Fifteen visual formats. Ten shown here, sorted by hit rate. The range is 6.1% to 9.8%, a 1.6x spread from bottom to top. That spread may look narrow, but at scale it is the difference between finding 6 winners per 100 launches and finding 10. Over a year of testing, that gap compounds into dozens of additional scaling assets.

Unboxing leads at 9.8%. This is not a coincidence. The unboxing format has a structural advantage that most formats lack: a built-in reveal delay. The viewer knows something is inside the box. The brain cannot disengage until it sees what. This is the Zeigarnik effect, the psychological tendency to remember and fixate on incomplete tasks. An unboxing ad is an incomplete task until the product appears. That is why it holds attention through the first 3-4 beats where most ads lose the viewer.

The winning unboxing ads in this dataset share a specific pattern: the product does not appear until Beat 4. The first three beats build anticipation (the packaging, the reaction, the context of why this product matters). The reveal is delayed just long enough to create investment. Show the product in Beat 1 and the curiosity loop closes immediately. Delay it to Beat 4 and you have held the viewer through the critical attention window.

Offer-First Banner is the scale format at 8.6%. Where Unboxing wins on hit rate, Offer-First Banner wins on everything else. It accounts for 21.9% of all creatives in the dataset and 29.3% of all spend. That means nearly 1 in 5 ads uses this format, and nearly 1 in 3 dollars flows through it. No other format comes close to this combination of volume and budget concentration.

Why does Offer-First Banner dominate? Three reasons. First, it is fast to produce. Motion graphics and text overlays do not require talent, locations, or post-production coordination. Second, it is easy to iterate. Change the headline, swap the offer, adjust the CTA, and you have a new variant in minutes. Third, it carries the offer in the first frame, which means the ad self-selects its audience immediately. People who are not interested scroll past. People who are interested have already seen the value proposition. This front-loading reduces wasted impressions and increases conversion efficiency.

Demo sits at 8.1%, the workhorse. Its spend use ratio is exactly 1.0, meaning it gets precisely the budget you would expect given its volume. Demo is the format that never surprises. It does not have the highest hit rate. It does not attract outsized spend. But it performs consistently across verticals, across seasons, and across brand sizes. Every creative program needs a workhorse format. Demo is it.

The remaining formats cluster between 6.1% and 7.8%. Headline (7.8%) and Montage (7.5%) outperform Before & After (7.2%) and Listicle (7.0%). Testimonial, often assumed to be a top performer, comes in at 6.5%. Still solid, but below the midpoint. The trust-building advantage of testimonials may be offset by the format's predictability. Viewers have been conditioned to recognize testimonial ads and apply a skepticism filter that newer formats have not yet triggered.

Hit rate tells you how often a format produces winners. Spend use ratio tells you how aggressively brands invest in those winners. These are two different signals, and they do not always align.

Hit rate is straightforward: of all ads in this format, what percentage became winners (10x account median spend and at least $500)? A higher hit rate means more frequent winners.

Spend use ratio is more nuanced. It measures the percentage of total spend a format receives divided by the percentage of total creative volume it represents. A ratio of 1.0 means the format gets exactly the spend you would predict given its volume. Above 1.0 means brands are concentrating budget on it, a signal that when this format wins, brands trust it enough to scale aggressively. Below 1.0 means brands use the format but do not back it with outsized budget.

The intersection of these two metrics reveals four format archetypes. Understanding which archetype a format falls into changes how you should use it in your creative program.

Scale Format: high hit rate, high spend use ratio. Offer-First Banner is the only true Scale Format in the dataset. It wins often (8.6%) and when it wins, brands pour budget behind it (1.3x). This is the format you build your creative program around. It is the foundation.

Niche Winner: highest hit rate, lower volume. Unboxing at 9.8% hit rate but only 2.1% of creative volume. It wins more often than any other format, but brands are not using it at scale. This could mean the format is harder to produce, more category-specific, or simply underexploited. For brands that can execute it well, Unboxing represents the highest-probability bet in the entire dataset. But it requires physical product, packaging, and authentic reactions. You cannot fake an unboxing.

Workhorse: solid hit rate, proportional spend. Demo at 8.1% hit rate and exactly 1.0x spend use ratio. It gets what it deserves, nothing more. Demo is the format that keeps your creative pipeline producing while you search for breakouts in other categories. You should always have demos running. They are your baseline.

