Facebook Dynamic Product Ads (DPA): Setup, specs & strategy

Facebook dynamic product ads show each person the products they're most likely to buy, pulled automatically from your catalog. Someone views a product on your store, and later sees that exact product - or similar ones - in their feed, with current price and availability. No individual ads created, no manual updates.

The mechanics have been around for years, but the naming and the campaign structure have changed. What advertisers still call DPA, Meta calls Advantage+ catalog ads, and the campaigns that deliver them have been consolidated into the Advantage+ campaign experience.

This guide covers how the system works today, what you need to launch, and how feed quality, event matching, campaign structure, and creative affect delivery.

What are dynamic product ads?

A dynamic product ad is a template ad connected to a product catalog. Instead of uploading images and writing copy per ad, you define the format - single image, carousel, or collection - and Meta fills it with products, prices, and links from your catalog in real time.

Delivery works in two modes:

• Retargeting: show people products they viewed, added to cart, or purchased, based on pixel events from your site

• Prospecting (broad audience): show relevant catalog products to people who haven't visited, based on Meta's interest and behavior signals

Because the ad content is generated per viewer, one campaign can effectively be thousands of personalized ads. If catalog ads as a concept are new to you, start with what are catalog ads.

What you need before launch

Three components, and the quality of each decides your results.

1. A product catalog

Your catalog lives in Commerce Manager and is fed by a data source - ideally a hosted feed URL on a daily schedule, not manual uploads. The required fields are id, title, description, availability, condition, price, link, image_link, and brand.

Feed quality is the ceiling on campaign quality. Rejected products can't be advertised, stale prices break trust at the landing page, and weak titles truncate mid-word. Set up the pipeline properly - product feed management covers the full workflow - and check common feed errors before you scale spend.

2. The Meta Pixel (with matching content IDs)

Retargeting depends on the pixel reporting which products people interact with. You need three events on your store:

• ViewContent on product pages

• AddToCart

• Purchase

Each event must pass a content_id that exactly matches the id in your catalog - the match is case-sensitive. This is a critical technical check in DPA setups: events can fire while mismatched IDs reduce product matching and retargeting delivery.

Shopify and most platforms handle this automatically, but verify it: open Events Manager, check that ViewContent events carry IDs, and spot-check them against your feed.

3. Product sets

Product sets are filtered slices of your catalog - by category, price band, margin, or label - and they're what campaigns actually promote. Even if you start with "all products", build sets early:

• Bestsellers for prospecting

• High-margin products for aggressive bidding

• Clearance excluded from prospecting entirely

Meta now recommends filtering sets by internal_label rather than custom labels, because editing custom labels re-triggers policy review on the products.

Setting up the campaign

In the current Advantage+ campaign experience, catalog ads run under the sales objective (they're also available for traffic, engagement, leads, and app promotion):

1. Create a sales campaign and attach your catalog, selecting the product set to promote

2. Choose audience mode: let Advantage+ audience run broad for prospecting, or define retargeting windows (e.g., viewed or added to cart in the last 14 days, excluding purchasers)

3. Pick formats: carousel is the workhorse; single image and collection have their places

4. Set the creative options: which fields render (price, overlays), and whether to use your enriched images

5. Launch, and give the system enough conversions to exit learning before judging results

Meta's own numbers explain why the consolidated setup leans broad and automated: in a 31-advertiser test, Advantage+ shopping campaigns improved cost per acquisition by 17% and ROAS by 32% versus business-as-usual campaigns, and Meta cites a 9% average improvement in cost per conversion for the current Advantage+ sales setup.

Treat those as directional, Meta-reported figures from the tested advertisers, not a guarantee. Use them as a reason to test consolidated delivery with a strong catalog against your existing setup.

Specs that matter

The specs that actually change outcomes, from Meta's image specifications:

• Images: JPEG or PNG, minimum 500 x 500 pixels, 1024 x 1024 recommended, under 8 MB

• Carousel consistency: keep every card in the same ratio - 1:1 or 4:5 - or Meta may auto-adjust all of them

• 9:16 placements (Reels, Stories): interface elements cover image edges; keep price and text in the safe area

• Up to 20 additional images per product, and up to 20 product videos (each under 200 MB, 4:5 recommended) via the feed's video fields

• Titles: under 65 characters render best; longer titles truncate

One operational quirk worth repeating: if you change a product image, serve it from a new URL - Meta doesn't re-fetch changed images at the same URL.

Creative: the actual competitive lever

Every store's DPA pulls from the same machinery. The targeting is Meta's; the format is Meta's; what's yours is the data and the pixels people actually see.

Default catalog ads render your website product photos. They're accurate, and they're identical to every competitor's. The fix isn't manual design - it's template-based enrichment: rendered creatives that add price, discount badges, brand framing, and campaign backgrounds, generated from the feed so they stay in sync.

That's the difference between a catalog ad that looks like a database row and one that looks like a campaign. Under The Sun applied campaign-ready catalog creative across 1,500+ products with one reusable template system and no per-product editing.

Best practices

1. Feed first. Daily syncs minimum, clean titles, valid identifiers, no rejected products. Everything else builds on this.

2. Verify pixel-catalog matching before spending. Events Manager → check content_ids against feed ids, case-sensitively.

3. Separate retargeting and prospecting campaigns. Different intent, different messaging, different expectations for ROAS.

4. Exclude recent purchasers from retargeting windows.

5. Use sale price fields for promotions. The strikethrough-plus-sale-price render is a clear, scalable treatment to test during promotions.

6. Segment by margin with internal labels. Bid on profit, not just revenue.

7. Refresh creative by rotating templates when frequency climbs and CTR decays - not by pausing the campaign.

8. Give learning phase room. Judge after conversions accumulate, not after 48 hours.

Troubleshooting quick hits

• Ads not delivering: check for rejected products in the set, a paused feed schedule, or an unbound pixel - the full walkthrough is in product feed errors

• Retargeting volume near zero: first check for a content_id mismatch, then audience size, pixel health, product approval, and stock status

• Wrong prices showing: feed lag; sync more often, and keep sale prices in sale_price

• Old images showing: image changed at the same URL - version the URL

Connecting the catalog and creative layers

In the three-part DPA system above, Flow supports two parts: the catalog and the creative it supplies. Product data can be imported, mapped, validated, and kept synchronized, while reusable templates turn that same data into reviewed format variations.

The signal layer still belongs to Meta and your site. Pixel implementation, content_id matching, audience eligibility, campaign setup, and Commerce Manager diagnostics must be verified there; Flow can improve the inputs, but it cannot repair a broken event join from outside the platform.

FAQ

Final thoughts

Dynamic product ads are the most automated ad format Meta offers, which is exactly why the inputs matter so much: the feed decides what can show, the pixel decides who sees it, and the creative decides whether anyone cares.

Get the data pipeline right, verify the matching, and put your effort where competitors don't - templates that make a thousand automated ads look designed.