Transcript:

Tony Myers: Hey everyone, I'm Tony Myers, director of paid search here at Green Line Digital. Today I'm joined by Cristina Priamo, our VP of paid search, and we are talking about something that's changing pretty much everything we know about how people shop online—and that is AI-driven search.

Cristina Priamo: Hey, Tony. I'm really excited about this one because I feel like every week there's something new happening in this space. And for retailers, it's really about understanding how your customers are changing the way they shop.

Tony Myers: Exactly. And I think that's an important distinction. We're not just talking about a few new Google features or another advertising platform. We're talking about a pretty fundamental shift in how people discover products, how they're comparing options, and ultimately making purchase decisions.

Cristina Priamo: Exactly. For years, I feel like we thought about the shopping journey as being relatively straightforward. You know, someone needs a product, they go to Google, they search for it, they click on an ad, and hopefully make a purchase. But that journey is just not linear anymore like that.

Someone might discover a product on TikTok, ask ChatGPT for recommendations, watch a couple YouTube reviews, get an ad on Meta, and then go to Google when they're finally ready to buy. And I think that's something retailers really need to be thinking about. Your customers could be interacting with your brand across six or more touchpoints before they even make a purchase.

Tony Myers: Yep. And the searches themselves are changing, too. We're moving away from those short-tail types of searches, like "espresso machine," and now we're seeing searches like, "what's the best compact espresso machine under $400 for a small kitchen?"

That's a much more specific question, obviously, and it tells us so much more about what the customer actually wants. That's a huge opportunity for brands that know how to take advantage of that. We're getting more context about the shopper's needs, what their budget is, their preferences, and the problems that they're trying to solve for themselves.

Cristina Priamo: Totally. But there's also like a really big disconnect between consumer behavior and retailer readiness. I think I read something the other day where research suggests around 70% of consumers plan to use AI for product research, but only 2% of retailers are fully prepared for that shift.

Shoppers are moving incredibly fast, and a lot of brands are still trying to figure out where to start. I think the biggest thing retailers need to understand is that this isn't just about adding more keywords anymore. If someone is asking AI to recommend an office chair for lower back support, your brand needs to be able to show up as a relevant solution even if the exact phrase isn't sitting in your keyword list.

Tony Myers: Exactly. It's about being part of that conversation while the customer is making their decision. It's not just showing up at the very end when they're ready to make that purchase.

That's really where I see two major opportunities right now, and that is Google's AI Max and then advertising within ChatGPT.

We can start off with AI Max first, because I think there's some confusion around what it actually is. One of those misconceptions is that AI Max is a completely new campaign type or that it's going to replace something like Performance Max, but that's not really how it works. AI Max is a set of AI-powered features within your existing Google Search campaigns—not separate campaigns. It helps Google go beyond your traditional keyword targeting to identify additional relevant searches based on things like the intent of the search, the conversation that was being had, the landing page content, and other signals like that.

Tony Myers: For sure. And I think it's important to distinguish that from Performance Max, because those two serve completely different purposes. PMax is really designed to find conversions across Google's different advertising channels—like Search, YouTube, Display, Discover, Maps, Gmail. AI Max is really focusing on how your search campaigns match to queries, how that can customize messaging, and direct users to relevant landing pages.

Tony Myers: Right. And so for retailers, it's not necessarily an either/or decision. You might have Performance Max campaigns working right alongside search campaigns with AI Max enabled.

Cristina Priamo: Absolutely. And I think the opportunity with AI Max really comes down to three things: incremental reach, more relevant messaging, and better alignment with how people are searching today.

If you think about how a retailer sells hiking boots, they might be bidding on keywords like "women's hiking boots" or "waterproof hiking boots." But someone could just search, "what shoes should I wear hiking in Colorado in October?" That shopper is a really great potential customer, even though their search looks nothing like your traditional product keyword.

That's where AI Max can help bridge that gap. It's looking beyond the literal words in the search to better understand what the person is trying to accomplish. It does that through a combination of three main capabilities: search term matching, text customization, and finally URL expansion.

Tony Myers: Let's talk about those, because I think the way they work together is what makes AI Max interesting.

Search term matching is really about expanding the types of relevant queries your ads can show for. The text customization can then adapt your ad messaging to better align with what that person is looking for. And then the final URL expansion part of this actually sends the user to the relevant page on your website, rather than just using that one landing page you originally selected.

Cristina Priamo: Yeah, and that can be incredibly valuable for an e-commerce brand with hundreds or thousands of products. If someone's specifically looking for waterproof hiking boots, we don't necessarily want to send them to a generic footwear category page. We want to get them as close as possible to products that meet their needs.

