How the Shopping auction works
What actually decides which of your products show on Google Shopping, in what position, and at what cost, the feed, your bid, and Quality, plus the second-price idea that sets what you really pay.
Every time someone searches, Google runs a split-second auction to decide which product listings show, in what order, and what each click costs. You don't enter keywords, your feed decides which searches you're eligible for, and then your bid and a quality signal decide whether you win. The second-price idea means you usually pay just enough to beat the listing below you, not your full bid.
The auction happens every time someone searches
There is no fixed "Shopping results page" waiting to be filled. When a shopper types a query, Google assembles a fresh auction in a fraction of a second: it works out which products in the world are eligible, ranks them, and decides which listings show, in what order, and what each click will cost. Your product is one entrant among thousands, and it has to earn its place every single time.
Three inputs decide the outcome:
- Your feed: determines which auctions you're eligible for at all.
- Your bid: how much you're willing to pay for a click.
- Quality: how relevant and clickable Google expects your listing to be.
Miss on any one and the other two can't fully rescue you. A perfect bid on a product that isn't eligible for the query never enters the auction; a cheap, eligible product with a weak title loses to better-matched rivals.
Step one: eligibility is decided by your feed, not by keywords
This is the part that trips up merchants coming from Search ads. In Shopping there are no keywords to bid on. You never tell Google "show me for waterproof hiking boots". Instead Google reads your product data (title, product type, google_product_category, attributes like colour and size, and the image) and infers which queries each product is relevant to.
So your feed is your keyword list, whether you meant it to be or not. If "waterproof" isn't in the title or attributes, the product may simply not be eligible when someone searches for waterproof boots. No bid, however high, changes that. This is why feed work is the highest-leverage thing you can do before touching bids: it sets the size of the pool you're allowed to compete in. See feed quality and quality impressions for how that eligibility turns into impressions.
Step two: ranking combines bid and Quality
Once you're eligible, Google ranks the entrants to decide who shows and in what order. Ranking is not "highest bid wins". It's your bid combined with a quality signal: Google's estimate of how likely shoppers are to click and how relevant and trustworthy your listing is. Broadly, quality is built from:
- Expected click-through rate: will people actually click this listing for this query?
- Feed relevance: how well your title, attributes and image match what was searched.
- Landing-page experience: does the product page load, match the ad, and let people buy?
The practical consequence: a well-matched product with a strong feed can outrank a competitor who bids more. Quality is the multiplier that lets a smaller budget punch above its weight: and a weak feed is a tax you pay on every auction.
Step three: the second-price idea sets what you actually pay
Winning the auction doesn't mean paying your full bid. Shopping uses a second-price style mechanism: you pay only just enough to beat the entrant ranked directly below you, adjusted for the quality difference between you. Your maximum bid is a ceiling, not the price.
Two things follow, and they're worth internalising:
- Higher quality lowers your cost to hold a position. If your listing is more relevant than the rival below you, you can outrank them while paying less per click than they would to take your spot. Feed quality doesn't just win auctions: it makes the ones you win cheaper.
- You're priced against your competitors, not against your own bid. Two products with identical bids can pay very different CPCs, because what matters is who's below them and how good that rival is.
Deep dive A worked auction, step by step
Imagine a shopper searches "kids waterproof wellies size 10" and three products are eligible. Google assigns each an internal rank roughly proportional to bid × quality (real ranking is more involved, but this captures the logic):
| Product | Max bid | Quality (relative) | Rank score |
|---|---|---|---|
| A (title "Kids Waterproof Wellington Boots) Size 10, Navy" | £0.40 | High (1.5) | 0.60 |
| B: title "Wellies" | £0.60 | Low (0.7) | 0.42 |
| C: title "Children's Rain Boots, Waterproof" | £0.45 | Medium (1.1) | 0.50 |
Product A wins the top slot despite bidding the least, because its feed matches the query closely: "waterproof", "size 10" and the product type are all present. Product B bids the most and loses position, because "Wellies" tells Google almost nothing and its quality drags its rank down.
Now the price. Under second-price logic, A pays only enough to stay ahead of C (the entrant below it), scaled by the quality gap: well under its £0.40 ceiling, say around £0.34. B, sitting bottom with poor quality, ends up paying near its full £0.60 bid to show at all. The merchant with the worst feed is bidding highest and paying most, for the lowest position. That's the entire argument for feed quality in one table: a clean title on Product A both won the slot and cut its CPC.
This is also why raising bids to fix a visibility problem so often disappoints. If the real issue is that your titles don't match the query (an eligibility and quality problem), more money buys you a worse position at a higher price. Fix the feed first, then bid.
What this means for how you spend your effort
The auction rewards, in order: being eligible, being relevant, then being willing to pay. Bidding strategy (Target ROAS, Maximise Clicks and the rest) only redistributes spend across auctions you already qualify for. It can't enter you into auctions your feed locked you out of, and it can't make a poorly-matched listing cheap.
So the sequence that works is: get the feed clean so you're eligible for the right queries and carry strong quality into them, structure your campaigns and product groups so you can bid different products differently, and only then tune bids. A strong Feed Health Score is, in auction terms, a wider eligibility pool and a lower cost to win each place in it. BidSmart's product-level bid analysis works on top of that: surfacing where each product's bid is leaving position or margin on the table, staged for your approval before anything changes.
Frequently asked questions
I never chose keywords, how does Google know which searches to show my product on?
Your feed does it. Google reads your titles, product types, categories, attributes and images and infers which queries each product is relevant to. That's why a weak title doesn't just look bad, it narrows the set of auctions you're even allowed to enter.
Does the highest bid always win the top spot?
No. Google ranks on bid combined with a quality signal (expected click-through, feed relevance and landing-page experience). A well-matched product with a strong feed can outrank a higher bidder, and it usually pays less per click to do it.
Why do two products with the same bid get different CPCs?
Because the auction is second-price and quality-weighted. What you pay depends on the competitor directly below you and their quality relative to yours, not on your own bid. Better feed relevance lowers what it costs you to hold a position.