Will a CSS partner cannibalise my own campaigns? The mechanics of second-price auctions

The objection that comes up in every meeting

If you run Google Shopping campaigns in-house, the first question you will ask any CSS partner is: "Won't you just bid against me and push my CPCs up?" It is a reasonable concern. On the surface, two entities placing bids for the same product on behalf of the same merchant sounds like a recipe for self-inflicted inflation.

The short answer is no, it will not happen — and the reason is structural, not commercial. Google's auction mechanics contain a specific rule that prevents a merchant from competing against itself. Understanding that rule is worth five minutes of your time, because it changes the entire risk calculus of working with a CSS.

How Google Shopping auctions actually work

Google Shopping uses a variant of the generalised second-price (GSP) auction. When a user searches for something like "men's waterproof hiking boots", every merchant with a matching product in their feed enters the auction. Each bid is associated with a CSS — either Google's own CSS (Google Shopping Europe) or a third-party CSS partner.

The auction determines which ads appear and in what order. Crucially, the price you pay per click is not your own bid — it is the minimum amount needed to maintain your position above the next-highest bidder. This is the second-price mechanic, and it is the same principle that underpins most of Google's ad auctions.

The deduplication rule

Here is the part that matters for the cannibalisation question. Google applies a deduplication rule at the merchant level. If the same merchant has bids entering the auction through multiple CSS partners (or through a CSS partner and Google's own CSS), Google does not let those bids compete against each other. Instead, Google selects the highest bid from among the merchant's entries and enters only that bid into the auction.

The other bids from the same merchant are simply discarded from that auction instance. They never push the winning bid's CPC higher, because they never participate in the price-setting mechanism.

This is not an optional feature or a setting you need to enable. It is how the system works by default, and it applies to every auction across every product.

A worked example

Suppose you sell a pair of running shoes at £85. Your in-house Shopping campaign bids £0.45 for this product through Google's own CSS. Your CSS partner, operating through an affiliate network, bids £0.38 for the same product through their registered CSS.

Here is what happens in the auction:

1. Google identifies both bids as belonging to the same merchant (matched via your Merchant Centre ID).

2. Google selects the higher bid — your in-house bid of £0.45.

3. The CSS partner's bid of £0.38 is dropped from the auction entirely.

4. Your £0.45 bid competes against other merchants in the normal second-price auction. Your CPC is determined by the next-highest bid from a different merchant — say, £0.41.

5. You pay approximately £0.41, not £0.45.

The CSS partner's bid had zero effect on the price you paid. It did not inflate your CPC by a single penny.

When the CSS partner's bid wins

Now consider a different scenario. Your in-house campaign does not bid on a particular product — perhaps it is a long-tail SKU that falls below your ROAS threshold, or your campaign structure simply does not cover it. The CSS partner bids £0.32 for that product.

In this case:

1. Only the CSS partner's bid enters the auction for your merchant.

2. If it wins, the click goes through the CSS partner's listing.

3. You pay nothing for that click — the CSS partner covers the CPC.

4. If the click converts, you pay a commission through your affiliate network, on a CPA basis.

This is the scenario where CSS partners add the most obvious value. They pick up traffic on products and queries where your own campaigns are not present, at no upfront cost to you.

What about overlapping bids at similar levels?

A fair question: what happens if both bids are very close, say £0.44 (in-house) and £0.43 (CSS partner)? The same rule applies. Google picks the higher bid (£0.44) and drops the other. The closeness of the bids is irrelevant — there is no scenario in which both enter the auction.

Some retailers worry about "auction pressure" from the mere existence of a second bid. This concern conflates how the auction works with how it might work if the deduplication rule did not exist. The rule does exist. It has been in place since the EU's 2017 antitrust decision reshaped how Google operates Shopping in the European Economic Area.

The real question: where does a CSS partner add incremental value?

Once you accept that cannibalisation is mechanically impossible, the question shifts to whether the CSS partner is actually generating incremental traffic and revenue.

There are three common scenarios where they do:

1. Coverage gaps

Most in-house Shopping campaigns do not bid on every SKU in the feed. Long-tail products, low-margin items, and seasonal stock often fall outside campaign structures. A CSS partner bidding across the full catalogue picks up clicks you would otherwise miss.

2. Query-level gaps

Even for products you do bid on, your bids may not be competitive for every search query. A CSS partner using different bidding logic — often informed by affiliate conversion data — may win auctions on queries where your in-house bid would not have shown.

3. The margin benefit

When a CSS partner places your products through a third-party CSS rather than Google's own CSS, the auction margin that Google retains is lower. This means a lower effective CPC for the same ad position. The saving is modest on a per-click basis, but across thousands of clicks per month, it adds up.

How to verify this yourself

If you want to confirm that cannibalisation is not occurring, the data is available:

• Check your Google Ads CPC trends before and after enabling a CSS partner. If cannibalisation were happening, you would see CPC increases on overlapping products. In practice, you will not.
• Ask your CSS partner for impression and click data broken down by product. Compare this against your in-house data to see where the CSS is winning auctions versus where you are.
• Look at overall revenue and ROAS, not just channel-level metrics. If the CSS partner is generating genuinely incremental sales on a CPA basis, your blended performance improves regardless of which entity won any individual auction.

A note on trust, not just mechanics

The deduplication rule is a mechanical safeguard. It prevents cannibalisation at the auction level. But it does not address every concern a retailer might have about working with a third party.

You should still ask your CSS partner how they set bids, which products they prioritise, and how they handle products where you are already dominant. A good CSS partner will be transparent about their bidding strategy and will actively avoid wasting their own budget on clicks you would have won anyway — because that is bad economics for them too.

The cannibalisation question is the right question to ask. The answer, once you look at the auction mechanics, is straightforward. The more productive conversation is about incrementality: where can the CSS partner win clicks you would not have won on your own, and at what cost?

That is where the real value sits.