Tony Myers: But I think we should also talk about the other side of that flexibility, because it's not all an upside.

Cristina Priamo: Oh, absolutely. More automation does not mean less management.

Tony Myers: That's a big one. Just because Google can expand your reach doesn't mean every search it finds is going to be a good fit. You'll need to be reviewing search terms, monitoring performance, and making sure you're not expanding into searches that don't make sense for your business.

That's especially important for retailers with very specific price points or product categories. Let's say you sell premium furniture. You probably don't want Google aggressively expanding your searches into searches for cheap or budget furniture if that's not what you offer. And the same goes for landing pages. We want Google finding the relevant product pages, not sending people off to the shipping policy or customer service-related pages.

Cristina Priamo: Right. And that's why things like negative keywords, URL exclusions, and ongoing search term reviews are so important. The technology is really powerful, but you still need guardrails.

So, let's talk about when it actually makes sense to test AI Max, because I don't think every retailer should just go into their account tomorrow and turn this on, do you?

Tony Myers: No, I definitely agree there. The first thing I'd look at is whether the existing campaigns have enough conversion data and budget to even support a meaningful test. If your search campaigns are already severely budget-constrained, expanding your reach might not be the first thing you really want to do. You may end up spreading an already-limited budget across more searches.

Cristina Priamo: Exactly. I'd also look at where there are opportunities for incremental growth. Are there product categories you're not capturing enough traffic for? Are your competitors showing up on searches where you're missing? I think that's where AI Max can really be interesting.

Tony Myers: And the conversion volume matters, too. As a general rule, I'd start with campaigns generating about 50 or so purchases a month, and that gives us a stronger foundation for testing and evaluating the performance.

Cristina Priamo: Yeah, that makes sense. Another big thing I would really want to stress is patience. (LAUGHTER) I know everyone wants to turn something on Monday and have results by Friday, but that's just not realistic with these types of automated solutions.

Tony Myers: (LAUGHTER) No, especially when leadership is asking for an update three days after launch.

Cristina Priamo: Exactly. We generally recommend giving an AI Max test at least 30 days before making major decisions, and accounting for that initial learning period when you're evaluating the results.

A little bit of a different mindset shift here is that we shouldn't be looking at whether clicks increased. Did we drive more purchases? Did CPA or ROAS improve? And most importantly, did we generate incremental business that we weren't capturing before? That's really the test, because if all we've done is shift conversions from one campaign to another, we really haven't done anything for the business.

Tony Myers: Exactly. Okay, so let's shift gears over to ChatGPT now, because I think this is where things can get really interesting for retailers. The biggest difference, in my opinion, is where we're reaching shoppers in their journey.

Cristina Priamo: I completely agree. With traditional Google Search, we're often focusing on capturing existing demand. Someone already knows what they want, and we're trying to get them to choose our brand. With ChatGPT, we have the opportunity to influence that decision much earlier.

Someone might not even know what product they need yet. They could be asking ChatGPT something like, "I need a gift for my wife for our anniversary, and I want to spend under $300." They're not searching for a specific brand or product; they're asking for help figuring out what even to buy.

Tony Myers: And that's an important distinction, because if you're a jewelry retailer, for example, you want to be part of that discovery process. You want shoppers to consider your brand while they're exploring their options, not just hope that they eventually search for your name on Google.

Cristina Priamo: Yeah, and from an advertising perspective, that means we have to approach campaign strategy differently. We're not building out traditional keyword lists like we would in Google Ads. Instead, we're thinking about the types of conversations where our products would be relevant.

We'll use contextual targeting guidance to help identify those conversations. So rather than targeting something like "gold-plated necklace," we might focus on conversations about gift ideas, comparing jewelry brands, or finding accessories for a special occasion.

Tony Myers: Yeah, and I think the creative strategy needs to change, too. We're so used to writing Google Search ads where we're trying to stand out against other advertisers on a results page. But in ChatGPT, the user is having a conversation.

The ad needs to feel relevant to what they're asking in that particular moment. The copy needs to be concise, useful, and directly connected to that shopper's needs.

If someone's comparing jewelry brands, we shouldn't just show them a generic, "Hey, shop our collection" page. We should be giving them a reason to consider our brand based on what they're actually researching.

Cristina Priamo: Yep. And another big piece for e-commerce brands is product data. A clean, detailed product feed becomes incredibly important as AI-driven shopping experiences evolve. But that's not unique to ChatGPT. Across AI-driven shopping experiences, product titles, descriptions, attributes, pricing, and availability all help platforms understand what you're selling. If your product data isn't complete or it's not accurate, you're making it harder for these systems to even match your products with your shoppers.

