How catalog ads work on Meta and Google today
Before diving into ChatGPT’s product spotlight format, it helps to understand the catalog-driven advertising model that already generates billions in revenue across Meta and Google. If you run e-commerce ads today, you are almost certainly using some version of this.
Meta Dynamic Product Ads (DPA)
Meta’s Dynamic Product Ads pull directly from your product catalog to serve personalized ads to users based on their browsing behavior and purchase intent. The standard approach splits budget roughly 60/40 between retargeting (showing products to people who already visited your site) and Advantage+ prospecting (finding new customers who resemble your buyers).
Advantage+ Shopping Campaigns, Meta’s fully automated campaign type, consistently deliver 15–25% better ROAS than manually targeted campaigns for catalog-based advertisers. The system tests thousands of product-audience combinations automatically and shifts budget toward what converts. For brands with 50+ SKUs, Advantage+ has become the default campaign structure.
The product feed is the foundation. Every product in your catalog needs a title, description, image, price, availability status, and product category. Meta uses this data to build the ad creative dynamically – matching the right product to the right user at the right time. Feed quality directly determines ad performance.
Google Shopping and Performance Max
Google Shopping works similarly. Your product feed in Google Merchant Center powers both standard Shopping campaigns and Performance Max (PMax). PMax extends your catalog across Search, Shopping, Display, YouTube, Discover, and Gmail – all from a single feed.
Custom labels are a critical optimization lever on both Meta and Google. Smart catalog advertisers use custom labels to segment products by margin tier (high, medium, low), seasonality (summer, winter, evergreen), bestseller status, or promotional pricing. This segmentation lets you allocate budget to the products that actually drive profit, not just revenue.
The key insight for ChatGPT ads is this: the feed infrastructure you already maintain for Meta and Google is the starting point for product spotlight campaigns on ChatGPT. You do not need to build a new catalog from scratch. You need to adapt what you have.
How ChatGPT product spotlight units work
Product spotlight units are ChatGPT’s shopping-focused ad format. They display as carousels within the conversation, showing product images, pricing, brief descriptions, and direct checkout links. The format is designed for purchase-intent conversations – when a user asks ChatGPT something like “what’s the best standing desk under $500?” or “I need running shoes for flat feet,” a product spotlight can surface relevant products alongside the AI’s response.
800M+
Weekly active ChatGPT users who may see product spotlight ads
The Shopify integration announced in March 2026 is the first native catalog connection. Shopify merchants can sync their product catalog directly, which enables real-time pricing and availability updates in the spotlight carousel. When a product sells out or a price changes in your Shopify store, the ChatGPT ad reflects the update automatically.
Product spotlights are triggered by conversational context, not keyword bids. ChatGPT’s ad system analyzes the full conversation thread to determine purchase intent and category relevance. If a user is asking detailed comparison questions about a product category you sell in, your products can appear in the spotlight carousel. The system matches products to conversation topics, similar to how Meta DPA matches products to user behavior – but using conversational intent instead of pixel data.
Only one ad unit appears per response, so product spotlights compete for a single placement. This means feed quality, image clarity, and description relevance directly impact whether your products get shown over a competitor’s.
Product feed quality requirements
Feed quality is the single biggest determinant of catalog ad performance on any platform. A weak feed produces weak ads, regardless of your budget or targeting. Here is how the requirements compare across Meta, Google, and ChatGPT.
| Element | Meta DPA | Google Shopping | ChatGPT Spotlight |
|---|---|---|---|
| Image minimum | 600×600px | 100×100px (250×250 non-apparel) | 256×256px |
| Recommended image | 1080×1080px | 1080×1080px | 512×512px+ |
| Title length | 150 chars max | 150 chars max | Clear, concise (no keyword stuffing) |
| Description style | Feature-focused | SEO-optimized, keyword-rich | Conversational, benefit-focused |
| Pricing | Required, real-time via feed | Required, real-time via Merchant Center | Required, real-time via Shopify sync |
| Availability | Required | Required | Required (auto-synced from Shopify) |
| Image background | Clean, white preferred | White required for Shopping | Clean, product clearly visible at thumbnail |
The biggest difference for ChatGPT is the description style. Meta and Google reward keyword-dense, feature-packed descriptions because their systems parse structured attributes. ChatGPT’s system operates in a conversational context, so descriptions that read like natural recommendations perform better than strings of specifications.
