Why Your Brand Is Invisible To AI (And How To Fix It) — Cole Casperson | Why AI Search Changes Ecommerce, Why AI Boosts Your Conversions, What Intent-based Discovery Means, Why Traditional SEO Is Not Enough, How AI Improves Traffic Quality (#468)

In this episode, we explore how AI is fundamentally changing the way people find and buy products online. 

Cole Casperson, Partner and Chief Data Officer at Crank Tank, discusses the shift from three-word searches to long, intent-driven AI conversations. He explains how brands can move beyond traditional SEO to ensure they are the ones recommended by AI tools. 

You will learn how to use diagnostic tools to see what AI thinks of your brand and how to win against bigger competitors in this new era of discovery.

Topics discussed in this episode:  

• How AI shifts search from keywords to intent.
• Why user queries are getting much longer.
• What makes AI discovery different from SEO.
• How to use diagnostics to see AI rankings.
• Why competing on ideas beats keyword stuffing.
• What Amazon can teach brands about AI commerce.
• Why ignoring AI creates a compounding risk.
• How semantic understanding boosts conversion rates.
• What the Universal Commerce Protocol means.
• How challenger brands can outrank giants.

Links & Resources 

Website: https://cranktank.net/
LinkedIn: https://www.linkedin.com/company/cranktank/
X/Twitter: https://x.com/WeAreCrankTank
Instagram: https://www.instagram.com/wearecranktank/

Get access to more free resources by visiting the show notes at https://tinyurl.com/ccpvanea

______________________________________________________

LOVE THE SHOW? HERE ARE THE NEXT STEPS!

Follow the podcast to get every bonus episode. Tap follow now and don’t miss out!   

Rate & Review: Help others discover the show by rating the show on Apple Podcasts at https://tinyurl.com/ecb-apple-podcasts   

Join our Free Newsletter: https://newsletter.ecommercecoffeebreak.com/   

Support The Show On Patreon: https://www.patreon.com/EcommerceCoffeeBreak   

Partner with us: https://ecommercecoffeebreak.com/partner-with-us/

Transcript

00;00;00;09 - 00;00;25;06
Unknown
AI systems from ChatGpt all the way down to rule based bots do is something different. They retrieve the content from across the web. They evaluate it and they generate a recommendation. And the process through which they do that is not a black box. What's the real cost of ignoring a Oh yeah.

00;00;25;08 - 00;00;48;07
Unknown
Hello and welcome to another episode of the eCommerce Coffee Break. A podcast search is entering a new era. For years, eCommerce brands fought for rankings on Google. You're optimized for keywords, you bought traffic, you tried to win the click, but the game is shifting and changing a lot. Consumers are typing longer questions. They're going straight to iTunes and expecting one clear answer, not ten blue links, one answer.

00;00;48;10 - 00;01;12;01
Unknown
So the big question is, is your brand even in that answer to unpack this, I'm joined by Cole Casperson. He is the partner and chief data officer at Crank Tank. He's a former ski industry executive, a brand owner himself a licensed CPA, was a data science background from M.I.T. and the creator of Reach, a diagnostic tool built to help brands see how I actually retreats and brings that content before customers ever see it.

00;01;12;04 - 00;01;39;01
Unknown
Welcome to the show. Thanks for having me, Claus. Really appreciate it. More and more people stop googling and start asking what's really going on behind the scenes. Yeah. So for years, customers have been used to interacting with the Internet based on these search engines, and they've been doing that. Those queries have averaged three words a query. But now with AI, where they can search based on meaning and intent, they're very much enjoying it.

00;01;39;01 - 00;02;00;20
Unknown
It's not a hate driven cycle. Customers really enjoy being able to search on multiple meanings and keywords and intent, and so now their queries are indeed, you're correct, they're 16 words long on average. Now at voice search queries are actually 29 words according to the latest dataset I saw. Yeah, that's impressive. And I think everyone is using it by now.

00;02;00;26 - 00;02;21;17
Unknown
So but obviously we want to find out what does that mean for e-commerce and DTC brands? What's your insight there? Yeah, so what I see is that everyone is very aware of this shift and what A.I. means for the shift, but a lot of brands are really approaching it the same way they approached SEO, and that's very understandable because that's the world we've lived in.