Spend Magnet: lower hit rate, disproportionate spend. Celebrity at 5.9% hit rate but 2.1x spend use ratio is the most extreme example. When brands invest in celebrity talent, they commit budget regardless of whether the ad wins. Production cost and partnership obligations drive the spend, not performance data. This does not make Celebrity a bad format. The absolute dollars behind each winner are significant. But the hit rate tells you that celebrity alone does not guarantee a winning ad. The execution still matters.

Coverage Format: moderate hit rate, proportional or below-proportional spend. Testimonial at 6.5% and 1.0x is the archetype. These formats serve a purpose (building trust, providing social proof, filling the middle of the funnel), but they are not where breakout performance comes from. Include them in your mix for coverage, not for breakthroughs.

Everything above is aggregated. It tells you which formats perform well across all 578,750 ads, all verticals, all brand sizes. That is useful context. But if you make creative decisions based only on aggregate data, you are optimizing for an average that may not reflect your category at all.

The formats that win in supplements are not the formats that win in fashion. This is not a subtle difference. The overlap between the top 10 formats in Health & Wellness and the top 10 in Fashion & Apparel is nearly zero. Category intelligence is not a nice-to-have. It is the difference between testing the right formats and wasting budget on formats that win in someone else's vertical.

Health & Wellness is dominated by authenticity formats. Stitch, Reaction Video, and Unboxing (the top three by hit rate) all share a common trait: they feel organic. They look like content someone would create naturally, not something a brand commissioned. In a category where trust is the primary purchase barrier (supplements, wellness products, health claims), the format itself signals credibility.

The spend use ratio tells a different story. Social Post Mockup, Letter, and Celebrity attract the most spend per use in Health & Wellness. These are not the highest hit rate formats. They are the formats that brands trust with scale budgets. When a Celebrity health endorsement works, it works at massive scale. When a Letter format (direct-to-camera, personal appeal) connects, the brand pours budget behind it. The hit rate may be lower, but the ceiling for each individual winner is higher.

Fashion & Apparel is a completely different landscape. The top formats by hit rate (Post-it, Quiz, Stylized Product Shot) are novelty-driven. Fashion audiences respond to surprise and visual disruption. A Post-it note format interrupts the feed because it looks nothing like a traditional ad. A Quiz format creates interactive engagement in a category where passive consumption is the default.

Fashion spend use tells yet another story. Podcast and Unconventional Text Placement lead. These are high-production formats that brands back with budget when they work. Billboard and Text Message formats also appear. Both are visual pattern-interrupts that stand out in fashion-heavy feeds where aesthetics are table stakes.

The practical implication: do not choose formats based on aggregate benchmarks. Start with your vertical. Look at which formats win in your category. Look at which formats attract scale spend in your category. Build your testing calendar around those formats first, then experiment with outliers from adjacent verticals.

Motion classifies ads into visual format categories (Unboxing, Demo, Testimonial, Offer-First Banner, and so on). This tells you what the ad looks like. It does not tell you why it won.

Heista's PatternMap taxonomy goes deeper. When you decode a video, PatternMap does not just classify the format. It extracts the beat structure, the psychological mechanism, and the linguistic patterns that make one execution of that format a 9.8% winner and another a forgettable scroll-past. The format is the container. PatternMap reads what is inside.

Each mapping represents a dimension collapse. Motion sees “Unboxing” as a single category. Heista sees the opening subtype (Curiosity Spike? Discovery Moment? Data Point Start?), the beat progression (how many beats before the reveal?), the psychological mission (curiosity resolution? social proof accumulation? identity validation?), and the linguistic patterns (imperative voice? question loops? sensory language?).

This matters because two Unboxing ads can have the same visual format and completely different performance. One opens with a curiosity hook (“I finally got my hands on this”), delays the reveal to Beat 4, and builds tension through three escalation beats. The other opens with a product shot, shows the unboxing in Beat 1, and spends the remaining beats on features. Same format. Radically different structures. Radically different results.

The benchmark tells you Unboxing hits at 9.8%. PatternMap tells you which Unboxing structures hit at 9.8% and which ones drag the average down. That is the difference between a benchmark and an actionable creative system.

Knowing Unboxing hits at 9.8% is useful context. But if you go produce an unboxing ad based only on that number, you are still guessing. You know the format wins. You do not know which execution of the format wins. The difference between a benchmark and a creative system is the difference between knowing the format and knowing the formula.

The benchmark tells you where to look. Decode a winning ad in that format and you see the internal architecture: the hook type, the beat progression, the psychological mechanism, the proof placement, the CTA timing. Then you can generate your brand's version of that specific structure, not just a generic version of the format.