The Expensive Problem with Google’s Search Partner Network (SPN): Insights from 1,500+ Campaigns

Episode Transcript

Mike 00:00:00: Hey, Chris. Hi, Mike. Welcome to another episode of Growing Ecommerce, brought to you by Smarter Ecommerce, also known as smec. I am one of your hosts, Mike Ryan, and with me is Chris, our other host.

Chris 00:00:15: Yeah, welcome back. Let’s jump right into it today. We have some great topics.

Mike 00:00:23: That’s right. Last episode, we were talking about how PMax is not a black box anymore, and we want to present some of the fruits of that transparency—even if Google might regret it. For those who missed the last episode, we were very positive about Google being open with data, and today we will be even more so.

Chris 00:00:49: Let’s make sure we don’t hold back. The numbers are really—what is the right word? Concerning.

Mike 00:01:05: Concerning is the word. And they’re not surprising at all, unfortunately. Network reporting, also called channel reporting, has been available for a long time in Google Ads. One of those networks that I’m not the biggest fan of is the Search Partner Network (SPN).

Chris 00:01:33: You are not the biggest fan? What does that mean?

Mike 00:01:36: The premise of the Search Partner Network is actually great. The idea is that Google is not the only search engine in the world. There are all these different publishers and websites that have built-in search engines, and if a commercially relevant search is made with intent, your text ads or shopping ads can appear on those properties. eBay was one of the big hallmark search partners at the start, and that’s cool.

Chris 00:02:08: There is nothing wrong with that.

Mike 00:02:11: The problem is that Google doesn’t really do a lot of quality control or gatekeeping. It’s the same problem we see with the Google Display Network: “more is more.” I don’t know if there’s an official number of search partners, but it’s likely in the millions.

Chris 00:02:36: We talked about that. I think this is one of the reasons why it feels like it’s not under control anymore—it might be too big even for Google.

Mike 00:02:49: We discussed that back with the CSAM scandal last year. Google didn’t even know everything that was in there, which is concerning. The Search Partner Network is our focus today. Again, shout out to Google; the reason we can even talk about this is because they opened up the black box.

Chris 00:03:13: Exactly. All good data and numbers to discuss today. One thing I’d like to put on Front Street is that you probably have similar issues with Microsoft. The numbers might not be as relevant for many clients, which is why we focus on Google, but the Search Partner Network concept is generally great; it’s just that big platforms struggle to ensure quality.

Mike 00:03:45: Yes, and it’s not just Google. We will see this topic rising more with Amazon and their off-sites. This is a problem within digital marketing at large. Google is by far the biggest player, which is why we focus on them.

Chris 00:04:03: Perhaps we should show the data?

Mike 00:04:06: First, I want to lead into this quickly. I’m going to put up data showing the development of this reporting over time. We pulled network data for hundreds of advertisers for all of 2025 and through January 2026. Until May 2025, that was the classic “black box” era where everything was categorized as cross-network. We knew that shopping and search—the Google Search Network—was the biggest part of PMax, but now we can see Google-validated numbers for the rest. We found that the scripts popular a couple of years ago were actually quite inaccurate.

Mike 00:05:06: Google has provided accurate data starting in June. Excluding the Google Search Network, it’s mostly Display and YouTube, which isn’t too surprising.

Chris 00:05:34: I would say that’s fine.

Mike 00:05:36: I have the numbers here. The median amount of impressions, excluding the Google Search Network, is 49% YouTube. So, it’s half YouTube. This has been significant for YouTube sales.

Chris 00:05:58: No surprise there, as it’s one of Google’s strategic focal points. YouTube is the channel where they want to tap into classic TV budgets; it’s the next growth story for Google. So, for me, it’s not surprising at all. And by the way, I love YouTube.

Mike 00:06:20: Absolutely. Then we have 17% on the Google Display Network, which includes a lot of remarketing, and 17% on the Search Partner Network. That’s why I started looking into this. We see 7% on Discover and 11% on Maps. For multichannel or omnichannel retailers, Maps plays a massive role.

Chris 00:06:56: Exactly. If you want to target local intent, it’s a great tool.

Mike 00:07:05: We see a very small amount, half a percent, for Gmail. It’s great that this data is available so we can see the performance of each. But I have a special request for Google: let us opt out of the Search Partner Network in PMax.

Chris 00:07:28: That’s reasonable. The data we are going to discuss makes it very reasonable. What do you think is the reason why Google is hesitant to give us the possibility to opt out? Is it as straightforward as revenue impact?

Mike 00:08:04: While the Search Partner Network generates a sizable amount of impressions, the intent is often low. That causes bad performance, but it also means it doesn’t generate that many clicks or costs. You can almost overlook it, but it’s just bloat for many advertisers. I think the reason Google doesn’t allow an opt-out is that the core identity of PMax is being a “one-stop shop” for all Google inventory. They want to hold the line there, because if they offer an opt-out for one, what’s next?

