Released:
Why did OpenAI quietly kill its shopping feature, and what does it mean for the future of Google Ads? 🚨
In this episode of Growing E-commerce, Mike Ryan and Chris Scharmüller tackle the reality of the “AI takeover” in paid search. Following Chris’s exclusive executive Think Tank with 20 leading CMOs in London, we are cutting through the noise to discuss the 2026 playbook.
First up: The ChatGPT Shopping disaster. We discuss why OpenAI had to walk away from its “thin Google Shopping wrapper” to save its burn rate, and why Google’s ad ecosystem dominance is stronger than ever. We also dive into the “Wild West” of agentic commerce, including Amazon’s massive legal battle with Perplexity over spoofed traffic and lost ad revenue.
But it’s not all drama! The second half of this episode is a pure, actionable deep dive into the future of Google Product Feeds. With AI search queries getting 2.5x longer, we break down exactly how to future-proof your account using Google’s new conversational attributes, semantic data, and business signals.
Balancing AI Hype and Revenue Reality
Chris Scharmüller emphasizes that while consumer interest in AI-powered search is rapidly increasing, it has not yet materially disrupted the core advertising ecosystem’s revenue contribution. He cautions ecommerce leaders against pivoting the majority of their top talent to agentic commerce too early, as foundational search campaigns and platform integrations will remain the dominant revenue drivers for the foreseeable future. Instead of a total shift in focus, businesses should prioritize future-proofing their data feeds with semantic knowledge to provide the “grounding” material necessary for artificial intelligence systems to accurately synthesize product information.
00:00:00:03 - 00:00:14:12
Mike: Welcome to Growing Ecommerce. I am one of your hosts, Mike Ryan, and with me, as always, is Chris Scharmüller, the man. But I leave out “the formula” because we agreed on it; it’s just not sexy at all.
00:00:14:14 - 00:00:36:21
Chris: Yeah, it’s just Chris. And I’m just plain Mike. But you have a great last name, Ryan. Do you know how hard it is to find an email address when your name is Mike Ryan? I had no issues with christian.scharmueller.
00:00:36:21 - 00:01:01:06
Mike: It’s always the longest email address there is. But Ryan has some vibe to it. I appreciate it, Chris. But again, Chris is fine. Let’s go into it. One day, on the Christmas episode this year, we’ll talk about email addresses, the sexiness of last names, or how long it takes you to fill out paperwork.
00:01:01:06 - 00:01:26:22
Chris: You must run out of boxes all the time. And how tough it is to let Austrians understand how to spell Ryan. I think I talked about that before; it’s very hard to explain. But you know what’s hard as well? To understand what is just hype and what is plain reality when we talk about AI and how much AI will impact search.
00:01:26:23 - 00:01:46:23
Mike: You dropped a hell of a transition there. That was so smooth. I had some time to prepare. What are we going to talk about? Well, Chris, let’s think about the last couple of episodes.
00:01:46:23 - 00:02:08:24
Mike: We talked about a report where I was the principal analyst called The Ultimate Guide to PMax. Then we talked about a report put together with some partners called the PPC Survey. Now it’s time to put you in the hot seat. You did a boardroom presentation recently at a think tank.
00:02:09:01 - 00:02:25:22
Mike: Excuse me, that sounds way cooler. I told him this was my intro, guys, the pressure is on. I feel like the CIA has think tanks. So tell us about this think tank session, Chris.
00:02:26:00 - 00:02:52:06
Chris: We talked briefly about it; it was in London as part of a convention we attended. I had the chance to talk to roughly 20 high-level execs of leading online retailers, most of them active in Europe. One of the major topics was to understand what is noise and hype versus what is reality.
00:02:52:06 - 00:03:20:00
Chris: What can we learn from it with regards to the paid search playbook for 2026 and beyond in order to be on the winning side? That was the frame. I think I did quite well because it was very engaging to talk with these CMOs. They were concrete conversations. We can go through a couple of items here because not everything is reality.
00:03:20:02 - 00:03:45:21
Chris: There’s a lot of hype going on. Tell us a bit about what’s hyped and the reactions you got. In one of our last episodes, we discussed three or four things I would like to go into today in more detail.
