When Should You Kill a Product in Performance Max?

Work out your max cost per sale for each product before you launch. Take the price. Take off what you pay for the product, shipping, payment fees and refunds. What is left is the most one sale can cost you in ads before you lose money.

Say a product sells for $50 and leaves you $20. Its kill line is $40 to $60 of spend with no sale. Hit that and you exclude it. Write the number down on day 1, not on day 20 when you are already in the hole and hoping.

There is a second trigger. A product that does sell, but stays under its break-even ROAS once learning is over and the feed is clean, is a cut too. Fix the price first if you can. If you cannot, it goes.

The most useful reply in the Reddit thread says: spent at least 3x your average order value without a sale, kill it. It is simple, and simple rules get followed. That counts for a lot.

The problem is the margin. On the $50 product above, 3x the order value is $150. The margin rule stops you at $40 to $60. That is $90 to $110 more on a product that has not sold once. On a thin dropshipping margin, order value is the wrong base. What one sale leaves you is the right one.

Another reply says never kill a product in PMax, just fix titles and the landing page. Half right. Fixing the listing comes first. But a product that got a fair test and still sells nothing keeps eating budget your winners could use.

PMax does not split money evenly. In one campaign, a few products take most of the spend and the rest get a few dollars each. So the campaign total tells you nothing about one product. Open the products report and read cost, clicks and sales per item.

A product that spent $4 in six weeks did not fail. It never got a test. Do not kill it. If you want a real answer on it, move it to its own campaign with its own budget, or leave it be. The kill line only applies to products that actually spent up to it.

Check the data before you trust a zero. Open Google's Tag Assistant, buy one product and check the purchase fires with the right value. Then remember that sales arrive late. Google files a sale under the day of the ad click, so the last few days always fill in later. A product with no sales and broken tracking did not fail. Your setup did.

Timing matters too. In our dropshipping guide, we plan the first 14 days as the learning stage at about $60 a day, and start reading products from day 15. That is our plan, not a Google promise. Google says learning takes longer with few sales or a long gap between click and buy. Judging a product while the campaign is still learning is judging Google's warm-up, not the product.

The thread asks what CTR, CPC and conversion rate should trigger a kill. We do not give one benchmark for every niche, because a cheap impulse buy and a $300 product live in different worlds. Compare each product with your own other products instead. Then read the symptom.

No impressions, or almost none: that is not the market saying no. Check Merchant Center for a disapproval or a feed problem first.

Impressions but few clicks: work on the listing. A sharper title with the words people search, a better main image, and your price next to the other listings in the Shopping row.

Clicks but no sales: look at the price against rivals, then the product page on a phone, then checkout. If all of that is fine and it still hits the kill line, it is the product. Cut it.

In Google Ads, open Campaigns, pick your Performance Max campaign, then Asset groups, then Listing groups. Click the green circle next to the product or group and pick Exclude. That is Google's own path.

Listing groups belong to one asset group, not the whole campaign. If your campaign has more than one asset group, another one can still show the product. Exclude it in every asset group that includes it. Then check the campaign's Final URL expansion setting. When it is on, Google can send ads to other pages on your site, not only the product pages from your feed. So a cut product's page can still pick up traffic.

With many products, a custom label is cleaner. Give the dead products a label like "excluded" in your feed, then exclude that label once in each asset group's listing groups. Tag each new loser and it drops out, with no listing group edit. Google allows five custom labels per product.

Keep the budget the same and start the next test. Do not rebuild the campaign or slash the budget the day you cut a product. Big changes can send the bid strategy back into learning.

Want a second pair of eyes on your kill list? Post your products report in our free Google Ads eCom Lab on Skool. It has 1,600+ members and a weekly Google Ads question thread. Once a store spends at least €10K a month on ads, our agency team can run the account for you.