Chris 00:09:01: That’s where my question comes from. In the Google “Power Pack”—PMax, Search, and Standard Shopping—there is one campaign where you have full control: Standard Shopping. You can literally say where you want to be. In PMax, you can’t.

Mike 00:09:34: They don’t want to open the floodgates.

Chris 00:09:40: You keep fighting for that, right? To all our fans—if we have some—one of them told me at a convention two weeks ago that if there is ever a possibility to opt out, it’s because of Mike.

Mike 00:10:03: I’m making a concerted effort now. I’m writing an op-ed piece for Search Engine Land. That’s why I’m preparing this data; it will help everyone.

Chris 00:10:25: Let’s talk about the numbers.

Mike 00:10:38: I tried to be fair here. I only looked at campaigns with 30 or more monthly conversions, which should be enough for smart bidding to learn. Google has been trying to make improvements to make the Search Partner Network more performant, but right now it’s not showing. I want to show the range of outcomes rather than a singular monolithic average.

Mike 00:11:45: I’m working with something called a “trimmed mean,” which excludes the top and bottom outliers to provide a robust and fair number. Unfortunately, the results are still garbage. Looking at the trimmed mean for 1,500 PMax campaigns: CTR is 40% lower, conversion rate is 63% lower, and CPCs are 14% lower. In Google’s defense, these clicks don’t cost as much, but it’s not proportional.

Mike 00:13:00: The cost per acquisition (CPA) is 73% higher in the trimmed mean case, and the ROAS is 63% lower.

Chris 00:13:11: That is horrific across the board. There are some advertisers for whom this works, and the only saving grace is that it’s usually not a huge spend. But having this data available and not being able to do anything about it feels like torture. You can exclude single placements, but there are limits. If Google is serious about the incrementality of your return, this doesn’t make sense.

Mike 00:14:39: Google is hesitant to disclose this data. Their help center suggests that these were the cheapest conversions available at the time. But my point is: don’t buy that conversion if the intent is too low.

Chris 00:15:22: These PMax campaigns typically meet or exceed ROAS targets, but this is just weird bloat dragging down the average. Based on the data, these conversions aren’t paying off. It’s costing money for no reason.

Mike 00:16:15: It’s horrific. Google has to do something. Even if it’s not a huge chunk of revenue for Google, the risk-reward ratio isn’t there, especially given past brand safety scandals. If I look at placement reports—and we have a script to help evaluate them—I don’t believe there is any human traffic on some of these websites.

Mike 00:17:08: I also looked at the invalid traffic rate per network. There are harmless bots, and then there is malicious invalid traffic. The Search Partner Network, unsurprisingly, has the highest share of malicious fraud or fraudulent clicks.

Chris 00:17:34: Super interesting data, Mike. Keep fighting the good fight. I want people to like PMax, but I want these reasonable objections resolved. We want people to like Google because it’s a great company, but this is worth the discussion.

Chris 00:18:20: Let’s move to our news section. Mike, you had a great webinar with Ginny Marvin recently. Shout out to her for joining; it was great to see so many registrants and discussions. What was the big news?

Mike 00:19:00: The biggest piece of news was when I asked Ginny about Dynamic Search Ads (DSA). There is a redundancy between DSA and PMax, and in many ways, PMax is superior. DSA is a popular product—it’s probably 20% of search—and knowing its future helps people decide whether to adopt PMax. Ginny confirmed that DSA’s time on Earth is limited.

Mike 00:19:45: We don’t have an exact date or a detailed migration plan yet, but DSA will eventually go the way of the dinosaurs. PMax is the future-proof technology for that kind of advertising.

Chris 00:20:23: I love the dinosaur analogy. The sunset of DSA might be for the greater good. It was a great campaign type, but it’s not surprising given the redundancy. I was surprised she was so concrete on that topic.

Mike 00:21:00: There hasn’t been an official announcement from Google yet, but Ginny speaks from a position of authority and knew this would give people clarity.

Chris 00:21:15: There is a timeline coming to an end. Google typically gives a one-year sunset period, and I expect there will be migration plans since the technologies are so similar.

Mike 00:21:47: It was a pleasure, Chris. Thanks to our audience for their time. As always, Growing Ecommerce is brought to you by Smarter Ecommerce, also known as smec. Please visit us at smarter-ecommerce.com to learn more.

Chris 00:22:13: Give us a follow, a recommendation, or a review; it means a lot to us.

Mike 00:22:20: See you next time.

Chris 00:22:24: Thanks, Mike. Bye.