Ep 623: Your Google Product Feed Is the Most Overlooked Lever in AI Shopping (with Pilothouse)

0:00Before we jump into today's all killer, no filler episode, a quick word about who makes this show possible. The D2C podcast is brought to you by Pilothouse, the performance agency behind some of the fastest growing D2C brands in the world. Creative, media, and customer journey all under one roof. Performance and brand without the trade-off. Every Friday, we hand the mic to a Pilothouse operator to break down what's actually working in their space right now. Want a team that treats your growth like their own? That's Pilothouse. Head to pilothouse.co and now on with the show.

0:30Doug:You definitely have reduced ability to serve for search demand higher in the funnel, right? The more conversational queries, the more problem-but-not-solution-aware queries are largely going to be handled by the AI overview. But the transactional queries are still very much in play. Why is the future keywordless? I'm going to my search bar. I'm not just typing three keywords. We are able to complete paragraph-long searches. You're not going to build millions of keywords around paragraph length terms. You're now going to try and direct towards interests, intent. So you're not just targeting based on audience.

1:16Hey guys, before we get started today, I just want to let you know this podcast is all about your Google feed and how to optimize it for discovery in the Google shopping environment. I want to let you know that this comes standard in all the audits that Pilot House does for clients looking for Google services. So if you are interested in having your feed audited to make sure that it's fully up to snuff, reach out at pilothouse.co on the forum and tell them the D2C podcast sent you. It's all killer, no filler. I'm Eric. Welcome to the D2C podcast. Today we are talking with Mr. Dougie, who leads the Google team over at Pilot House.

1:52Welcome to All Killer, No Filler. Doug, how are you doing? I'm doing well.

1:56Doug:Thanks for having me, as per usual. Yeah. I want to start with a question off the top. I'm seeing a lot of discourse right now about Meta and the challenges people are having on Meta. It's one of the toughest times, a lot of folks are saying, to be advertising on Meta right now, unless you're a really big, established brand. There's a lot of functions that aren't working correctly. The conversion rates are low. Is this something we're seeing in the Google world as well? Yeah, maybe not to the same extent, but to some degree, I think we're just on the more macro level and like a transitional moment.

2:29Doug:I think DTC advertisers continue to undervalue the impact of true awareness. And I think that true awareness is very difficult to attribute performance to. So I think with all the changes regarding Andromeda on the creative side for Meta, that's been a transition for a lot of folks. So maybe not for the same reason, but there are also challenges during this particular time on Google as we navigate what AI discoverability means for paid placements on Google and what have you. So I would say maybe not for the same reasons, but to some degree, seeing a bit of that on Google, it doesn't feel like the world is or the sky is falling to any degree.

3:08Doug:But yeah, I think it just feels like a very transitional moment in DTC e-com. So what are the – like, I don't know. You don't know if you know the stats on hand, but it's like the amount of search volume that's been taken by the AI summaries at the top of Google. Are there – is there any impacts that you're seeing on volume when it comes to – because I guess they're showing a lot of ads right under that still where they normally would have been. But I imagine that AI answer solves a lot. What are you seeing in terms of impact of that change? Yeah, so you definitely have reduced ability to serve for search demand higher in the funnel, right?

3:44Doug:The more conversational queries, the more problem but not solution aware queries are largely going to be handled by the AI overview, right? But the transactional queries are still very much in play. And I think in DTC e-commerce, we haven't strayed too far from transactional queries, right? Like we're selling blue shoes, someone searching specifically for$100 blue shoes. That's a pretty transactional query that we're still able to show up for and convert well for. But if you're asking maybe what are the best shoes for wide feet, that's maybe a query that you could have targeted more specifically in the past that you've had to move away from in the present.

4:22Doug:So it certainly has shifted the behavior, but for lower funnel e-com, which is typically what Google Demand Capture services, it hasn't been quite as impactful. But for the blog sites and what have you that are showing you the top 10 of this, that, and the other, I think those have been hit the hardest from the change to AI overviews. That makes sense. And then what do you see in your results when your brand gets featured in one of those AI summarized queries? Do you see that the purchase comes from Gemini? Yeah, so there is like standard UTM tracking. So even for our own business, I can't speak too specifically to it because we have another team that manages that.

