Quick answer: how should an online store run AI ads?
Start with a clean product feed, because Shopping, Performance Max and catalogue ads build ads from it. Send every purchase with its value from browser and server. Run one broad Meta sales campaign, catalogue retargeting and Google Shopping or Performance Max. Then improve the offer and product pages, and review weekly on profit.
Why the product feed comes first
For a service business, the ad is a video or an image you design. For an online store, much of the advertising is built automatically from your product feed: Google Shopping listings, Performance Max product ads and Meta catalogue ads all pull titles, images and prices from it. For the wider operating plan, see our complete guide to AI ads.
What Google requires
Google's product data specification lists the attributes every product needs, including id, title, description, link, image link, availability and price, plus brand for most new products. Titles can be up to 150 characters and descriptions up to 5,000.
What Meta does with it
Meta's catalogue tools use the same kind of feed for Advantage+ catalogue ads, which show products in different formats as personalised ads to each viewer.
Feed rules that help the AI
- Put the words shoppers search for at the start of the title: product type, key feature, size or colour, then brand.
- Use clear product images on a plain background for the main image, and lifestyle images as extra images.
- Keep price and availability in sync with the store, so ads never send people to out-of-stock items.
- Group products with custom labels such as margin band or best seller, so you can bid or report by group.
A simple campaign structure for online stores
Most stores grow faster with fewer, clearer campaigns. This structure suits a store spending roughly ₹1 lakh to ₹10 lakh a month.
| Campaign | Platform | Job |
|---|---|---|
| Broad sales campaign with 4 to 6 creative angles | Meta | Find new buyers |
| Advantage+ catalogue ads for viewers and cart abandoners | Meta | Bring back interested shoppers |
| Search on brand and top product terms | Catch people already searching | |
| Shopping or Performance Max with the product feed | Show products across Google inventory |
Where Performance Max fits
Google describes Performance Max as a goal-based campaign that reaches all Google Ads inventory from one campaign. For stores, it combines Shopping placements with YouTube, Display and more. Our guide to Performance Max and asset generation explains the setup.
Split products by margin
If some products earn 60% margin and others 15%, put them in separate product groups or campaigns with different targets. Otherwise the AI will happily sell lots of low-margin items.
Keep retargeting small
Catalogue retargeting often shows the best ROAS in the account, because it reaches people who were already close to buying. Keep it to a modest share of budget and judge growth by the broad campaign, which is where new customers come from.
Teach the AI which orders make money
The AI optimises for the purchases and values you send. Two habits make that signal much better.
Send purchases from browser and server
Meta recommends a redundant setup with deduplication: the pixel and the Conversions API both send each purchase, with the same event ID so it counts once. Shopify and most store platforms offer this through their Meta integration.
Send the right value
Pass the actual order value after discounts. If cash-on-delivery orders often fail, consider optimising for prepaid purchases or sending completed orders later as offline conversions, so the AI learns from orders that actually paid.
Check it monthly
Compare platform purchases with completed orders in your store. Our guide on what AI needs from your pixel, CAPI and store covers the checks step by step.
The offer and product page do half the work
The AI finds people likely to buy, but whether they buy depends on what they find when they land.
Offer
Free shipping thresholds, bundles and first-order offers raise conversion rate and average order value. A clear reason to buy today helps the AI find buyers faster.
Product page
Fast loading, clear photos, size or usage guidance, real reviews and visible delivery and return terms. Every point of conversion rate you gain lowers cost per purchase on every campaign at once.
A quick page check
Open your top product page on a mid-range phone on mobile data. If it takes more than a few seconds to show the price and the add-to-cart button, fix that before raising budget. For store tools that help, see our list of AI marketing tools for Shopify stores.
If the ad shows a bundle offer, land people on the bundle, not the home page. Mismatched landing pages waste the AI's targeting work.
A worked example: pricing a bundle for ads
Worked example: a tea brand
The single product
One pack sells for ₹499. Product cost ₹120, shipping ₹70, payment fees and GST ₹90, returns ₹20. Margin before ad cost is ₹199, about 40%. Cost per purchase from ads is ₹260, so each order loses ₹61.
The bundle
A three-pack sells for ₹1,299. Product cost ₹360, shipping ₹90, fees and GST ₹235, returns ₹40. Margin before ad cost is ₹574. With the bundle as the main offer, cost per purchase rises to ₹340 because fewer people buy.
Profit per order = margin before ad cost − cost per purchase
Single pack: ₹199 − ₹260 = −₹61. Bundle: ₹574 − ₹340 = ₹234.
The same ads become profitable once the offer carries a higher order value. That is a change only the brand can make; the AI then finds the buyers. Test the bundle as the hero offer in new creatives first, and keep the single pack available for people who want to try before buying more.
A weekly routine for store owners
- Monday: read the numbers. Spend, purchases, cost per purchase, cost per new customer and profit after ad spend for last week.
- Tuesday: check the feed. Disapproved products, out-of-stock items and price mismatches in Merchant Center and Commerce Manager.
- Wednesday: add creative. Two new angles into the existing Meta sales campaign.
- Thursday: one budget decision. Move 15 to 20% toward the campaign with the best profit, or hold.
- Friday: one page fix. Improve the product page with the most ad traffic.
At higher spend, the same rhythm still applies; our guide for brands spending ₹10 lakh+ a month adds team rules on top.
What most e-commerce ad advice tends to get wrong
It starts with targeting
For stores, the feed, data and offer matter more than audience settings. The AI handles most targeting now.
It treats all products the same
Products with very different margins need different targets or groups, or the AI will optimise for revenue instead of profit.
It ignores failed orders
In India, cash-on-delivery failures and returns can turn a good platform ROAS into a loss. Judge campaigns on completed orders.
ScaleWin, our product, is an AI ad manager for Meta and Google Ads. The playbook above works with or without any third-party tool.
Read next in this series
- AI Ads in India: Costs, Tools and Case Studies for 2026
- Can AI Run Your Ads Fully? What It Handles Alone, Where It Fails and What to Keep Human
- AI Ads and Privacy: What Changed After iOS Tracking Rules and the Cookie Debate
- AI Ads Targeting: Broad vs Interest Targeting in 2026
- AI Ads Budget: How Much Should You Spend, and How Fast Should You Scale?
Key takeaways
- Your product feed is the raw material for Google Shopping, Performance Max and Meta catalogue ads, so fix it first.
- Google requires id, title, description, link, image link, availability and price, with titles up to 150 characters.
- Purchase data with real order values teaches the AI which products and buyers make money.
- Offers, prices and product pages decide conversion rate, which the AI cannot fix for you.
- A fixed weekly routine beats daily changes on every campaign.
Ai ads: frequently asked questions
What is the most important part of AI ads for an online store?
What product data does Google require?
Should an online store use Performance Max or Shopping campaigns?
How do I stop the AI from pushing low-margin products?
Do catalogue ads work for small stores?
What should a store owner check every week?
Sources
- Google Merchant Center Help: Product data specification
- Meta for Developers: Catalog
- Meta Business Help: Conversions API best practices
- Google Ads Help: About Performance Max campaigns
Let ScaleWin read your ad data every day
ScaleWin checks your Meta and Google Ads daily, judges every ad on profit and tells you exactly what to change.
Start free, no card



