Meta Advantage+ Catalog Ads for Large-SKU DTC and Fashion Brands: SKU-Level Optimization After Andromeda
If you run a large catalog on Meta, this piece is about one specific problem: getting products beyond your top sellers in front of the right shoppers after Andromeda reshaped how Meta picks which ad to show. You will walk away knowing why the concentration happens, how to structure product sets so the long tail gets a fair test, and how catalog ads differ from Advantage+ Shopping so you run the right tool for the job.
This runs on the same engine we covered in our Meta Andromeda update breakdown. Here we stay on the catalog side of it: product feeds, product sets, and SKU-level delivery, not campaign objectives.
What Are Meta Advantage+ Catalog Ads?
Advantage+ catalog ads are Meta's automated format for advertising a product feed instead of a fixed set of creatives. You connect a catalog, define a product set, and Meta assembles the ad on the fly, choosing which products to show each person based on their behavior and the signals in your feed. Meta renamed these from dynamic ads, and most media buyers still call them DPA out of habit.
The requirements are simple to list and easy to get wrong: an active product catalog, a pixel or Conversions API firing clean product-level events (ViewContent, AddToCart, Purchase with matching content IDs), and at least one product set to target. Get the content IDs out of sync between your feed and your pixel and the whole thing quietly underdelivers, because Meta cannot connect the product a person viewed to the product in your catalog.
Two modes matter. Retargeting shows people the exact products they viewed or added to cart. Broad audience (prospecting) lets Meta show catalog products to people who have never touched your site, using the feed and behavioral signals to guess fit. The prospecting mode is where large catalogs live or die, because that is where Andromeda decides which of your thousands of SKUs are worth an impression.
Why Does Meta Only Promote the Same 20 to 30 Products From My 1,500-SKU Catalog?
Because Andromeda is a retrieval engine, and retrieval rewards certainty. Before any auction happens, Meta has to narrow tens of millions of ad candidates down to a few thousand in roughly the time it takes a page to load. Andromeda is the layer that does the narrowing, and Meta reports it lifted recall at that stage by 6% and ad quality by 8% (Engineering at Meta, Dec 2024).
For a catalog, each SKU is a candidate. A product with a rich purchase history is a safe bet the model can retrieve with confidence. A brand-new SKU with no conversions is a coin flip, and a retrieval engine tuned for precision does not spend its narrow shortlist on coin flips. Meta has not published this as a catalog mechanic, but it is the pattern we see on large accounts and it is widely reported across the industry: the products that already sell keep getting retrieved, keep converting, and keep reinforcing their own head start, while the long tail never accumulates the data to break in.
The volume context makes it worse. Meta says more than a million advertisers now use its AI tools to generate over 15 million ads a month (Engineering at Meta, Dec 2024). The candidate pool exploded, so the pressure to retrieve only the surest bets went up with it. Your unproven SKUs are competing for a shortlist slot against every other advertiser's proven ones.
The takeaway is not that the automation is broken. It is working exactly as designed. The job is to give it a structure where testing the long tail is the path of least resistance, not a fight against its instincts.
Advantage+ Catalog vs Advantage+ Shopping: Which One Are You Actually Running?
These get mixed up constantly, and running the wrong one is a common reason large-catalog performance stalls. Advantage+ Shopping (ASC) is a full-funnel campaign objective that bundles prospecting and retargeting and hands Meta the audience decision. Advantage+ catalog ads are a product-feed format built to serve the right SKU to the right person. You can run catalog products inside an ASC campaign, but the levers you pull to fix SKU concentration live on the catalog and product-set side, not the campaign objective. Our Advantage+ Shopping guide covers the campaign-type decision in full.
| What you're deciding | Advantage+ Catalog Ads (DPA) | Advantage+ Shopping (ASC) |
|---|---|---|
| Core unit | A product set from your feed | A whole-funnel sales campaign |
| What Meta automates | Which SKU to show which person | Audience, placement, creative, budget |
| Your main lever | Catalog structure and product-set rules | Existing-customer cap and creative pool |
| Best fit | Large catalogs that need SKU-level control | Scaling spend with high creative volume |
| Key data input | Clean product feed plus content-ID-matched events | Pixel or CAPI purchase signals plus creative |
| Where SKU concentration is fixed | Here, via product sets and feed quality | Not here; it inherits your catalog structure |
The short version: if your problem is "the wrong products get shown," that is a catalog problem, and ASC will not fix it for you. You fix it in the feed and the product sets.
How Should You Restructure a 1,000+ SKU Catalog After Andromeda?
The goal is to stop treating your catalog as one undifferentiated pile and start giving Andromeda smaller, cleaner pools where a new SKU can actually earn a test. Here is the approach we use for large-catalog and fashion brands.
