How to make Meta static ads that perform

A good static does one job. A stranger scrolling on a phone names the product and the benefit in one second. Everything below serves that test, from what goes on the image to how you judge it two weeks after launch.

What every static must include

Use these as a starting checklist. A text-only card, a product demonstration and a customer quote can each work differently.

Copy around the image continues the argument instead of repeating it. Primary text opens with a second proof point (only the first 125 or so characters show before “more”), the headline field carries the offer, and the landing page repeats the hook.

Leave out a second idea, small text, more than one proof point, invented reviews or faces, and any headline that tells the viewer they have a condition (“Struggling with acne?”). That last one breaks Meta's personal attributes rule.

Match the static to how aware the buyer is

Before you pick a format, decide who the static is for and how much they already know. Each concept targets one persona, one value prop and one stage.

Which formats to lead with in your category

S means lead the batch with 2 to 3 concepts. A means include 1 to 2. B means test now and then. Format names link to Motion's gallery of real ads in that format.

Hit rates are Motion's data. The health and wellness column follows Motion's published vertical leaderboard; the other four columns are our placement, based on the overall data and category judgement. Motion's free PDF has the exact table for each vertical.

• Text-only statics have the highest hit rate of any asset type, and product image with text comes next. Both beat UGC.
• Offer-first banners produce the most winners in absolute terms, but the data window included Black Friday, so weight them by season.
• Before and after leads where Meta allows it (cosmetics for 18+, cleaning results) and is risky for supplements and anything body-related.

Check every static before it ships

Zoom to 200% and run each static through this list. Anything that fails goes back, however good it looks.

Product

• Label text is exact: product name, variant, net weight and claims.
• Scale is right next to hands, surfaces and other objects.
• Geometry is right: cap shape, silhouette, proportions and pack count match the real SKU.
• Consistent across the batch: the same product looks like the same product in every asset.

Brand

• Audience fit: the person, styling, setting and mood match who the brand sells to. A beautiful image for the wrong audience is a reject.
• Fonts are the brand's own where guidance exists. No “close enough”.
• Colors are the brand palette where guidance exists.
• Tone reads as the brand, not a generic ad with a logo dropped on it.
• Must-avoids honored: banned claims and category restrictions from the brief.

AI tells

• Hands and fingers: count them, check joints and nails.
• Grip: the product is held the way a person would hold it.
• Posture makes sense for what the person is doing.
• Physics: shadows and reflections match, liquid behaves like liquid, nothing floats.
• Background: no gibberish text, melted objects or impossible architecture.
• Faces and skin: no waxy texture, dead eyes or odd teeth. Not clockable as AI in three seconds.

Every time

• Safe zones respected in every placement, and layout holds (centering, graphic structure).
• Legible on a phone at feed size, nothing cut off in any required crop.
• One message and one CTA. Two messages is zero messages.
• No invented claim, award, rating or comparison.
• Offer, price and discount code match the brief character for character.
• Not the AI-default look: golden glow, marble surface, generic bokeh. Any two together is a reject.

Run every static project in 10 steps

About two working days from intake to launch for a first batch, then a two-week loop.

If you only keep three rules: lead with your category's S-tier formats, make every concept visibly different, and judge by spend plus CPA, never by CTR alone.

Questions