CSS Google Shopping: How It Works & Why It’s Still Mandatory

I’m hearing a dangerous myth crop up more and more. It started towards the end of last year—first from one prospective client, then another, and then a third, all repeating advice from different agencies: “CSS doesn’t matter anymore.”

They are absolutely wrong.

At Savvy, we’ve run every possible CSS configuration you can think of: solo CSS, hybrid setups, and countless head-to-head tests against Google. I’ve consulted with CSS providers and affiliates. I come at this from every single angle, and here’s what’s actually happening: CSS still provides a 16% to 18% auction advantage in EU markets. But because this advantage is completely invisible in your Google Ads account metrics, many agencies assume it has stopped working.

Let’s set the record straight. If you’re running competent Shopping campaigns in the EU and you’re not using a CSS, you are giving your competitors a structural advantage. Full stop. However, a quick disclaimer: using a CSS will not fix a broken campaign structure. If your Shopping setup is fundamentally flawed, fix that first.

Let’s Get a Few Things Straight: What We Know for Certain About CSS

There’s a lot of talk about conspiracies and theories, but let’s stick to the facts. Here’s the foundation of how this all works.

First, Google was forced by the EU to make Google Shopping a separate business entity with its own revenue requirements. To comply, Google began giving 20% of each Shopping ad click’s CPC to this new “Google Shopping” entity. This is the key.

When you use a third-party Comparison Shopping Service (CSS), that 20% margin doesn’t apply. This is the source of the auction advantage. This only applies in the European Union (plus Switzerland and the UK), not Europe as a whole.

So what’s the actual saving? When this first rolled out, Google’s own documentation stated the margin was exactly 20%. That documentation has since vanished. My sources in the industry (who I will not name) say the number is now closer to 16-18%, but that it is absolutely still in effect today. The only reason the documentation was removed is because Google is no longer forced to disclose the exact percentage publicly—not because the advantage is gone.

Finally, there is no fundamental difference between CSS provider A and CSS provider B in terms of performance. The only differences are the additional services they might offer (like feed optimization) and their name. I personally avoid the ones with cheap, scammy-sounding names, but that’s just me.

What’s Changed? Dispelling Old CSS Myths

The process of using a CSS used to be a lot more cumbersome, which is where some of the bad advice comes from. Here’s what used to be true but isn’t anymore:

• You DON’T need a new Merchant Center. You can simply convert your existing one and retain all your product history, disapprovals, and performance data. It’s a simple switch.
• You DON’T need two Merchant Centers. Free listings can now run directly through your CSS Merchant Center. This is a relatively new development that cleans things up immensely. No more managing two separate accounts.
• You should NEVER pay a percentage of ad spend for a CSS. This was an old, predatory model. Today, you should only pay a small flat monthly fee (think €20-€30). If someone is trying to charge you a percentage, run.
• There is no “kickback” anymore. In the very early days, there was a spend-match incentive. That program has been gone for years.

Why You Can’t “See” the CSS Discount in Your Account

This is the main reason so many people are confused. The advantage isn’t a line-item refund in your billing or a visible “discount” metric. It’s an auction advantage, and it gets absorbed in three key ways.

1. It’s a Bidding Advantage, Not a Refund
The 16-18% is not an amount that gets returned to you. Instead, it allows you to bid more competitively. In an auction, a €1.00 bid from a CSS advertiser is effectively treated as a €1.20 bid (using the old 20% for simplicity) compared to a non-CSS advertiser. You either win more auctions at your current bid level or win the same auctions for a lower cost.

2. Smart Bidding Immediately Reinvests the Savings
As soon as you make the switch, Smart Bidding notices you’re hitting a higher ROAS than your target. What does it do? It starts increasing your bids to capture more volume at your target ROAS. For most advertisers, the result isn’t a higher ROAS, but rather an increased ability to bid for more auctions and gain more impression share. You *could* technically increase your ROAS target by about 10% when you switch and likely retain your current volume, but most just let the algorithm push for more volume.

3. The Advantage Diminishes in Highly Competitive Auctions
If all your primary competitors are also using a CSS, the playing field is leveled again. There’s no effective “saving” because everyone has the same advantage. However, this doesn’t make CSS less important. It makes it mandatory. If everyone else is using it and you are not, you are operating at a permanent 16-18% disadvantage.