4:59Doug:But we are able to track, for instance, what B2B leads are coming through from ChatGPT. So depending on where the query is surfaced and on what LLM, you can still use UTMs to attribute at least a portion of the traffic to these non-Google search behaviors. Interesting. Okay, cool. On to the meat of the topic, which is Google Marketing Live 2026, the What's Next Google Ads Conference. What were some of the exciting things that came out of that event? Yeah, there's a few things that I was pretty keen on. I would say the two highlighted areas for me as a DTC e-com advertiser is the direction they're going with regards to AI and demand capture, and then the direction they're going to try and bridge some of the gaps that are existing when it comes to a couple of things we talked about already in that awareness marketing and awareness attribution is pretty difficult, especially on Google slash YouTube.

6:01Doug:So those are the two primary areas of focus that I keyed in on during the presentation. And I think I am most excited about there's some lead gen stuff and what have you, but just less our realm. So I'm pretty excited to see what particular features from the AI side on demand capture and the measurement side on demand generation come next because I think what they've teased is pretty interesting. Okay, well, tease away what does the DTC audience need to get excited about on the AI demand capture side. And just for any newbies listening, we are kind of talking about what we were just talking about, which is how to show up in the AI summaries.

6:39Is that about right? Or high level, what does AI demand capture in this context mean?

6:43Doug:Yeah, so it is that for sure. Google is going to monetize Gemini and their AI overviews. They're doing it to some extent right now. It's not to, I think, the full extent that we anticipate seeing in the future. They are moving into a direction similar to ChatGPT in terms of monetizing their platform for advertisers. But what that means in terms of actual features for Google, right now, there's something called AI Max, which to some extent allows you to push search campaigns more towards those conversational queries that are largely surfaced via Gemini or through AI overviews, they're also rolling this out for shopping campaigns now.

7:24Doug:So that's particularly interesting because you don't just have a bit of text reading out about a particular product. You have the price, the product ratings, an image, and what have you. And this is typically where most DTC advertisers are getting the majority of their conversions anyways. So it's interesting to see this component or this type of placement for Google being filtered in to AI forward placements. So what does that mean? It means you're showing more product focused ads during kind of agentic shopping slash conversational queries. And I think what advertisers are going to have to increasingly focus on, particularly on Google, is the quality of their feed.

8:08Doug:We've maybe touched on it in the past, but I think the product feed that you filter through to Merchant Center and then through to Google Ads is probably one of the most overlooked undervalued components that you can manipulate to impact performance. Because it's not like you click a button and you spend more on these adjustments. You are trying to provide as much detail and value to the AI that's crawling your listing and your site and your feed with all the information you're providing. And it is then digesting all that information and determining which queries are most relevant for your particular product.

8:44Doug:So in the same way that people are developing post-click to try and be discoverable for LLMs, I think there will be a similar phenomenon for making sure that your feed is as robust as possible to ensure that when the Google algorithm is trying to determine, hey, should we serve this particular shopping product for this particular conversational query, you are being surfaced because your feed provides so much information and relevant content for particular searchers. And is that just a matter of sort of metadata diligence and just making sure you're maximizing the value of each of the available fields?

9:22Or are there other dimensions you need to think about?

9:24Doug:There's certainly a large component of not metadata manipulation, but making sure that you are providing as much relevant information for each of the relevant fields, I would say that there's, in many clients and many advertisers, they are just kind of copy pasting whatever's on site into their feed, or they're using an automated integration with their site. And typically, that is suboptimal, because you are not fully fleshing out each of those attributes. So you might have a really strong back end in Shopify, but if you're not mapping that correctly into the feed, you're not even filtering that through for Google to digest and take into account when serving to more conversational queries.

10:08Doug:So yeah, I think in short, we want to maximize the current attributes within the feed, but you also want to make sure that you're setting yourself up via either a third party management tool or through an automated integration such that whatever work you do on the backend towards LLM discoverability is also being accurately mapped to the different attributes you're utilizing in the feed. Are there any tools that you'd shout out? Right now, I mean, there's 101 standard tools you can use for feed management. I would say I haven't come across one that is particularly well suited for the kind of shifting landscape for AI discoverability yet.

10:49Doug:I would say most of your feed management tools are going to be very similar to one another, just with different levels of interface and customer support. And is just like doing an audit of your feed, all the data in your feed, that's like a perfect job for Claude or Gemini or any of these tools? Is this something you really have to really think about from a human perspective? Or should an LLM be able to provide you the perfect feed data? Yeah, I think Claude could probably get you most of the way there. And we've utilized AI to analyze feed data and what have you. But it's also understanding best practices for each of the individual attributes, how that influences what shows up for consumers, how that influences what attributes Google takes into account.