00;02;21;19 - 00;02;45;19
Unknown
But A.I. represents a fundamental shift in how that actually works and what search actually is. And so with Google, the algorithm was proprietary and hidden. It was opaque. You could see your rankings go up or down. You could track your traffic, but she could really never see inside the machine. To understand how you related to the question that was being asked.

00;02;45;21 - 00;03;08;01
Unknown
So the entire industry and what we're used to doing is e-commerce people is working from the outside in. You know, you track the symptoms, you try things and kind of hope for the best and that was fine. And there was a lot of really cool ways of doing that. But with air systems from Chad, GP all the way down to rural based bots do is something different.

00;03;08;04 - 00;03;40;06
Unknown
They retrieve the content from across the web, they evaluate it and they generate a recommendation. And the process through which they do that is not a black box. And what we found is it's something we can actually observe, replicate and test. So what REACH is, is for the first time in the Internet era, we can read real diagnostics, not just, you know, your traffic went down or something like that, but actually why I recommended your competitors product instead of yours for a specific query.

00;03;40;08 - 00;03;57;27
Unknown
And hopefully by the end of our conversation today, you'll have a clear picture of what that means for brands. Mm hmm. Very interesting. I want to dive a little bit more what you said, traditional SEO. And I think that's where a lot of people out there, marketers are still stuck there optimizing for keeper keywords. They're right their content in a certain way.

00;03;57;29 - 00;04;20;17
Unknown
And is that wasted time or what's what's the best way to show up in a drive and such? Yeah. So definitely, I mean, SEO is still important because the first step of AI retrieval, what we call retrieval, is and becoming a candidate, still relies on web search. So basically are you want to be top 200 results that could have shown for this.

00;04;20;17 - 00;04;45;13
Unknown
And and then what's really different about large language models or AI is that they don't match keywords or look for co-occurrence. They actually it's really interesting. They understand meaning and they can and they can do this because they convert language into math and mathematical representations of concepts. So similar ideas exist right next to each other in the systems understanding of the world.

00;04;45;15 - 00;05;11;03
Unknown
And so just first grounded example here in the old world we use a jogging shoe example women's trail running shoe and ladies, outdoor jogging sneaker. They don't have any words in common. And because of that, you know, you might have had to optimize for each one separately it on different phrase matches, maybe even different ad groups. But to an eye system, those represent the same idea.

00;05;11;06 - 00;05;32;06
Unknown
And so I understands that the idea of what's going on here is that a person is looking for a rugged, lightweight shoe designed for women who run on trails. And so we're not competing on keywords, we're competing on ideas. And that gives brands a lot of opportunity to think about their content. It's not stuffing the right keywords into the right fields anymore.

00;05;32;08 - 00;05;53;20
Unknown
It's how clearly and completely you're communicating. The idea of your product, what it does, who it's for, and what problem it was designed to solve. And so if you can understand and measure how well a systems understands that you're going to have a tremendous advantage right now and and tomorrow and even more the day after that. From your background.

00;05;53;20 - 00;06;17;13
Unknown
I like that you I was very deep into e-commerce and also have this really high front insight into I know what a lot of online sellers merchants brands have is a product detail page that's sort of limited and most more likely than not based on features. What's the strategy there to get you really your brand be seen in air search?

00;06;17;16 - 00;06;48;26
Unknown
Yeah. One underlying tension that currently exists in the marketplace is, is that web indexing, right? How you become the candidate. Web indexing has always been more inclined towards home pages and collection pages, but the way a classical collection page is designed is it's not really content heavy. It's more more of a grid of projects. Right. And and so there's some very good ways that you can, you know and how you might talk about your product to your category that differ based on whether it's a collection page or or a home page.