A Quick Look at the Data: Old vs. New Tests

Back in the days of manual bidding, it was easy to prove this. We would run two identical campaigns, one through Google Shopping and one through a CSS, with the CSS campaign having a 10% lower bid. The results were clear: the CSS campaigns consistently produced a better ROAS.

This kind of test is no longer possible with Smart Bidding, but we saw the effect clearly in a 2024 account takeover. We had to manually switch Merchant Centers. As soon as we moved the campaigns over to the CSS Merchant Center, we saw CPCs drop while the algorithm maintained the target ROAS. We were able to produce better results simply by making the switch.

The Problem with CSS Affiliates

This is a topic that agencies often have a hard time articulating. The concept is that you allow an affiliate partner to set up their own CSS Merchant Center and run Shopping ads on your behalf, alongside your own campaigns. They typically work on a commission or “no cure, no pay” model.

The only time this is a good idea is if you are not confident in your own Google Shopping setup. In that rare case, a highly skilled affiliate might create a more efficient structure that can outperform what you can do. In my experience, this is the exception, not the rule.

At Savvy, we do not allow CSS affiliates. We’re on the blacklist of several affiliate networks because of it. Here’s why it’s almost always a bad idea.

The Illusion of “Free” Performance

Let’s say you give an affiliate a 10% commission. They will set a very high ROAS target—say, 1,100%—so their 10% commission is covered and they make a profit. Meanwhile, you might be running your main account at an 800% ROAS target.

You might look at the 1,100% ROAS from the affiliate and think it’s fantastic. But that’s a mistake. The affiliate CSS is effectively just cherry-picking the auctions your own campaign could have easily won at a 1,100%+ ROAS and outbidding you for them. The problem is even worse if you haven’t switched to a CSS yourself. The affiliate gets a 16-18% bidding advantage while stealing your best auctions.

Loss of Control and Spotty Data

With an affiliate, you have no control. They can decide to lower their bids overnight to increase their margins, causing your overall volume to drop without warning. I’ve seen this cripple businesses during audits.

This leads to the biggest issue: your data becomes unreliable. When you see a drop in volume, is it because your affiliates are taking over, or are you genuinely losing ground to competitors? The actions you should take are completely different for each scenario, but you can no longer trust the data to tell you which is which. For any skilled PPC manager, losing data integrity is a dealbreaker.

The Right Way to Handle It

If your boss or client is convinced that a second CSS will deliver incremental revenue, just do it yourself. There’s no need for an affiliate to manage it for you unless you lack confidence in your own abilities.

When You Actually Shouldn’t Bother with CSS

After spending this entire article telling you CSS is mandatory, let me take the opposite side for a moment. There are three specific situations where it might not be worth the minimal effort.

1. You have a very small budget. If you’re spending under €5,000 per month, the gains will be so small that your time is better spent elsewhere. I would still do it on every account, but if setting it up feels like a struggle, don’t worry about it.
2. You only advertise in non-EU markets. Obvious, but I’ve seen people pay for a CSS for campaigns in the US or Australia. It does absolutely nothing there.
3. You already dominate your auctions. If your data shows you are consistently bidding 50-100% higher than your competitors, then you’re not at a competitive disadvantage. You’re already winning the auctions you want to win.

Even in that last case, I would still do it. The process is so simple today. Just pick a reputable CSS provider for under €50 a month, convert your existing Merchant Center, and get back to actual Google Ads work. There is no downside, and the advantages are clear.

[TL;DR]

• CSS is not dead. It still provides a real 16-18% bidding advantage in EU markets, but it’s invisible in your account metrics because it’s an auction advantage, not a refund.
• The advantage comes from an EU ruling that forces Google Shopping to operate as a separate entity with a profit margin that third-party CSS providers don’t have to charge.
• Smart Bidding automatically uses the CSS advantage to bid more aggressively and win more impressions, which is why you see more volume, not necessarily a higher ROAS.
• Avoid CSS affiliates. They create data integrity issues and often just cherry-pick the auctions your own campaigns could have won, giving you an illusion of performance.
• Implementation is simple: convert your existing Merchant Center for a small flat monthly fee. Unless your budget is tiny or you’re outside the EU, it’s a mandatory part of a competitive Shopping setup.