One of the most common mistakes we see when we take over an ad account isn't bad creative — it's bad decision-making about the creative. Ads get killed before they've had a fair shot, and underperformers get left running out of habit.
Give every ad a real sample size first
Judging an ad after a few hundred impressions is judging noise. Most platforms need enough spend and delivery to leave the learning phase before performance data means anything. Pulling the plug too early just resets the clock on your next attempt.
Signals that say kill it
Once an ad has had a fair sample size, a few signals reliably point to "stop": cost per result climbing well past your target with no sign of settling, click-through rate meaningfully below your account average, or frequency creeping up while results decline — a sign the audience has seen it too many times.
Signals that say scale it
The opposite signals are just as clear: cost per result holding steady or improving as spend increases, engagement (saves, shares, comments) outperforming your other creative, and results holding up as you widen the audience rather than tapering off.
- Let an ad reach a real sample size before judging it
- Rising cost per result with no plateau → kill it
- Below-average click-through rate → kill it
- Steady or improving cost per result as spend grows → scale it
- Strong engagement relative to your other creative → scale it
The goal isn't to fall in love with an ad or to panic at the first bad day — it's to let the data make the call.
Build the habit, not the guesswork
The advertisers who win long-term aren't the ones with the single best ad — they're the ones who reliably spot which ads deserve more budget and which ones deserve to be retired, week after week.
