5 ways to lift your catalog ad performance

Catalog ad performance is won in aggregate. A single creative improvement, multiplied across thousands of SKUs and millions of impressions, moves the whole account in a way no individual ad ever could. That is the mental shift: you are not optimizing one ad, you are optimizing a system that prints thousands of them.

The targeting side of catalog ads is largely automated and largely solved. The platform matches product to person well. What it does not do is make the ad worth acting on. That job lives in the creative layer, and it is where the levers below apply. If you are still mapping the basics, what catalog ads are sets the context; this piece is about pushing the numbers up once you are running.

1. Lead with the offer

The most common and most expensive mistake is shipping ads that show only a product name and image. The price is the strongest commercial signal you have, and in a raw feed it is invisible.

Surface it. Show the price, and when there is a discount, show the strikethrough and the sale price together. A viewer who sees a jacket has mild interest. A viewer who sees the same jacket at thirty percent off has a reason to click now. The feed already carries price and sale_price; the work is bringing them into the creative where the eye lands, not leaving them buried in the caption.

This single lever, applied across a catalog, usually produces the largest immediate lift, because it changes the ad from a product display into an offer.

2. Add social proof

People trust what other people have already bought. A star rating and a review count are among the cheapest trust signals to add and among the most effective, because they answer the silent question every catalog ad provokes: is this any good?

If you have ratings, put them on the creative. A "4.8 stars, 2,400 reviews" badge does more to earn a click than another adjective in the title. The data usually already exists in your review platform; the work is surfacing it on the ad rather than hoping the viewer goes looking for it.

3. Test layouts, not button colors

Teams new to creative testing often burn cycles on cosmetic tweaks: a different button shade, a slightly bigger logo. At catalog scale this is noise. The tests that move performance are about concept.

Test the things that change the message:

• Offer framing. "30% off" versus "Save $40" versus "2 for $60."
• Layout and hierarchy. Product-led versus offer-led versus benefit-led composition.
• Background and setting. Studio white versus lifestyle versus bold brand color.
• Brand treatment. How prominent the logo, palette and typography are.

A winning concept does not lift one ad. It lifts every SKU it applies to, which is what makes concept testing worth the effort in a way button-color testing never is. We show what a strong concept looks like end to end in from plain feed to branded ads.

4. Build for each placement

A single square image stretched across every placement is optimized for none of them. Meta feed, Stories, Reels and TikTok each want their own frame, and the difference between a well-fit 9:16 and a letterboxed square is the difference between full attention and a scroll.

Build per placement from a single source product image:

• 1:1 square for feed placements.
• 4:5 portrait to claim more vertical space in feed.
• 9:16 vertical full-bleed for Stories, Reels and TikTok.

The product image is the same; the framing, safe zones and text placement are not. Fitting each placement is tedious to do by hand per product, which is exactly why it gets skipped and exactly why doing it is an edge.

5. Refresh automatically

Two kinds of staleness quietly erode performance. The first is creative fatigue: an audience that has seen the same look for weeks stops noticing it. The second is data staleness: the platform keeps serving a cached image after your price dropped or the item sold out, so the ad shows a promise you are no longer keeping.

Both are solvable. Rotate and refresh creative so audiences see movement, and use cache-busting URLs so that when price, stock or imagery changes, the platform is forced to fetch the new version instead of serving the old one. We go deep on the stale-data half of this in why your catalog ads show the wrong price.

The hard part is doing it per product

Every lever above is straightforward for one ad. The difficulty is applying all of them to every product, continuously, as prices and stock change underneath you. Doing it by hand does not scale past a few hundred SKUs, and hiring for it does not scale at all.

That is the work FeedForce automates on Meta catalog ads: you set the strategy once, as brand rules, and the catalog keeps itself on-brand, offer-forward and placement-fit across every SKU. Your team decides the concept; the system applies it thousands of times.