How do you know which ad is actually carrying the account?

Why does a scoreboard of ad-level numbers lie?

Not because the numbers are wrong. Because the number an ad shows depends on more than what that ad did. A retargeting ad runs against people who already know the brand, because a warm-up campaign primed them and a conversion campaign already proved some of them buy. When that retargeting ad posts a big ratio, the scoreboard credits the ad. Some of that ratio belongs to the two campaigns underneath it that built the demand in the first place.

Attribution windows compound the problem. Meta's reporting generally credits the last meaningful touch, so the ad someone saw right before they bought gets the purchase, even if three earlier touches did most of the actual persuading. And small spend inflates ratios mechanically: an ad that's spent $200 and closed two big-ticket sales can show a ratio no larger-spend ad could realistically sustain, simply because two sales on $200 is a tiny, volatile sample.

What's the strongest example of this happening?

My own best number. The Moveable Coops account badges a 36.0x lifetime ROAS, and that figure belongs specifically to the 1% lookalike retargeting layer of a three-stage funnel: warm-up engagement campaigns season the pixel, a conversion campaign builds a real buyer signal, and the retargeting layer compounds on both. Read alone, 36.0x looks like one campaign, or one ad, carrying the whole account.

It isn't, and the honest version of the number says so directly: it's the retargeting layer of a full-funnel system, which is exactly why it's high, and the whole funnel's economics get shown on the call, not just the prettiest campaign. If the best number in the whole case study needs that context to be read correctly, an ordinary ad in an ordinary account needs it more, not less.

So what should you actually look at instead of the ranked list?

The unit of analysis isn't "the ad." It's a hook, a body, and a CTA, running a specific angle at a specific funnel stage. Two ads with identical headline ROAS can be doing completely different jobs: one starting a conversation at the top of the funnel, one closing a warm buyer at the bottom. Ranking them against each other on the same number treats two different jobs as one competition.

Decomposing an ad down to hook, body, and CTA, and to the angle and funnel stage it's actually running, is what makes a real comparison possible. Which pain or desire is the hook working from. Which funnel stage was this creative built to serve. What was live upstream and downstream of it when the number was captured. Without that breakdown, you're comparing scoreboard entries that don't share a scoring system.

What does the UK/AU lending account add to this?

Scale, and the same lesson from a different angle. That account has analyzed and systematized more than 13,000 ads under a weekly testing pipeline, and the way that pipeline behaves is consistent: most creatives die in testing, and the winner compounds. Compounds is the operative word. A winning ad in an account that size isn't an isolated genius creative; it's the survivor sitting on top of everything the pipeline already learned and discarded to get there. Crediting the winner alone, without the pipeline underneath it, is the same misread as crediting a retargeting ad without the campaigns that warmed the audience for it.

That's also why "which ad is carrying the account" is often the wrong question on its own, and "which ads had to fail first for this one to work" is the more useful version of it. A scoreboard shows you the survivor. It doesn't show you what the survivor is standing on, and in a regulated pipeline running that many ads, standing on something is close to guaranteed.