Query sculpting across Shopping campaigns with priorities

Standard Shopping campaigns have no keywords. Google matches a query to product data, and the advertiser’s only levers over which query triggers which product are campaign priority and negative keywords. Used together they let the same product carry a low bid on a generic query, a higher bid on a category query, and the highest bid when the query names the product or the brand. This is the mechanism, the build, and where it stops working.

The mechanism in four rules

1. When one product is eligible in several Shopping campaigns in the same account, the campaign with the higher priority serves it, whatever the bids are.
2. When priorities are equal, the higher bid serves.
3. A campaign stops being eligible for a query when a negative keyword in that campaign matches the query, or when the campaign’s daily budget is exhausted.
4. A query that a higher-priority campaign cannot serve falls through to the next priority down.

Rules three and four are the whole trick. A high-priority campaign with negatives for brand and product terms will catch every query except those, and pass those down to a campaign that bids more for them. The negative keywords documentation covers match behaviour, and About Shopping ads is the platform reference for how products are matched.

The three-tier structure

Tier	Priority	Queries it should catch	Bid level	Negative lists applied
Generic	High	Broad category words with no brand or product name	Lowest	Brand terms, product and model terms, waste terms
Category	Medium	Product type plus attributes, no brand or model	Middle	Brand terms, waste terms
Brand and product	Low	Your brand, product names, model numbers	Highest	Waste terms only

Every product sits in all three campaigns. A query for “trail running shoes” enters the generic campaign at the lowest bid. A query for “waterproof trail running shoes wide fit” also enters generic unless you push it down; adding attribute terms to the generic campaign’s negatives routes attribute-rich queries to the category tier. A query for “Northline Ridge 4 trail shoe”, with a made-up brand and product, matches the brand and product negatives in the top two tiers and lands in the low-priority campaign at the highest bid. Same product, three bids, decided by what the shopper typed.

Budgets are part of the routing

Rule three has a consequence people miss. If the generic campaign runs out of budget in the afternoon, every generic query for the rest of the day falls through to the category tier at a higher bid, and then to the brand tier at the highest bid. The generic campaign’s budget must be large enough that it never caps, or the sculpting reverses itself daily. Shared budgets across the three tiers defeat the purpose for the same reason.

Building the negative lists

Three shared lists, applied as the table shows.

• Brand terms. Every spelling and misspelling of your brand, plus the brand names of products you stock if brand queries should reach the high-bid tier.
• Product and model terms. Product line names, model numbers, SKU-like strings shoppers type. Built from the catalogue and from the search terms report.
• Waste terms. Words that signal a query you never want: free, used, repair, manual, jobs, recall, and whatever your category adds.

Use phrase match for multi-word terms and exact match where a word is ambiguous, since a broad negative in the generic tier can block more than intended. Build the first version from the catalogue, then refine weekly from the search terms report for the first month and monthly afterwards. Every query appearing in a tier it does not belong to is a missing negative.

Bidding by tier

The structure was designed for manual and enhanced bidding, where the bid on each tier is set directly and the sculpting is visible. It also works with Smart Bidding: target ROAS is set per campaign, and priorities still decide which campaign is eligible for a query before bidding runs. Set each tier’s target from its role. The brand tier converts best and carries the lowest risk, so it can run a modest target and take the volume. The generic tier is acquisition, where the store decides how much return to give up for new customers, and the break-even ROAS calculator gives the floor below which no tier should go. The target ROAS bidding entry explains how the target translates into bids.

Where it breaks

• Performance Max in the same account. Performance Max takes precedence over Standard Shopping for the same product, so a Performance Max campaign holding the whole catalogue makes the tiers irrelevant. Split products between the two by custom label, and keep the sculpted set out of Performance Max. The setup for the other side is in feed-only Performance Max.
• Weak titles. A product whose title lacks the category and attribute words cannot match category queries in any tier. Sculpting routes queries; the feed decides whether the product is in the auction at all. Feed title optimisation comes first.
• Too little volume. Three campaigns on a small account leave each with too few conversions for Smart Bidding. Two tiers, brand and everything else, are enough for most small stores, and the split still removes the worst inefficiency, which is paying generic prices for brand queries.
• Negative lists that drift. A new product line launched without its name in the product list serves from the generic tier at the lowest bid until someone notices.

Reading the results

Open the search terms report for each campaign. The generic tier should contain nothing with a brand or model in it; the brand tier should contain nothing without one. Anything else is leakage. Then compare impression share and click share by tier: a brand tier below full impression share is leaving the cheapest, most reliable conversions on the table, and why is Shopping impression share low covers the causes. Rising costs in the generic tier are usually competition rather than structure, and Shopping CPCs rising and how do you lower CPC in Google Shopping go through the responses.

Setup checklist

1. Confirm Performance Max holds none of the products you intend to sculpt.
2. Create three Standard Shopping campaigns with identical product groups and priorities set high, medium and low.
3. Build the three negative keyword lists and apply them per the table.
4. Set bids or targets per tier, with the generic tier lowest.
5. Give the generic tier a budget it will not exhaust.
6. Review search terms weekly for a month, then monthly, moving leaked terms into the right list.
7. Report performance by tier, and compare the brand tier’s cost per conversion with what those queries cost before the split.

For catalogues large enough that product segmentation and query segmentation both matter, how to structure Shopping campaigns for a large catalogue compares the options, and should an ecommerce brand bid on its own name settles the question the brand tier depends on. The Google Shopping service runs this structure alongside the feed work it relies on.

---

Sources

• Google Ads Help: About Shopping ads
• Google Ads Help: About negative keywords
• Google Ads Help: About the search terms report