What separates stronger advertisers is not sharper intuition. It’s how their testing cadence works. They create enough new ideas to give wins a chance to appear, and they notice when something starts to take off.

This pattern shows up even when comparing advertisers with similar budgets. Bigger budgets allow more testing, but the pace of launching new ads explains the difference between top-performing brands and everyone else. 

In simple terms, volume helps because it creates more opportunities to get winners. It doesn’t make the average ad better. It increases how often an advertiser runs into something exceptional. 

In a system where big wins are uncommon but crucial, taking more chances matters more than being certain about any individual ad.

Winners are rare — and that’s okay

If spend was more evenly spread out, improving results would mostly be about optimization. The data shows that’s not how it works. Most ads don’t spend, or they spend very little. Meanwhile, a small number of ads receive far more spend than the rest.

This isn’t just a holiday effect. It’s not a sign of bad creative decisions. It’s how performance advertising behaves. 

In this report, we use a strict definition of a winner . In order to become a winner, an ad must spend at least 10× more than the account’s median ad. This helps us find ads that truly stand out, not just ads that run a bit longer.

As accounts get larger, the bar gets higher. Bigger brands run more ads and spend more overall. That raises the level of spend an ad must reach to stand out compared to the account average. Winners grow unevenly as volume increases.

This helps explain a common frustration. When results are predictably inconsistent, it doesn’t always mean an ad that doesn’t spend is “weak.”

It suggests that advertisers are much better off asking themselves why an ad worked than why it didn’t. Because the majority of ads will predictably fail, understanding your winners has significantly greater impact on your account than interrogating “losing” ads.

This is why high hit rates need context. They could mean strong judgment. They can also mean limited testing. The metric itself cannot tell you which is true. Lower hit rates often appear in accounts that test more ideas. More testing naturally produces more average ads as well as more winners.

Hit rate is still a valuable metric. But it is not a proxy for success. It just describes how often rare events happen within an account.

These mid-range spenders play a different role than winners. They don’t spike performance. But they help keep spend stable.

Larger accounts often behave like volatile portfolios. Some ads aim for big gains. Others provide consistency. Winners can drive growth, but they are unpredictable. 

Smaller accounts, on the other hand, depend more on stability. They have less budget for big swings and rely more heavily on sure bets.

Mid-range spenders make high testing volume easier to manage. When teams launch many ads, some will fail quickly. Without stable ads in the mix, that churn can create volatility. Mid-range ads act as a ballast. 

A common mistake is to treat mid-range ads as failed tests. In a healthy account, they connect testing and scaling. They help keep performance steady while new ideas compete for spend.

How many ads should you be testing?

There’s no universal testing volume that’s “best” for all advertisers. The right testing volume depends on budget, team size, and how quickly an advertiser can produce new ideas.

Creative output varies widely across accounts. Some advertisers launch only a few ads per week. Others launch many more. This reflects resources and workflow as much as strategy.

Below are the median testing volumes listed out by vertical and by spend tier.

Looking across both hit rate and spend use, certain themes appear repeatedly.

Hooks that signal immediacy, clarity, or a concrete reason to act tend to surface often. Language tied to price framing, offers, urgency, and product newness reduces the effort required to understand the value of the ad. 

Other hooks rely less on direct offers and more on attention. Curiosity, confessional framing, bold claims, or unexpected statements can interrupt scrolling and create tension or surprise. 

It is also telling what appears less often near the top of these rankings. Broad lifestyle statements and vague benefit claims are common in advertising, but they surface less frequently among winners in this dataset. Messages that delay clarity or require more interpretation may struggle when attention is limited. 

Keep in mind that these patterns are time-bound. The dataset includes heavy promotional periods (BFCM & the holidays), when shoppers are primed for deals and product announcements. 

Hooks and headlines that emphasize immediacy or value naturally align with what shoppers expect during this window. In a different season, it’s reasonable to expect that this looks very different — a point that warrants future analysis.

Text-forward assets create winners more often than many teams expect. Text-only ads, product images with text overlays, and simple GIFs frequently rank well. Their strength is speed and clarity. They can communicate value quickly and allow small framing changes without reimagining entire campaigns. 

Because they are fast to produce, these assets also make experimentation easier. A team can test multiple variations in a short period, which increases exposure to different outcomes. In a system where rare events matter, the ability to try more variations can influence results.

Higher-production assets play a different role. In some cases, they signal credibility or brand quality. But they are slower to change. When iteration slows, the number of ideas tested often slows as well. That does not make these assets ineffective, but it changes how they function during testing.

Structuring your strategy for success

One major takeaway from this research is that advertising — and creative strategy — can be humbling.

For all the planning, optimizing, and iteration, creative performance resists tight control. The system is shaped by human attention, platform dynamics, and competition that never sits still. Outcomes cluster. A few ads carry accounts. Many barely move the needle.

Performance advertising looks less like careful optimization and more like broad exploration. Not blind guessing, but structured searching in an environment where winning is rare.

The role of a creative strategist is not to predict the future perfectly. It’s to build conditions where winning ideas can surface, where signals are noticed, and where decisions are made with evidence rather than attachment. 

There is also something reassuring in the benchmarks themselves. They show that uneven win rates are not a personal failing or a team-specific problem. They are a shared reality across thousands of advertisers. The pressure to “crack the code” often oversells how deterministic performance advertising really is.

The advertisers who win out will not be the ones who eliminate failure. They will be the ones who make it survivable and routine. They’ll keep shipping high volume. They’ll keep adjusting. And they’ll keep looking for a signal in the noise.

*Key Benchmarks Summary
*Full Methodology.
*Download LLM files