Example of a keyword-stuffed description (bad for ChatGPT): “Adjustable Standing Desk Electric Height Motorized Sit Stand Desk 60x30 Bamboo Desktop Home Office Ergonomic Workstation.”
Example of a conversational description (good for ChatGPT): “Standing desk that adjusts from 28 to 48 inches with one button. 60×30 bamboo top fits two monitors. Rated 4.8 stars by 2,400 buyers.”
Both descriptions contain the same information, but the conversational version reads as a recommendation rather than a search result. In a ChatGPT conversation where a user asked “what standing desk should I buy for my home office?” the second version fits the tone of the interaction.
Image requirements for product spotlight
Product spotlight images display as thumbnails in the carousel, which means clarity at small sizes is critical. Upload at 512×512px or higher for sharpness. Use clean backgrounds with the product centered and clearly visible. Avoid lifestyle photography with busy backgrounds, text overlays, or multiple products in a single image – these become unreadable at carousel thumbnail scale.
If your existing product images are 1080×1080px for Meta, they will work for ChatGPT after cropping or resizing. The key check is whether the product remains identifiable at 256×256px. Open your image, scale it down, and verify the product is still clear. If it is not, reshoot with a tighter crop.
Step-by-step: Shopify to ChatGPT ads
Shopify merchants have the most direct path to ChatGPT product ads thanks to the native integration. Here is the process from start to launch.
1. Connect your Shopify catalog
Navigate to your Shopify admin and enable the OpenAI sales channel. This syncs your full product catalog – titles, descriptions, images, pricing, and availability – to ChatGPT’s ad system. The sync is real-time, so inventory changes and price updates reflect automatically in your product spotlight ads.
Before connecting, audit your catalog for completeness. Every product needs a title, at least one high-resolution image (512×512px+), a description, a price, and an availability status. Products missing any of these fields will not be eligible for spotlight placement.
2. Configure product sets by margin and category
Not every product in your catalog should get the same ad budget. At $60 CPM and roughly $12 effective CPC, the unit economics favor products with higher average order values and stronger margins.
Create product sets segmented by:
- Margin tier – prioritize high-margin products that can absorb the cost per click
- Category – group products by category so you can match them to relevant conversation topics
- Price point – products with AOV above $100–$200 produce better unit economics at current CPMs
- Bestseller status – proven sellers with strong reviews convert at higher rates
This mirrors the custom label strategy that top-performing Meta DPA and Google Shopping advertisers already use. If you have custom labels set up in your existing feeds, use the same segmentation logic for ChatGPT.
3. Write conversational product descriptions
This is the step most Shopify merchants will need to spend the most time on. Your existing product descriptions are almost certainly written for SEO or for on-site shoppers scanning bullet points. ChatGPT product spotlights need descriptions that read like a friend recommending the product.
Rewrite each product description with these principles:
- Lead with what the product does for the buyer, not what the product is
- Use natural language – write as if explaining the product in conversation
- Include one to two specific differentiators (material, dimensions, ratings, price)
- Skip keyword repetition, all-caps labels, and promotional language
For catalogs with hundreds of SKUs, rewriting every description manually is impractical. Lapis can generate conversational product descriptions in bulk from your existing feed data, reformatting keyword-heavy listings into ChatGPT-ready copy.
4. Set topic cluster targeting
ChatGPT ads use topic clusters for targeting, not keywords. Define the conversation topics where your products should appear. For a home office furniture brand, topic clusters might include “standing desks,” “home office setup,” “ergonomic furniture,” and “work from home equipment.”
