Shopping Campaign Priority Settings: Does the Old Trick Still Work?

The priority tiering trick still functions mechanically, but Smart Bidding has made it the wrong default for most accounts.

October 8, 2026 · QueryCut

Shopping campaign priority settings (high, medium, low) still work exactly as they always did: when two Standard Shopping campaigns both carry the same product, the higher-priority one enters the auction, regardless of bid. The classic trick used this to route cheap, broad traffic away from expensive queries. It still functions, but for the vast majority of accounts running Target ROAS or Performance Max, it is now the wrong move and often actively harmful.

What the priority setting actually controls

Priority is a tiebreaker, not a bid. When the same product (same item ID) is eligible in two or more of your own Standard Shopping campaigns, Google does not let both compete. It picks the campaign with the higher priority to enter the auction, and the other sits out for that query. Within a single priority tier, bids decide. Across tiers, priority decides first and bid is irrelevant.

This is the mechanism the old trick exploited. It has nothing to do with quality, relevance, or how good the match is. It is purely an internal routing rule to resolve self-competition, because Google will not auction your product against itself.

The critical consequence: priority only matters when the same product lives in multiple campaigns. If every product sits in exactly one Shopping campaign, the priority setting does literally nothing. People change it expecting a performance lift and see no change, because there was no overlap to arbitrate.

The old trick, in full

The classic setup, often called tiered or waterfall Shopping, used three Standard campaigns carrying the same products at descending priority, with negative keywords wiring traffic between them.

The high-priority campaign got a low bid and broad, generic negatives removed only the terms you did NOT want cheap. Actually the standard build is the reverse of what people remember, so here is the exact wiring: High priority = broad, generic, cheap traffic with low bids, and it holds negatives for your high-value branded and long-tail terms so they fall through. Medium priority = mid-intent terms. Low priority = your money terms (brand, specific SKU, buy-intent) with high bids, catching whatever the upper tiers rejected via negatives.

Worked example. Say you sell running shoes and 'nike pegasus 41 size 10' is your money query. In the HIGH-priority catch-all campaign you add the campaign-level negative 'nike pegasus 41' so that query cannot enter there. In the LOW-priority campaign (high bid, your closers) you leave it eligible. Now a search for the generic term 'running shoes' enters the high tier at a low bid, while 'nike pegasus 41 size 10' is blocked up top and falls to the low tier where you bid aggressively. Notice the cost already: one query, managed across two campaigns, and the negative must live in the right tier or the routing breaks.

That mirroring is the real tax. Every time you discover a new high-value term, you must add it as a negative to every tier above the one you want it to land in. Miss one, and the term enters the cheap campaign at the wrong bid. With three tiers and a growing negative list, this is a maintenance job that never ends.

Why Smart Bidding broke the use case

The trick solved a problem that manual bidding created: you could not set different bids for the same product based on query intent, so you faked it with campaign structure. Target ROAS and Maximize Conversion Value solve that problem directly. Smart Bidding sets a query-time bid for every auction using signals you cannot see or encode as negatives: device, time, user list membership, query specificity, and dozens more. It already bids less on 'running shoes' and more on 'nike pegasus 41 size 10' inside a single campaign.

So the tier structure is now redundant at best. At worst it starves the bidding algorithm. Splitting the same products across three campaigns fragments conversion data. Each campaign's bidding model learns from a thinner slice of signal, so all three optimize worse than one consolidated campaign would. You are paying a learning penalty to recreate, badly, something the algorithm does automatically and better.

There is also a reporting trap. Because priority routing happens before the auction, your tier campaigns will show wildly different CPCs and conversion rates that reflect the routing, not genuine performance differences. People read the low-priority 'money' campaign's high ROAS as proof the structure works. It is not: that campaign only ever sees pre-qualified high-intent traffic you fed it. The ROAS is an artifact of the plumbing.

When tiering still earns its keep

It is not dead everywhere. The structure still works mechanically and is defensible in specific cases, all of which share one trait: you genuinely need hard, deterministic control that Smart Bidding's probabilistic bidding cannot give you.

Clear-out and margin control: if you must cap spend on low-margin or clearance SKUs regardless of predicted value, priority plus manual CPC gives you a hard ceiling Smart Bidding will not. Brand defense versus prospecting split: some advertisers want brand queries handled at a fixed low bid (since they would convert anyway) and everything else bid up, and they do not trust tROAS to keep those separate. Query-type isolation for reporting: if you specifically need to see generic versus branded Shopping spend as separate line items for internal reporting, tiering forces that split.

Even in these cases, weigh it against Performance Max, which has absorbed most Shopping inventory and does not expose priority at all. If your decision is PMax versus Standard Shopping, priority tiering is only a reason to stay on Standard if the deterministic control is worth losing PMax's broader placements.

The failure mode nobody warns you about

The trap is the orphaned negative. You build three tiers, wire up your negatives, and it works for a month. Then you launch a new product line, or a seasonal query spikes, and a high-value term enters the cheap top-tier campaign because you never added the mirrored negative. It converts at a low manual bid, your top tier reports great, and your actual closer campaign sits idle on that term. You do not notice because the aggregate numbers look fine.

The second trap is leaving Smart Bidding on inside a tiered structure. If your low-priority campaign uses tROAS while the high-priority one bleeds it of the high-intent traffic, the bidding model is optimizing against a distorted, pre-filtered stream. Either commit to manual bids across tiers (the honest version of the trick) or abandon the structure. The hybrid is the worst of both.

Honest limit: Google does not tell you, in the UI, which campaign won a given query when overlap exists. You infer routing from search terms and campaign-level reports (Campaigns > Insights and reports > Search terms), but there is no 'this query was routed here by priority' flag. That opacity is exactly why orphaned negatives go unnoticed. If you cannot commit to auditing your negative lists across every tier on a set cadence, do not build the structure.

Questions people also ask