Tony Myers: Yeah, and that's why feed optimization is no longer just something you do for Google Shopping. It's becoming a much bigger part of making sure your products can be discovered across different AI-powered experiences.

Cristina Priamo: Yep. Now, one question I think we're going to hear a lot from retailers is whether they should start shifting budget out of Google Search to test ChatGPT. And my answer here is generally no. (LAUGHTER)

Tony Myers: Yeah, definitely the same here. If your Google Search campaigns are driving profitable conversions, you don't want to pull money away from something that's working. ChatGPT should really be viewed as an opportunity to expand your reach and expand your influence in the consideration period, but it shouldn't be looked at as replacing your existing lower-funnel advertising.

Cristina Priamo: Exactly. And that's especially important as we head into Black Friday and Cyber Monday. There's a real opportunity to get your brand in front of shoppers earlier while they're researching gifts, comparing products, and planning purchases.

Tony Myers: Yeah, and that brings up the measurement piece, too, because I think that is where marketers need to set their expectations internally. If you're trying to compare ChatGPT directly against a branded Google Search campaign on last-click ROAS, for example, you're not making a fair comparison.

Cristina Priamo: Exactly. They're doing completely different jobs. A branded search campaign is capturing someone who's already decided that they want your product, and ChatGPT may be introducing someone to your brand for the first time. So, we need to look beyond immediate conversions.

I'd want to understand what those visitors are doing after they click. Are they engaging with the site? Are they viewing multiple products? Are they coming back later through other channels? And where we have the right measurement in place, can we see evidence that ChatGPT is contributing to future conversions?

I think that's a really important part of the conversation to have with leadership before launching a test. You have to define what success looks like upfront. If the goal is building awareness and consideration, don't evaluate the campaign solely on immediate ROAS.

Tony Myers: 100%. Otherwise, we're just risking shutting down something that could be creating valuable demand simply because you're measuring it against the wrong objective.

Tony Myers: So, we've covered a lot. The search landscape is changing quite a bit. Google's introducing more AI-powered capabilities, and ChatGPT is creating new ways for shoppers to discover brands. But I think the most important question is: what should retailers actually be doing right now?

The good news is that you don't have to completely reinvent your marketing strategy overnight. A lot of the prep comes down to getting the fundamentals right, and I'd start with measurement and first-party data.

Cristina Priamo: Definitely. AI is only as effective as the signals we're giving it. You want to make sure conversion tracking is accurate, that you're optimizing towards meaningful business outcomes, and that you're using first-party data where it makes sense. If you're telling Google that every website visit is equally valuable, you can't expect it to automatically understand which customers are more likely to drive revenue for you.

The second thing I'd focus on is campaign and product feed structure. We've talked a lot about automation, but automation needs good inputs. If you've got dozens of campaigns with very little conversion data or product feeds with missing attributes, those are things I'd be addressing now.

I'd also add website content as the third priority here. This is something I think retailers sometimes overlook. Your product pages need to clearly explain what you're selling, who it's for, and what makes it different. Think about those conversational searches we discussed earlier. If someone is asking for a compact espresso machine under $400, your website should make it easy for search engines and AI systems to understand the machine's dimensions, prices, features, and use cases. Clear product descriptions, specifications—all of that is really good information, and it needs to be accessible.

Finally, the fourth really is testing.

Tony Myers: Absolutely. Don't wait until everyone else has this figured out. Set aside a reasonable test budget, identify where AI Max and ChatGPT ads could complement your existing strategy, and establish those clear goals before launching.

It's very important to give those tests enough time to actually learn something, too. I think one of the biggest mistakes marketers can make right now is just jumping into every new AI feature without a strategy, or waiting so long that they fall behind their competitors. You don't need to be everywhere, but you do need to start understanding where these opportunities fit into your business.

Cristina Priamo: Yep. And I think if there's one thing I want retailers to take away from this conversation, it's that AI isn't just changing how we advertise—it's changing how people shop. The brands that are going to actually be successful are the ones that adapt to how their customers want to discover and evaluate products.

Tony Myers: Yeah, and at the end of the day, the fundamentals of good marketing haven't changed. We need to understand your customer, make sure your products are relevant to their needs, and give them a great experience. AI is just changing how we connect those dots now.

Cristina Priamo: Exactly. And that's what makes this such an exciting time to be in search.

Tony Myers: All right. Well, I think that's a great place to wrap it up. Thanks for joining me, Cristina, and thanks to everyone listening, and we'll see you next time.