Map your product sets to topic clusters. High-margin standing desks go into the standing desk and ergonomic furniture clusters. Accessories with lower AOV might target narrower clusters where competition is lighter and CPMs are lower.
5. Launch product spotlight campaigns
Set your daily budget (the self-serve platform is expected to require $500/day minimum), select your product sets and topic clusters, and launch. Start with your highest-margin, highest-AOV product set to generate conversion data quickly. Expand to additional product sets once you have enough data to evaluate performance.
Monitor early performance closely. Key metrics to track: impressions per product, click-through rate by product set, and cost per acquisition by margin tier. If a product set is generating impressions but not clicks, revisit the descriptions and images for those products.
Step-by-step: Amazon and DTC to ChatGPT ads
Amazon sellers and DTC brands on platforms other than Shopify do not have a native integration yet. The process requires more manual work, but the opportunity is the same.
1. Export your product data
Start by exporting your product catalog from Amazon Seller Central, your DTC platform (WooCommerce, BigCommerce, Magento), or your existing product feed. You need: product titles, descriptions, images, pricing, categories, and any custom attributes like margin tier or bestseller status.
From Amazon, download your inventory report and active listings report. These contain the product data you need, though Amazon-formatted descriptions will require significant rewriting for ChatGPT’s conversational format.
2. Reformat descriptions for conversational tone
Amazon product descriptions are heavily optimized for Amazon’s A9 search algorithm – dense with keywords, bullet-pointed, and structured for scanning. This format performs poorly in ChatGPT’s conversational context.
Rewrite each description to read naturally. Remove keyword repetition, convert bullet points to flowing sentences, and lead with the benefit to the buyer. The goal is a description that would sound natural if ChatGPT recommended the product in conversation.
| Amazon format | ChatGPT format |
|---|---|
| “Wireless Bluetooth Earbuds, Active Noise Cancelling, 36H Battery, IPX5 Waterproof, Touch Control, USB-C Charging Case, Premium Sound Quality for iPhone Android” | “Wireless earbuds with active noise cancelling and 36 hours of battery life. IPX5 water-resistant for workouts. Charges via USB-C and works with both iPhone and Android.” |
| “Cast Iron Skillet 12 Inch Pre-Seasoned Frying Pan Oven Safe Cookware Kitchen Indoor Outdoor Camping BBQ Grill Stovetop Compatible” | “12-inch cast iron skillet, pre-seasoned and ready to use. Works on stovetop, in the oven, on the grill, or over a campfire. Lasts a lifetime with basic care.” |
3. Create images at 512×512px or higher
Amazon product images are typically 1000×1000px or larger on white backgrounds, which is a good starting point. Resize or crop your hero image to 512×512px for ChatGPT upload. Verify the product is clearly visible at 256×256px (the minimum display size).
If your Amazon images include text overlays, infographics, or lifestyle contexts with busy backgrounds, use your clean white-background hero image instead. The carousel thumbnail format demands simplicity.
4. Set up campaigns manually
Without a native integration, you will need to upload product data to OpenAI’s ad platform manually or through a tool like Lapis that can ingest your product feed, reformat it for ChatGPT specifications, and push it to the platform. Define your product sets, assign topic clusters, and set budgets the same way Shopify merchants do.
The manual process is viable for catalogs up to about 100 SKUs. Beyond that, the description rewriting and image reformatting work scales linearly and becomes a bottleneck. Automation tools become essential at scale.
What Meta DPA teaches us about ChatGPT product ads
Meta Dynamic Product Ads have been running at scale since 2015. A decade of optimization data from DPA campaigns offers concrete lessons for ChatGPT product spotlight strategy.
Let the system match products to context
The biggest lesson from Advantage+ Shopping is that algorithmic product-to-user matching outperforms manual curation at scale. Advertisers who tried to hand-pick which products to show to which audiences consistently underperformed those who gave the system a full catalog and let it optimize.