00;06;48;28 - 00;07;20;24
Unknown
And specific to what you're talking about with product display pages, that becomes very powerful. And if you think about who has been on the frontier of ecommerce, that's that's Amazon, because Amazon is nothing but product display pages, right? It's the there's no home page, there's no collection page. It's all product display pages. And Amazon, furthermore, doesn't have to worry about what happens on the larger web where people are searching for, you know, what is the local orthodontist or how do I get from here to here on Amazon?

00;07;20;24 - 00;07;43;29
Unknown
You're only there to look at physical e-commerce products. And so we took a lot of the inspiration in designing reach from the actual patented processes that Amazon used to drive. Rufus. And so a lot of people that, you know, a lot of the people listening to this that are familiar with Amazon know that they have a much smaller playing field with the product display page.

00;07;43;29 - 00;08;07;17
Unknown
It's descriptions, it's bullet points, and it's how effectively you can get that message across in a limited space you have what's the real cost of ignoring? I one of the big problems that when you kind of ignore and it'll kind of rhyme with a lot of what we saw you know 20 years ago with Echo is that the risk of waiting is you fault is isn't that you fall a little bit behind.

00;08;07;17 - 00;08;27;16
Unknown
It's that the advantage compounds against you in a way that gets harder and harder to reverse. And this is because air systems have reinforcement patterns. So once they learn to trust inside a source, that source tends to keep getting cited and studies have proven that it is very much a winner take most dynamic and it is a self-reinforcing dynamic.

00;08;27;18 - 00;08;47;19
Unknown
So the brands that are visible now are building the authority that makes them more visible tomorrow, which makes them more visible next month, etc.. And and again, we saw that with their CEO, right. Is is that the brands that jumped on SEO in the early days built domain authority and backlinks profiles and their competitors spent a lot of time trying to catch up to that if they ever did.

00;08;47;22 - 00;09;08;05
Unknown
But at least with traditional SEO, you could kind of brute force your way back into the game, right, with enough budget and enough time. But with air discovery, that reinforcement loop runs a lot faster. And so every month you way the early movers are getting cited more building more authority and the air systems are learning to trust them more.

00;09;08;08 - 00;09;31;28
Unknown
But the good news is this is all still very, very early. The playbook isn't written yet. And so a Challenger brand that moves right now has a genuine shot that they're putting themselves as a authoritative source in the category before the bigger players. And that's especially true if you feel like you've built a better product and are excited to explain exactly why and how that product is better.

00;09;32;01 - 00;09;53;04
Unknown
So that's the good news for our listeners. Still, early games, you can actually survive and get into the game very quickly if you start now. Now you decided to build a system around this, a tool around this which is called Reach. Tell me more about this. What what is it? What does it do? So Reach is a diagnostic platform and it's really built specifically for e-commerce.

00;09;53;06 - 00;10;17;11
Unknown
And what it does is something that sounds simple, but nobody else is really doing it. And what it does is it runs the entire A.I. Discovery pipeline, personalized to the question that is being asked. And so customers are asking longer questions, and they may be doing that even indirectly because these alarms are synthesizing the questions as you're interacting with them.

00;10;17;14 - 00;10;37;23
Unknown
And so it's not a large industry report or report on what keywords are trending. It's a focused inside look into exactly what A.I. sees when it's generating an answer that you think should include your brand as part of that answer. And what I think is going to surprise a lot of people is, is that they hear a discovery.

00;10;38;00 - 00;11;03;16
Unknown
And so you think, no Chad, GPT or Gemini or perplexity and and those do totally matter, but they're just the visible tip of something much bigger. A.I. Discovery is already embedded across the infrastructure. Meta, of course, uses A.I. models to decide what products to show in your ad campaigns. Instacart uses AI to decide which brands to show when someone searches for like trail mix or electrolyte powder.

00;11;03;19 - 00;11;30;20
Unknown
Strava does the same thing for gear for runners and cyclists, and everybody's got a new browser extension that's interrupting your shopping session, giving you an alternative Capital one. You know, does that to me all the time. And so these this A.I. infrastructure is across the landscape and it's and where the people are actively looking to buy. And so and AI is the one that is quietly and sometimes not so quietly deciding which brands they see.