00:03:55:23 - 00:04:16:03
Chris: We talk all the time about how AI is going to disrupt paid search. As a matter of fact, I think it requires a nuanced discussion. If we think of disruption in a way that people start to use generative AI services to get information for their purchasing journey, I think that statement is true.
00:04:16:05 - 00:04:43:06
Chris: There’s a lot of data which supports that claim. For instance, there’s a McKinsey study done last year in August. The question was: what is the preferred information source for your purchasing journey? The results were heterogeneous.
00:04:43:08 - 00:05:01:08
Chris: The two dominating layers were classic search engines at 31%, but already 44% of the people asked claimed that AI-powered search is their preferred information source for the purchasing process. This is attributed to Google and others.
00:05:01:08 - 00:05:42:12
Chris: So, 44% clearly prefer it. This supports the claim that people are adopting these new AI services to process information. That claim is reality. We can discuss how you see it, but it’s certainly not supporting the claim that it disrupts the way you should do your search marketing right now, because there is data indicating that this is not happening yet.
00:05:42:15 - 00:06:14:22
Chris: I think it’s hard to draw a line between disruption and transformation; it might be a matter of perspective. If you’re the benefactor, it’s a transformation. If you’re on the losing side, you feel disrupted. I agree there’s a big shift happening that we have to call a transformation.
00:06:14:24 - 00:06:33:14
Mike: To call it a disruption, we don’t know that yet. There needs to be a certain amount of speed and impact that is not clear yet.
00:06:33:14 - 00:06:59:02
Chris: Just because we look at the massive trajectory here doesn’t mean it impacts your core campaign revenue contribution significantly enough to massively shift your priorities. This was one of the conversations I had with the key players. They asked how to shift the balance now.
00:06:59:02 - 00:07:16:18
Chris: We have our core campaign setup, which is PMax, standard search, and all that, and we see this trajectory of people using other surfaces. This is the major question for online retailers: how to balance that shift.
00:07:16:19 - 00:07:43:06
Chris: The statement I really believe in is that balancing focus is more important than ever because the future of search will look different. But we are right at the beginning. The ecosystem you’re operating in is not changing overnight. That is the main claim. Prepare yourself.
00:07:43:06 - 00:08:18:00
Chris: Your core campaigns will be the major revenue drivers for the foreseeable future. Google and advertisers both have skin in the game here. This is a large majority of their advertising revenue, so it has to be managed.
00:08:18:00 - 00:08:42:09
Mike: The more dominant they become, the more they control that pace and the less they have to react to challengers. There are concerns about an incumbent becoming more powerful, but it’s always a trade-off. There are benefits as well.
00:08:42:11 - 00:09:06:16
Chris: Alphabet is a Nasdaq-listed company; top and bottom lines matter to them. They won’t do massive disruptive things to their advertising ecosystem. So, yes, there is a massive trajectory where people adopt these new generative AI services, but that doesn’t mean we are flipping the script from a classic advertising ecosystem to a full agentic commerce world overnight.
00:09:06:18 - 00:09:30:10
Chris: We have to understand that. There’s one case I found really funny which is very telling of where we are right now: the Amazon versus Perplexity situation. It shows that agentic commerce isn’t fully solved yet.
00:09:30:10 - 00:09:51:16
Chris: In this case, Amazon is not the villain, but the victim. Perplexity has a product that offers an end-to-end shopping assistant, and that assistant is buying on Amazon.
00:09:51:18 - 00:10:14:15
Chris: Amazon claims there are areas of conflict. First, their terms of use are for billing. Amazon says whatever Perplexity does is very close to an item, and on top of that, it’s spoofed. Amazon doesn’t even know it’s an agent.
00:10:14:18 - 00:10:41:04
Chris: It’s spoofed as Google Chrome traffic. Someone has given authorization, so it looks authorized, and it’s bypassing all the safety guardrails of Amazon. The identity is not clarified. More importantly, Amazon’s motivation for going to court is their ad ecosystem.
00:10:41:04 - 00:11:10:16
Chris: Recommendation engines don’t work with agents. If there’s no human being clicking a buy button and everything is running through an API, billions of dollars of advertising revenue are at risk. Then there’s the question of accountability if something goes wrong.