11:35Doug:I would say if you are an advertiser or a brand who maybe doesn't have as robust of an ability to leverage Cloud or use additional AI tools internally, depending on your size, you are also able to rely on Google reps to provide some of that analysis as well. So Google reps are also there to provide some degree of a health report on your feed. And you get some of it naturally in your merchant center as well. But to get deeper, you can probably get 80 % of the way there with Cloud, but there's going to be that additional probably human layer to ensure that you are taking into account how the AI is digesting that information.

12:16Question, why is the future keywordless?

12:19Doug:Well, it's been moving that direction for a while, right? Like broad match terms are more semantic, right? We lost modified broad match many years ago. Exact match isn't truly exact match. So we've been moving in this direction. Google's been training us to understand that, hey, we're not going to be as prescriptive in our targeting moving forward. But the future is keywordless because search behavior has already changed significantly with the utilization of AI, right? Like you are no longer searching blue shoes to use a previous example, or like you're only searching blue shoes. If you want to see, Hey, here are all the blue shoe products that maybe I can purchase right now.

12:56Doug:More and more people each day are searching and discovering that, okay, via this crazy tool in AI and Gemini or Chatshubit, whatever you want to use, you can get more customized. Wait, are you actually recording now? Just wait a second. Are you recording? Okay. Now you're recording. Okay. I want blue shoes that are for squash. and they need to have really good grip and they need to, they're just yapping, you know, just yapping is the new keyword. Exactly. And I found it myself, right? I'm going to my search bar or LLM of choice and I'm not just typing free keywords because that wouldn't have given you a relevant result in years past.

13:30Doug:But now given the evolution of technology, we are able to complete paragraph long searches. So with that in mind, you're not going to build millions of keywords around paragraph length terms. You're now going to try and direct towards certain interests, intent within the terms. So you're not just targeting based on audience. But yeah, I think the technology has changed the way people search and therefore we have to adapt and adjust how we are targeting. And Google is adjusting what tools they're providing us to target people given this change in behavior. Very cool. What else was gleaned at this What's Next for Google Ads conference?

14:15Doug:A couple more interesting things, or maybe one more on the AI front. They are working on bypassing the landing page experience, which I think for some brands that have poor technical debt on the PostClick side might be helpful. I do see it being most effective in the lower funnel, right? An example where you show a YouTube ad to a completely cold audience, you probably don't want them, unless it's a true impulse product, clicking and resulting in a checkout right then and there. They probably have to do some more research and understanding the product before they go full conversion. but google is trying to play around with expediting some of the checkout process and i think what we'll see in the midterm future is more and more of this shoppable within the platform and i think this is not going to be limited to google we're already seeing it to some degree in other platforms right but i think we will increasingly see the ability to shop and check out within the platform itself as opposed to being pushed through a PDP and then you have to go add to cart and then you go to your cart and then you go to your checkout and take action from there.

15:31Doug:I think we are going to increasingly see that transition squeezed into the platforms themselves. You don't put speed bumps on a runway is what they say, right? When you don't need to, you just get right to the purchase. Yep, for sure. So they're moving in that direction, which is not surprising given where we're seeing other platforms move as well. But that was another interesting note on the AI side of things. But some of the most exciting stuff was probably on the measurement of demand generation, just because that's where I feel Google has had the greatest gap. That's where our contacts at Google have told us, hey, this is our biggest problem.

16:09Doug:This is where brands and advertisers are coming to us. And we don't have a good answer for them because we haven't built anything in the backend to meaningfully support this for self-serve advertisers. And for years, so for anyone listening new, Google's considered a demand capture platform more where people are problem aware and they're looking, searching for things, but the ability to actually generate demand, which is what can be done on social advertising. But they've always said, Google's always said that YouTube is their demand generation platform in a lot of ways. So what does this mean?

16:46So measurement is key, I guess, right? So if you've got YouTube as this other major piece, measurement of how it works would be super important. Yeah.

16:55Doug:And they've been pushing YouTube and I've been on board with YouTube being a meaningful channel for demand generation, but it's been so hard to tie the thread between that initial upper funnel impression and conversions in your backend. Part of the reason it's so difficult on Google is because even within their optimization for particular conversions, if you have a click on a YouTube ad that then later on a few days down the road, Google's your brand and converts via brand search, that conversion is being divvied up between those two signals. And as much a brand search click is a much clearer signal and a much more trackable signal than a YouTube view as an example right so that's been the difficulty is that google ads in totality has this interesting phenomenon more so than other platforms where you truly have the lowest of low funnel and people are actively searching for your brand ready to convert to the upper uppermost of upper funnel where people have never heard of your brand uh and are consuming creative content akin to tv so uh that is what they're trying to bridge the gap on right now.