00;11;30;22 - 00;12;02;16
Unknown
And so we designed reach to show you how your products perform across that landscape. And it is a very detailed look, layer by layer of retrieval because I has these different mathematical layers that represent, you know, how you're vectoring towards the find ability and how you're vectoring towards the answer ability of what's going on. And so when we follow that question, we can see the third party sources that are part of that answer generation, the trade publications, forums, your competitors.

00;12;02;18 - 00;12;26;06
Unknown
And we can really get an inside look with objective scoring that says, Does I understand what your product is and who it's for? Because going back to like the semantics I talked about before, if you have designed a rain jacket that is specific for trail running in the Pacific Northwest and AI systems, just think that you have a generic waterproof jacket.

00;12;26;08 - 00;12;52;02
Unknown
You're basically invisible for the exact kind of queries that you should be crushing it on. Mm hmm. Obviously, once you rank their, you know, your position, how you do start optimizing to get a better position, is that something that you directly can conclude from the results that Rich gives you? Yes. Yeah. If we built it that way. So what happens when you run one of our rich reports, you run it for a very for a query.

00;12;52;02 - 00;13;12;17
Unknown
And as part of a process, we are very intentional about the queries we select that we can come back to talk about more later. But we run that query through and you will see exactly where you score on these layers. On the web search candidacy, on the what's called the by encoding step, you can kind of consider this like a quick airport scan, right?

00;13;12;20 - 00;13;41;19
Unknown
All the way down to the final step, which is called cross encoding. And this is more like an MRI, right? But then again, we don't have to get into the details. But at each of those layers, you're actually getting a score of how well your content embeds relative to the question that's being asked. And so if a competitor is beating you on that query, we can we can surface your content and we can surface the competitors content and we can say how they're winning and by how much they're winning for.

00;13;41;22 - 00;14;05;12
Unknown
And as part of the REACH program, you can actually test your content, your, you know, the content that you want to replace, your old content. You can test any content you want and get that scored in real time against the same candidate set so that you can be ensured that you're getting higher scores along all these different vectors as you as you basically get gated towards being part of the generative answer.

00;14;05;14 - 00;14;24;21
Unknown
Very interesting. And you already answered my question in regards can I spy on my competitors and see what they're doing better than me? So that's that's great that this is already implemented there. Now, obviously, our listeners want to know is like, how can they make more money from this? Are there any ways to see that you have a increasing revenue?

00;14;24;21 - 00;14;57;23
Unknown
Do you have any case studies on this or in the examples? One really fun one is, is that and I've heard it cited before by guests on your show is that I air driven customers convert at a significantly higher rate and abso that's absolutely true. So everybody wants higher conversion. But here's a really good example is that we were working with a fishing bridge and anybody out there familiar with fishing space is, is that they have enormous crossover in their imagery and their aspirational language, regional interests.

00;14;57;23 - 00;15;19;21
Unknown
And you've been the places where you can go buy it, but there is zero overlap in product, had very little overlap and customers fly. Fishermen and conventional fishermen are completely different consumers. And so the brands have always struggled with attracting the wrong audience. Right? Is someone looking for fishing gear, lands on a fly fishing site and they're going to bounce almost immediately.

00;15;19;21 - 00;15;45;04
Unknown
And so now you've paid for that click. Andrew Conversion metrics are going down. And of course, if you don't have the negative keywords dialed, that can kind of backfire on you. But so anyway, when we started with this brand reach gave us a chance to look inside the box as we talked about, showing us how to make sure that the agents on, on all of these vectors understood exactly what the products were designed for and what problems they were designed to solve.

00;15;45;06 - 00;16;17;13
Unknown
And well over the first 150 days, we saw a 31% increase in conversion rates. This was yeah, this was really fascinating to us. And and again, you hear this over and over again. Air driven consumers purchased at a higher rate. And the interesting thing that we haven't heard discussed where we dug into the data, the visitors arriving, we're much better qualified because we had done a semantic understanding was reducing unqualified traffic significantly.