00:11:10:22 - 00:11:34:08
Chris: We’re in a Wild West situation. As long as there is no clarity, agentic commerce won’t happen overnight. It’s the future of search to a second degree, but it’s not happening tomorrow.
00:11:34:10 - 00:12:03:03
Chris: It’s funny with Amazon as well because there was another news story about small business owners complaining that Amazon is selling their products without permission. Amazon says that’s not cool, but then they do it themselves. If I were Perplexity’s legal team, I would use that to undermine Amazon’s case.
00:12:03:03 - 00:12:23:01
Mike: It’s a mess. To give context, Amazon has a feature called Rufus. They have a shopping assistant that searches Amazon, but if the product is not available, it goes to another website. Those websites are stunned because they get sales from Amazon without being listed on it.
00:12:23:07 - 00:12:42:02
Chris: I think it’s stupid. Maybe Amazon just feels better being the villain. Jeff Bezos looks like Lex Luthor. We’re in this permissionless era where everyone’s taking liberties and then complaining when liberties are taken against them.
00:13:27:18 - 00:13:55:08
Mike: I read about Anthropic and Claude; they’re supposed to be the moral high ground. They had a program to buy and destroy books back when they were training their models. A team would buy books, rip the covers off, and scan the pages at scale.
00:13:55:14 - 00:14:15:01
Mike: They turned all of human knowledge into a text slurry for their models. Those are supposed to be the “good guys” in all this.
00:14:15:03 - 00:14:39:07
Chris: Every business with limited resources—not just money, but smart people—shouldn’t do a full switch. Don’t focus 90% of your best people on agentic commerce because the data shows it’s not happening yet.
00:14:39:08 - 00:15:00:04
Chris: Let’s talk about ChatGPT and Sam Altman. You know my feelings. I worry about that guy’s personal safety.
00:15:00:06 - 00:15:22:17
Mike: After just a few months, they stepped back from shopping, which was one of their only paths to making money. It turns out they were basically just scraping Google Shopping.
00:15:22:17 - 00:15:48:01
Mike: Imagine you’ve invested tens of billions in a company and their main monetization strategy is to be a thin Google Shopping wrapper. Your competitor could stop that at any time. I don’t know why they allowed it.
00:15:48:03 - 00:16:04:19
Chris: Maybe Google sent them a cease and desist behind the scenes. It doesn’t make sense to shut it down after only a couple of months. That was their only path to reduce the burn rate. If they thought e-commerce was going to be easy, it takes time.
00:16:04:20 - 00:16:28:08
Chris: Anyone who thinks these new players will arrive and immediately build an ecosystem for advertising on scale is mistaken. It took Google years to perfect this.
00:16:28:10 - 00:16:47:18
Chris: Google and Amazon are the two major players in agentic commerce. This leads me to the third layer of noise: everyone who thinks there will be new players on the block is right, but anyone who thinks they will eat significant market share from Google is wrong.
00:17:11:19 - 00:17:40:08
Chris: One year ago, I said Google faced real competition. It turns out the competition isn’t moving the needle much. The dominance of Google and how they embed AI into their advertising ecosystem is second to none. I’m more bullish on Alphabet than ever.
00:17:40:14 - 00:18:03:08
Chris: I want to talk about the new attributes and Google’s strategy. Google announced a while back that they would have thousands of new conversational attributes.
00:18:03:08 - 00:18:28:03
Mike: I can’t go into too much detail because it’s not all public yet, but at a high level, it feels like an admission that AI alone is not up to the task. It emphasizes how important the infrastructure layer really is.
00:18:28:04 - 00:18:55:18
Mike: Google has years of advantage with Merchant Center and product data. They need more structured data. When you look at these attributes, they are closing gaps with the OpenAI and Meta feeds.
00:18:55:20 - 00:19:14:01
Mike: They want the Google feed to be authoritative. You should focus your effort on your Google feed because it will cover other channels. If you’re investing in OpenAI or Meta attributes, Google should be able to ingest and benefit from that data. It furthers their flywheel.
