74 to 2.27. The product went from 13 percent of the store’s revenue to 31 percent.
Here is why this belongs in a piece about conversion volume: every point of conversion rate is signal. In a starved account, doubling the page’s conversion rate doubles what the algorithm gets to learn from, without spending another dollar on media. And on Google it pays twice, because landing page experience feeds directly into quality, and better quality buys back impression share at the same bid. The page is the one optimization that makes both machines smarter at once.
Challenger brands are not small versions of big brands. They are a different discipline. The agencies that figure that out will do very well over the next few years, because the platforms certainly are not building for them.