00;16;17;16 - 00;16;48;10
Unknown
And so that's a powerful tool, a tool that attracts the right customers should be just as good at protecting you from the wrong customers. So all of that money, they get spent on ads delivering the wrong customers, just think how that can compound in your direction once you're finding these customers in a better way. Mm hmm. Now, that's very interesting because everyone who is in marketing and e-commerce marketing for a while knows this pain points that you're spending money, but you're attracting the wrong traffic now, which is part of great tank and you're on.

00;16;48;12 - 00;17;11;21
Unknown
This is what I really like your also Shopify Pop and so Klaviyo partners, you're coming from the e-commerce side of things. Who's your perfect customer who starts working with Reach and sees the most results the fastest? The perfect customer for us is a brand that wants to win based on the passion for their products. They know who they built it for, why they built it, and why their solution or product is different.

00;17;11;23 - 00;17;38;17
Unknown
And honestly, that the fun ones are the ones that are kind of a little bit frustrated, right, with the old world because for decades, you know, Google and search engines, they cared more about backlinks and domain authority keyword density more than they cared about whether you made a great product and you know, that was the game. But it's kind of stunk because it meant that authenticity and product knowledge weren't really discovery assets, But that's changed.

00;17;38;20 - 00;18;12;16
Unknown
And so the brands that we're most excited to work with are the ones that are ready to dig in to identify the questions and ideas that surround each product and their content, update their content to what the AI systems are actually evaluating. And so what this really does is we've always specialized in challenger brands, and this really supercharges Challenger brands because we get to see all the time how A.I. gives them a real chance to compete for those brands that either have that Head Start in SEO or they just have those unlimited budgets to really just brute force their way to the top of the pile.

00;18;12;19 - 00;18;36;20
Unknown
But I will add one thing is, is that we were presenting this to a an executive from a ten figure brand and his first comment was he's like, Well, the heck with my ecommerce team. I want to show this to my product development team. And that's that's a really good instinct because what Rich Root really reveals about how I understand each your products isn't just a marketing insight, it's a product insight as well.

00;18;36;23 - 00;19;03;15
Unknown
And we've had brands approach us about actually helping them prepare for product launches once they see how well this all accumulates in the favor. Mm hmm. Giving you a take on that a lot of Amazon or integrating like Shopify, Amazon directly in the conversation into the chat, what kind of challenges and opportunities do you see there? Yeah, I think this is a great opportunity.

00;19;03;17 - 00;19;32;23
Unknown
I think it might be the most important development in e-commerce in 2026 and specifically, you know, the one that we've all heard about or maybe some of your listeners hasn't. Is this Google's universal commerce protocol? Walmart's Etsy target, Wayfair, have all endorsed it themselves and they're kind of the early adopters on the way. And what it does is it's going to create a standardized way for agents to discover, evaluate and then sell your products on behalf of the merchants.

00;19;32;23 - 00;19;59;18
Unknown
Right. So that that clicklist that you were talking about. This is how that's going to be empowered. And the really cool part about that is, is that you're going to get a again, it kind of goes back to that Amazon example. You're going to get maybe four lines to compete on a display page. We've looked at some of the ways that they're going to be building out UCP and it might just be a 180 character description field on which you're going to compete.

00;19;59;21 - 00;20;24;19
Unknown
And so what we can do with Reach is again on both A by encoding and across encoding factor and all of these factors that relates to exactly how A.I. does semantic understanding. We'll let you have the most impactful 200 characters or whatever the limit is that most precisely, help A.I. systems find your product and understand that it's answering the customer question.

00;20;24;21 - 00;20;43;21
Unknown
Yeah, it's important to to know what you need to do because you don't know what you don't know. And Rich helps you in finding out exactly the right steps to get your right and your products in front of the right customer. Walk me through the typical onboarding process of a new user. What steps are involved? How long does it take to get up and running?

00;20;43;23 - 00;21;04;26
Unknown
Yeah, I mean, the fun part is, is that since we're competing based on ideas and meaning and not just technical infrastructure, we can start generate generating insights almost immediately and we like to get content updates onto your pages within the first few weeks of watch. Okay. That's very, very quick. Next question, obviously is how is your pricing structure working?