Segment product sets by margin and velocity, not by category alone. A sale set, a hero set, a new-arrivals set, and a long-tail set behave differently and deserve different budgets. Lumping a 40-point-margin bestseller in with a clearance item teaches the model nothing useful about either.
Ring-fence a prospecting budget for new and long-tail SKUs. If everything competes in one set, the proven products win every time and the tail starves. A dedicated set with its own budget forces the automation to spend some impressions where it otherwise never would. It is the same logic as capping existing customers in ASC: you are protecting the growth line from the algorithm's short-term instincts.
Fix feed quality before you touch bids. Titles, primary images, product type, availability, and price are the signals Andromeda reads to decide fit. A SKU with a vague title and a placeholder image is invisible to a retrieval model no matter how much you bid. Clean the feed and you widen the pool of products the model is willing to consider.
Match content IDs across feed, pixel, and Conversions API. This is the boring one that breaks the most accounts. If the ID a shopper's AddToCart fires does not match the catalog ID, retargeting and prospecting both lose their signal for that product.
Refresh availability and pricing on a real cadence. Out-of-stock and stale-price SKUs poison delivery and waste budget on products people cannot buy. For fashion brands cycling seasonal inventory, feed freshness is not hygiene, it is performance.
For brands that run the same catalog on Google, the feed discipline carries straight over. We cover the Merchant Center side in our Google Shopping guide for DTC and CPG brands, and the SKU-level thinking is the same on both platforms.
How We Approach Large-SKU Catalog Optimization at jetfuel.agency
When we inherit a large catalog that is only moving its top 20 or 30 products, the first thing we do is stop blaming the algorithm and audit the structure feeding it. Most of the time the account is running one giant product set against one budget, the feed is half-optimized, and the content IDs are out of sync. The automation is doing the only sensible thing it can with the inputs it has.
Our process is to rebuild the product sets around margin and velocity, protect a prospecting budget for the long tail, and clean the feed so more SKUs are eligible to be retrieved in the first place. Then we judge it on catalog coverage and new-SKU sell-through, not just blended ROAS, because a great ROAS on your existing bestsellers can hide a catalog that never grows. So we track how much of the catalog is actually earning impressions before and after the rebuild, not just the headline return.
Back when I used to buy media, a catalog was a spreadsheet you fought with by hand. Now the machine fights you back. It wants to spend on the sure thing, every time. Our whole job on a big catalog is building the guardrails that make it test the products it would never touch on its own.
Frequently Asked Questions About Meta Catalog Ads for Large Catalogs
Why does Meta keep showing the same products from my catalog?
Because Andromeda retrieves the products it is most confident will convert, and your proven sellers have the history to earn that confidence while new SKUs do not. Left in one big product set with one budget, the bestsellers win every impression and the long tail never gets the data to compete. The fix is to give new and long-tail SKUs their own product set and protected budget so the automation is forced to test them.
Do I need to rebuild my catalog campaigns after the Andromeda update?
You rarely need to rebuild the campaign, but you almost always need to rebuild the product sets and feed underneath it. Andromeda changed which products get retrieved, so a catalog structured as one undifferentiated pile now concentrates harder than it used to. Segmenting product sets by margin and velocity and cleaning the feed is usually higher-leverage than any campaign-level change.
What is the difference between Advantage+ catalog ads and Advantage+ Shopping?
Advantage+ catalog ads are a product-feed format that decides which SKU to show which person. Advantage+ Shopping is a full-funnel campaign objective that decides the audience, placement, and budget for you. You can serve catalog products inside an ASC campaign, but the tools to fix which products get shown live on the catalog and product-set side, not the campaign objective.
How many SKUs should be in a Meta product set?
There is no single right number, but the mistake is one enormous set for everything. Group SKUs that share margin, velocity, and lifecycle so the model learns something coherent from each set. A dedicated set for new arrivals or the long tail, even a large one, will out-test a single catch-all set because it forces spend toward products that would otherwise be starved.
Does feed quality really change delivery, or is that overblown?
It genuinely changes which products are eligible to be retrieved. Andromeda reads titles, images, product type, availability, and price to judge fit, so a SKU with a weak title or a placeholder image is effectively invisible to it. Cleaning the feed widens the pool of products the model will consider, which is the whole point when you are trying to move beyond your top sellers.
We rebuild large catalogs so the long tail gets a fair test, not just your proven SKUs. Let us take a look at your product sets and feed.
Still have questions?
Let your AI
pressure-test us.
Ask the assistant you already trust.
What does Jetfuel Agency's analysis in "Meta Advantage+ Catalog Ads for Large-SKU DTC and Fashion Brands: SKU..." show about its expertise?
Launch into Success
Tell us a bit about yourself and your business. We are just one message away from the perfect partnership!