Google Shopping Management Agency for Ecom Brands

Target ROAS is the most misused setting in Google Shopping. Google's target ROAS documentation says what it does plainly: you give the bidder a return goal and it bids to hit that goal on average. The trap is what happens under the average.

Set the target too high and the bidder does the easy thing. It pours budget into the handful of products and searches it already knows convert, and quietly stops bidding on the rest. Your reported ROAS looks great. Your coverage of the long tail collapses. Revenue flattens while the dashboard says the account is healthy.

So we do not run one target across the whole catalogue. Each margin tier gets its own, worked out from the real margin on those products, not from a number someone heard was good. Break-even ROAS is 1 divided by your gross margin. A 60% margin product breaks even around 1.7. A 12% margin product needs about 8.3 before it makes you a cent. Point the same target at both and you either burn the thin one or leave the fat one under-bid.

Then we move slowly. Target changes go in small steps, and never in the middle of a learning period. Google's own Smart Bidding help says a significant change restarts that clock. Restart it every week and the account stalls for a month with nobody noticing. Between changes we watch impression share and how many distinct products still got clicks. If that product count is shrinking, the target is too tight, whatever the ROAS number says.

Shopping campaigns have no keywords. You cannot tell Google which searches to enter. You can only tell it which ones to stay out of, which makes negatives the main steering wheel you get.

The input is the search terms report, and we read it every week on every account. Three things come out of it. Junk goes negative: the how-to searches, the repair searches, the competitor part numbers, the free-download hunters. Brand gets separated, so branded traffic stops inflating the ROAS of campaigns that did not earn it. And the terms that actually convert get their own campaign or product group. There they carry their own bid, instead of being averaged in with everything else.

That last move is what people mean by query sculpting. On Standard Shopping you do it with negatives plus campaign priority, and priority works backwards from what most people guess. Google checks the High priority campaign first, whatever the bids are. So High priority becomes the cheap catch-all, you negative your money terms out of it, and those searches fall through to a Low priority campaign carrying your real bid. The full build is in our post on the three-campaign priority structure. Performance Max gives you far less of that control, which is exactly why we keep Standard Shopping in the mix instead of retiring it.

Every negative goes into one shared list per account, so the same junk term never has to be blocked twice in five campaigns. We phrase-lock each one, log who added it and when, and can take it back out. If a negative turns out to be blocking real demand, that is a five-minute undo, not an archaeology project.

Every account gets a written update in a shared Slack channel each week. Not a slide deck, not a PDF nobody opens. A short read in plain words: what the numbers did, what we changed, what we are doing next, and what we need from you.

The numbers come from Shopify first and Google Ads second. Google Ads reports what it thinks it drove. Shopify reports what actually got paid for. When the two disagree, the update says so instead of quietly using the flattering one.

On top of the weekly you get a live dashboard. Spend, revenue, ROAS, product-level performance, Merchant Center health, and the current disapproval list, refreshed daily. If you want Tuesday's numbers on Tuesday, you do not have to ask anyone for them.

Once a month it goes deeper. Which products moved between Champion, Sleeper, and Waster. What the feed work changed. Where the wasted spend went and what we did about it. What the plan is for next month. Every claim in it traces back to a number you can open yourself.