00;21;04;26 - 00;21;28;28
Unknown
That's always something in the minds of marketers. Yeah, no, this is great. What we do is, as you've talked about, we are a very integrated approach and you can imagine how traffic takes approach of, you know, powered by reach that can influence how well we may run your, your Google campaigns, your downstream, how we can build out your Shopify store to syndicate appropriately for this oncoming you, Google's Universal commerce platform.

00;21;29;00 - 00;21;57;24
Unknown
So we do have pricing structures that integrate reach into a larger approach, becomes very affordable for smaller brands. And then some of these larger nine 9 to 10 figure brands we work with, they just buy it as a standalone product. So we do have pricing and availability for all sorts of brands. And you know, and if you're a small brand, you know, small targeted approach can be the very impactful and powerful right out of the gate for you.

00;21;58;01 - 00;22;31;13
Unknown
Yeah, I'm a big fan of small brands of founder start ups, and I think that's a huge opportunity right now in the market to get in early and do it the right way before a coffee break comes to an end today. Is there anything you want to share with our listeners that we haven't covered yet? Yeah, I wanted to talk about a little bit, you know, exactly how they can understand what's going on right now in that, you know, Google search is still, of course, very important and one thing that's really important to understand about Google search is, is a Google search is already doing this a bit already because Google search is now layering

00;22;31;13 - 00;22;51;16
Unknown
on a by encoding model as part of their as part of their answer generation. And that's one of the steps that we measure directly. So as brands are, you know, especially some of the smaller brands, we're like, Hey, is this a bit of a leap? Or, you know, hey, I'd really like to get search dialed. This does have a direct recursive effect on on search.

00;22;51;24 - 00;23;22;01
Unknown
And that was the first thing that we saw as we started beta testing with these brands is we had to challenge our brands were firmly stuck in second place on search behind two very big brands. And so very respectful place to be. And once we started putting these fixes in place and maximizing for the by encoding that occurs with large language models, both of them independently within weeks vaulted ahead of these much bigger competitors in classical Google search.

00;23;22;04 - 00;23;45;00
Unknown
So this is something that can start to have an immediate impact for you right away, whether you're, you know, feel confident about optimizing for all ends at a larger scale or you just want to start closer to home. Mm hmm. Very interesting. So, yeah, before we come to an end, I think it's a very interesting time. I think it's the biggest change in search in the last, I don't know, 25 years or since the interwebs is around.

00;23;45;00 - 00;24;04;08
Unknown
And it's so important to be part of it. And as I said before, I think smaller brands have a huge chance there to really make a difference. Where can people go and find out more about you? Yeah. I would encourage them to go to our new microsite that's dedicated to this. It's called Reach dot, Crank, Tank, dot Net.

00;24;04;11 - 00;24;23;07
Unknown
And you can go there and find out a little bit more and then book a free demo where you can give us your URL. And the question that you are most curious about. Give us any question. You know, if you're I think you had a good idea earlier if you're a little bit shy and you'd rather have us take a look at one of your competitors.

00;24;23;09 - 00;24;49;11
Unknown
We're happy to do that, too. And I think once we show you how this works and how you can grasp it and how, you know, once, once you get past the complexities, how simple and transparent it is. And that's the real the real change is that I would encourage everyone to change their mindset on, is that, you know, of course Google is going to keep their algorithms set, you know, PageRank hidden because that's their you know, that's proprietary to them.

00;24;49;13 - 00;25;09;06
Unknown
But with large language models being open source and this highly competitive landscape, you can and truly see how you can increase your impact within these large language models. Mm hmm. I will put the links in the show notes, as always, and just one click away. And I hope a lot of listeners will take you up on your offer and just give it a test.

00;25;09;07 - 00;25;26;21
Unknown
I think it will open your eyes to see what's really happening out there. And is that Google is there, Google will be there, but they're not the monopoly sort of candidate they were in the past. And it's opens to market for many, many other competitors and chances to get your word out there. Cool. Thanks so much for your time today.

00;25;26;21 - 00;25;32;17
Unknown
Keep you updated on what's happening in the air and hope to talk to. So I totally will. Thank you so much, class. Appreciate it.