00:20:21:24 - 00:20:51:09
Chris: Given this ecosystem won’t dramatically change, you must do everything in your power to focus on the things you know. That will be decisive for your outcomes. How do we make the data feed future-proof?
00:20:51:11 - 00:21:18:17
Chris: The data feed will be at the center of everything. What is your take on it? Google wants to take things that exist on your website or in your business knowledge and bring them into the feed.
00:21:18:17 - 00:21:47:13
Chris: Is there stuff from your Q&A or reviews that could be surfaced? Google’s product taxonomy is not that detailed, so it’s important to have clear product types and descriptions. Think of every piece of semantic knowledge that exists and how to transport that into your feed.
00:21:47:13 - 00:22:10:06
Mike: By giving that data in a structured way, you’re alleviating some of the heavy lifting Google would have to do otherwise. It’s a super smart move to outsource that.
00:22:10:08 - 00:22:39:00
Mike: This semantic conversational layer will be the biggest key, along with native commerce requirements like precision with total costs, taxes, and return fees.
00:22:39:04 - 00:23:14:22
Mike: In Dublin, I asked if they see a change in search patterns. They said the length of search queries is already 2.5x compared to classic search. People are adopting more natural language. You have to meet that with your data structure.
00:23:15:03 - 00:23:35:23
Chris: I did an interview with Ginny Marvin where she shared statistics on how search is changing. We no longer have to repeat that “15% of searches are new” line; there is a material change happening.
00:23:35:23 - 00:23:57:02
Chris: These attributes provide “grounding” for AI so it doesn’t hallucinate. It helps the AI get details right when people ask specific questions. If the AI doesn’t have the detail, it will just invent a plausible answer.
00:23:57:04 - 00:24:27:20
Chris: This grounding material means the AI doesn’t have to guess. I asked a CTO of a major UK player what the most important interface is for future product information. If you have a siloed approach where a product manager defines information without talking to marketing experts, you will lose.
00:24:27:22 - 00:24:47:09
Mike: The market will tell you how people search now. It remains to be seen how much transparency we’ll get into those conversations. With PMax, they’re working on “synthetic keywords” to take long queries and create something digestible for learning.
00:24:47:09 - 00:25:13:03
Mike: You have to empower your marketing people to have a say in data feed structure. Google is the market. One last thing: I’ve quoted you many times, Mike. You said algorithms think per auction with no view longer than a month, budget-wise.
00:28:05:22 - 00:28:28:13
Chris: That’s a great quote. Anyone who thinks shopping is “set and forget”—to quote Kobe Bryant, “job’s not finished.” It starts with making data actionable. This data is your context window for algorithms to learn on.
00:28:28:18 - 00:28:58:12
Chris: You need to provide signals that algorithms can work with by segmenting products according to your business goals. Use budget and target ROAS settings as controls, not just goals. I’m emotional about this because it’s often too shallow.
00:29:18:05 - 00:29:39:06
Chris: Having the right data structure is just the beginning. You have to think about how to put that into signals so algorithms work as hard as possible for you.
00:29:39:08 - 00:30:15:17
Chris: These systems are like your distributed salesforce, having conversations with potential customers. There’s a wealth of knowledge in reviews and Q&A. AI can do the digging for the customer, which is a beautiful thing.
00:30:33:00 - 00:30:57:09
Chris: Think beyond that to return rates, basket data, and price competitiveness. You have to think about it in a unified way and then make it actionable. This is huge. The CMOs I spoke to really understand this challenge.
00:30:57:09 - 00:31:19:23
Mike: Thanks for sharing those insights from the “CIA think tank.” We discussed global world domination. If anyone is interested in the slide deck I presented, give us a shout out.
00:31:43:06 - 00:32:03:16
Chris: Let’s wrap it up for this week. It was a pleasure, sir. This has been another episode of Growing Ecommerce, brought to you by Smarter Ecommerce, also known as smec. You can learn more at smarter-ecommerce.com.
00:32:03:21 - 00:32:25:18
Mike: If you go to our website navigation, you’ll find a content repository where Chris’s talk and other reports are unified. Go check it out. Thanks again, see you next time.