Which ad metrics actually tell you the creative worked.

A campaign ends. The dashboard is full of numbers: impressions, click-through rate, cost per click, conversions, return on ad spend. The instinct is to read all of them as a report card on the ad. Most of them are not that. Most of them are a report card on who you paid to see the ad and how much you paid for the privilege.

That distinction matters because it changes what you do next. If your click-through rate is low, the fix might be a new headline, a clearer photo, a different offer. Or the fix might be that you were showing a lunch special to people three states away. The number alone can't tell you which. You have to know what each metric is actually built to measure before you can act on it.

Why click-through rate lies to you more than it helps you.

Click-through rate feels like the obvious creative scorecard: did the ad make people click? But CTR is downstream of at least three things happening at once: who Facebook or Google decided to show the ad to, how much competition there was for that person's attention at that moment, and what the ad itself said. Change the audience or the bid and the number moves, with the creative untouched.

This is the trap for a plumber running a $30-a-day local campaign. A tight, well-targeted audience of homeowners with an aging water heater will produce a high CTR on a mediocre ad. A broad, lazy audience will produce a low CTR on a genuinely sharp one. The number is real. It just isn't measuring what you think it's measuring.

Treat CTR as a targeting-and-offer signal first, a creative signal a distant second. It earns creative-diagnostic value only when you hold the audience and budget constant and swap nothing but the ad itself, which is closer to a real test than a glance at last week's dashboard. Most small-business owners can't run that kind of controlled test before they spend their budget, and Knowing is built for that moment, after the ad is made and before it runs.

The metric that actually separates creative from noise: how the effect decays.

The single most useful thing you can look at is not one number but a shape: how a metric behaves over time after the ad runs. This is the core insight behind Les Binet and Peter Field's long-running analysis of the IPA's case study databank, roughly a thousand campaigns spanning decades. Brands splitting spend close to 60% long-term brand building and 40% short-term sales activation get the best long-run results: higher market share, stronger pricing power, lower cost of acquiring the next customer, as Growth Method's breakdown of the framework summarizes it.

The reason that split matters for judging a single ad is what it implies about timing. Activation effects, the kind that show up as a sales bump right when the ad runs, are sharp and short-lived. Brand-building effects are slower to appear and last far longer. Long-term brand effects were typically 2 to 4 times larger than short-term activation effects, and could persist for up to five years, according to a separate read of the same research published by INVRSN, which analyzed roughly £1.4 billion in spend.

If a sales spike disappears the week the ad stops running, you didn't build anything. You rented attention.

Practically: a jump in sales during the campaign and gone the week after is a sign the ad was working as a coupon, not as an argument for your business. If interest, direct search for your name, or repeat visits keep climbing after the ad stops, that's a sign the creative did something a discount code can't: it made people remember you.

Sales metrics are the easiest to fake and the shallowest to trust.

James Hurman and Peter Field ran a related but separate analysis in 2019, studying nearly 5,000 award-entrant case studies from 2011 to 2019 to build what they called the Creative Effectiveness Ladder, described in detail by Criterion Global. The ladder runs six levels: short-term sales spikes at the bottom, then measurable behavior change, then brand perception shifts, business growth, and finally cultural icon status at the top, described in detail by Criterion Global. The point of the ladder isn't the top; it's the ordering. Sales-based metrics sit at the bottom because they're the easiest signal to produce and the easiest to fake with a discount, a seasonal tailwind, or dumb luck. Behavior change and brand-building effects sit higher because they're harder to manufacture and harder to fake by accident.

The practical lesson isn't that sales don't matter. It's that a good sales number, on its own, doesn't tell you whether the creative caused it or whether the price cut, the season, or the competitor's stockout did. A sales metric read in isolation is close to noise. Read alongside a brand-lift survey, a search-volume trend, or a holdout test, it starts to mean something.

What's actually worth checking, and in what order.

Start with reach and frequency only as a sanity check, not a verdict: did enough of the right people see the ad enough times to have a fair shot at noticing it. Then look at engagement quality over engagement volume: video completion rate, time spent, or saves tell you something closer to whether people found the ad worth their attention. For static ads, that's comments, shares, or how long someone lingers before scrolling past. Either way, it's closer to a creative question than a raw click count is.

Then look at the shape over time, not the total. Did whatever lift happened hold for a few weeks after spend stopped, or did it vanish the day the budget did. That decay curve is doing more diagnostic work than almost any single-day number.

Finally, treat cost-per-result and return on ad spend as budget-and-market metrics, not creative metrics, unless you've run a real test: same ad, same audience, same spend, only the creative changed. eMarketer's data, cited by Criterion Global, shows the industry moving this direction on its own: 53.5% of US marketers used marketing mix modeling by 2025, and 47% of buy-side decision-makers planned to lean more on attention metrics, up from 36% two years earlier. The industry has moved past reading a single cost-per-click number as a creative verdict, and the math backs that up. You shouldn't either.

The honest limit, and what to do with it.

There's a reason no single metric fully answers "was this ad good." John B. Ford's 2020 review in the Journal of Advertising Research frames quantifying creativity itself as still an open problem in the field, one academics and practitioners are narrowing but haven't closed. If researchers who study this full-time can't reduce creative quality to one number, a dashboard with a single green arrow can't either.

That's not a reason to skip measurement. It's a reason to stop expecting one number to carry the whole verdict, and to check the shape of the effect, the timing of the decay, and whether the metric could be explained by targeting or budget alone before you credit or blame the ad itself. The metrics that matter are the ones that survive that question. Most of the ones on your dashboard don't.