18:08Doug:So a few things that they've come out with, one being demand gen campaign AB uplift experimentation. So in theory, this would allow us to be able to measure the impact on lower funnel conversions more accurately. They have some of this in platform, but honestly, it's kind of, they feel like band-aids slapped on the solution over and over again, and none of them really provide that much value. In theory, this would allow us to be able to truly measure the impact of demand gen campaigns on our down funnel work. They are also rolling out options to change how your attribution is allocated that is more similar to how social platforms operate, as well as including view through attribution to the actual optimization of the campaign, which you of course want to be careful about including view through, But assuming you have strong exclusions in place in your upper funnel targeting, that could be quite meaningful, especially given the fact that YouTube is primarily a view-based platform as opposed to socials, which are typically more click-based.

19:14Yeah, and people don't often – like I'm on social media quite often to go somewhere else to shop or do whatever, but a lot of people on YouTube are there for their primary entertainment experience. 100%.

19:25Doug:And I've been parroting the idea that YouTube is the new TV. And that's evidenced by the amount of adoption across North America. It's evidenced by behavioral patterns in terms of what people are doing after work. They're not going and they're flicking on the cable. No, they're going and they're watching their subscribers and what came out most recently on YouTube. I mean, there's YouTube TV now, which is just actually a TV, a true TV subscription through the YouTube platform. So the more it kind of takes that direction, I think the greater the difficulty because most of that consumption is view-based as opposed to click-based.

20:04Doug:I don't know about you. I've personally never clicked on a YouTube ad in my entire life. Just to close it. Just to close it. I've certainly had meaningful impressions from advertisers via YouTube, though that impacted my awareness, consideration, and conversion intent for a brand. Every time I hear view through, because I go way back in display advertising and they were always trying to get me to sort of like factor in view through in my hardcore performance marketing affiliate campaigns. And I was always so wary of it. So I think being able to, I guess, de-dupe it as well, right? Like make sure you're triangulating with a measurement partner essentially to make sure that you're not – like I guess they're not trying to ding you by taking into account view through attribution here, I guess, right?

20:45They're just trying to show the value of views a little more.

20:48Doug:Yeah, they're trying to provide the campaign with the signal that, hey, this was meaningful. And there's a difference between setting up your campaign such that you're optimizing to a 90 day view through window on your retargeting list like that is not useful, right? But if you are using a one-day view-through on a true cold audience and you're finding that, okay, I'm getting view-through conversions from people that I know have never interacted with my brand to this point, that's a meaningful signal. So there's the huge barometer that you have to control for as the advertiser. But I think that because true top-of-funnel YouTube fits so high up, it is certainly worth accounting for and optimizing for.

21:32Doug:Just be careful around attribution windows and how you abuse it. And then the last one here is qualified future conversions. And that just makes me think about pre-crime and minority reports. So they're basically counting their chickens before they've hatched. Yeah, I'm interested to see how this one unfolds. What they showed was pretty interesting on it in terms of the visuals. where they're already doing a bit of this work is there's betas going around where you're able to see the number of brand searches that your ads have have driven on youtube and demand campaigns so i find that super interesting because we've had to triangulate it via search console and google trends and smrush and what have you in the past when we've run more upper funnel efforts but now having the ability to track that in platform assuming it's accurate is quite interesting to me.

22:24Doug:So that's just an example of them being able to forecast, okay, we've driven these brand searches. Typically, brand searches convert at XYZ rate over this window of time. So we can anticipate this resulting in XYZ factor. A qualified future conversion. That's what they're calling it. Yeah. So I think there is legs to that. I don't think it's all kind of smoke and mirrors, trying to grab attribution because it's a philosophy that we've adopted to try and measure upper funnel awareness to this point already. Cool. Well, Dougie, thanks for coming on the D2C podcast to talk about these exciting AI developments.

23:04You mentioned feed. Is it something that the pilot house does? Like if someone's out there and like my feed's a mess, they want to work with you directly to fix their feed. Is that something you do? Could you do feed audits?

23:14Doug:Yeah, we absolutely can audit the feed. It's a key component in the Google ad audits that we perform. We aren't at this current point auditing for LM discoverability on your post-click, but we are auditing your feeds for how your products show up in shopping listings. So absolutely a major part of what we do, I think an undervalued part for many advertisers and brands out there that I think has been overlooked by some. And I think it will only become increasingly important as we see more AI development and AI placements utilizing information from that feed. Love it. And what do you think about the Taylor Radish signing?