ChatGPT product spotlights work the same way. Upload your full eligible catalog (filtered by margin and AOV thresholds), define broad topic clusters, and let the system determine which products surface in which conversations. Resist the urge to micro-manage which product appears for which query.
Segment by AOV and margin
At $60 CPM and roughly $12 effective CPC, the math is straightforward. A product with a $30 AOV needs to convert at 40% just to break even on ad spend – that is unrealistic for cold traffic. A product with a $200 AOV only needs a 6% conversion rate to hit a 1:1 ROAS, which is achievable from high-intent ChatGPT conversations.
| Product AOV | Eff. CPC ($12) | Break-even CVR | Verdict |
|---|---|---|---|
| $30 | $12 | 40% | Not viable at current CPMs |
| $75 | $12 | 16% | Difficult, needs high intent |
| $150 | $12 | 8% | Viable with strong feed |
| $300 | $12 | 4% | Strong unit economics |
| $500+ | $12 | 2.4% | Ideal for ChatGPT spotlight |
This does not mean low-AOV products should never run on ChatGPT. It means you should segment your catalog and allocate the majority of ChatGPT budget to products where the math works. Use custom labels (the same ones you use for Meta DPA and Google Shopping) to separate your catalog into margin tiers and bid accordingly.
Feed quality equals ad quality
In Meta DPA, advertisers who invest in feed quality – better images, more accurate titles, complete product attributes – consistently outperform those who upload a bare-minimum feed and try to compensate with targeting. The same principle applies to ChatGPT, arguably even more so.
ChatGPT users are mid-conversation, actively evaluating options. A product spotlight with a blurry image, a keyword-stuffed title, and a generic description will lose to a competitor’s spotlight with a crisp image, a clear title, and a conversational description – even if the competitor’s product is objectively worse. The ad format does not give you space to overcome a weak first impression.
Use custom labels aggressively
Custom labels saved DPA advertisers from the trap of treating every product equally. The same applies to ChatGPT. Label your products by:
- Margin tier – allocate more budget to high-margin products
- Seasonality – increase bids on seasonal products during peak periods
- Performance history – products that convert well on Meta or Google are likely to convert well on ChatGPT
- Review rating – products with 4.5+ star ratings convert at higher rates across every platform
Turn your catalog into ChatGPT ads with Lapis
The hardest part of moving from Meta DPA or Google Shopping to ChatGPT product spotlight ads is the creative reformatting – rewriting hundreds of product descriptions from keyword-dense to conversational, resizing images, and configuring product sets for a new platform. Lapis automates this workflow.
Import your catalog. Connect your Shopify store, upload a product feed CSV, or paste product data directly. Lapis ingests your existing titles, descriptions, images, pricing, and product attributes.
Generate ChatGPT-format creatives per product. Lapis rewrites product descriptions from keyword-optimized to conversational, maintaining all the critical product details (specs, pricing, differentiators) while shifting the tone to read as natural recommendations. Each product gets multiple description variations for testing.
Auto-size images. Lapis automatically crops and resizes your product images to 512×512px for ChatGPT spotlight placement, verifying that the product remains clearly visible at thumbnail scale. If an image fails the clarity check at 256×256px, Lapis flags it for manual review.
Multi-platform output. The same catalog import generates ad creatives for Meta DPA, Google Shopping, ChatGPT product spotlight, LinkedIn, and TikTok. You do not need to maintain separate creative workflows for each platform. One import, multiple outputs.
Try Lapis for free and start converting your product catalog into ChatGPT-ready creatives today. Whether you are a Shopify merchant with native integration access or an Amazon seller building campaigns manually, Lapis handles the reformatting work that would otherwise take weeks of manual effort.
For more on ChatGPT advertising strategy, read our complete guide to ChatGPT ads. To learn how to create ad images that work at ChatGPT’s thumbnail scale, see our guide on ChatGPT ad images at scale. And for a step-by-step walkthrough of building your first ChatGPT ad creative, read how to create ads for ChatGPT.
