The Evolution of Ecommerce Search — Arv Natarajan | Why Conversational AI Simplifies Queries, What Makes A Perfect Search Page, How Visual Search Improves Discovery, Why "Revenue Per Search" Rules, Why Search And Browse Both Matter (#365)

Enjoying the Ecommerce Coffee Break Podcast? Here are a few ways to grow your business: https://ecommercecoffeebreak.com/level-up/
---

In this first episode of our special mini-series on ecommerce search, Arv Natarajan, Director of Product at GroupBy Inc., explores how artificial intelligence is revolutionizing digital customer experiences. 

From early keyword matching to today's sophisticated AI-powered solutions, Arv breaks down the transformation of search technology and its critical role in creating personalized shopping journeys.


Topics discussed in this episode:  

• How early e-commerce search failed due to basic keyword matching. 
• Why AI understands shopping intent vs just matching words. 
• How search personalization works without compromising privacy. 
• Why both search and browse experiences matter equally. 
• What makes the perfect e-commerce search page layout. 
• How visual search transforms product discovery. 
• Why conversational AI revolutionizes complex shopping queries. 
• What makes unified AI shopping assistants the future. 
• How search analytics reveal hidden inventory issues. 
• Why "revenue per search" is the only metric that matters. 

Links & Resources 

Website: https://www.groupbyinc.com/integrations/shopify
Shopify App Store: https://apps.shopify.com/groupby-ai-search-discovery
LinkedIn: https://www.linkedin.com/in/arvnatarajan/
X/Twitter: https://x.com/groupbyinc 

Get access to more free resources by visiting the show notes at
https://tinyurl.com/3fftw9mw 

MORE RESOURCES

Download the Ecommerce Conversion Handbook for store optimization tips at https://tinyurl.com/CRO-ebook

Best Apps to Grow Your eCommerce Store: https://ecommercecoffeebreak.com/best-shopify-marketing-tools-recommendations/

Become a smarter online seller in just 7 minutes
Our free newsletter is your shortcut to ecommerce success. Every Thursday. 100% free. Unsubscribe anytime. Sign up at https://newsletter.ecommercecoffeebreak.com

Rate, Review & Follow
Enjoying this episode? Help others like you by rating and reviewing my show on Apple Podcasts. Rate here: https://podcasts.apple.com/us/podcast/ecommerce-coffee-break-digital-marketing-podcast-for/id1567749422

Follow the podcast to catch all the bonus episodes I am adding. Do not miss out. Hit that follow button now!

Transcript

Welcome to the E Commerce Coffee Break Podcast, where we share actionable tips and strategies to help you grow your online store and sell smarter. I'm your host, Claus Lauter. This is the first episode of a mini series where we talk about the evolution of e commerce search. Today we will talk about the power of AI and how it impacts the digital customer experience.

Joining me for this mini series is Arv Natarajan. He is the Director of Product at groupbyinc.com. So let's dive right into it. Hello and welcome to another episode of the eCommerce Coffee Break podcast. Starting with the new year, we want to do something new. We will have a mini series and we want to dive a little bit more in the evolution of eCommerce search.

And it has a lot to do with user behavior. It has a lot to do with AI. And we want to find out how a personalized path to purchase can help you with the help of AI. And we will dive really deep into search because every customer journey most likely starts with a search somewhere on the interwebs. And we want to find out what makes your store more accessible.

Joining me on the show for the mini series, a guest that has been on the show before, Arv Natarajan. He's the director of product at groupby, and he's a passionate entrepreneur currently responsible for product management at groupby. So let's welcome him back. Hi Arv, how are you today? I'm well, thanks for having me again.

So when we look at the early days of e commerce search, what do you consider the biggest limitations that frustrated merchants, retailers, and customers the most? I think one of the biggest limitations with the early days of search was the TikTok technology. And the way that it used to work was based on pretty simple keyword matching.

So when someone searched for a basic query like jeans or dress or shoes, The search engine would try and find products in your catalog that contained those keywords and return those products to the top of the list. The more times the keyword was contained in the product report, in the details of the product, it might increase the importance or the relevance of that particular product.

And so then that might be boosted towards the top of the list. But you run into problems that way. A simple problem might be where if you sell, let's say, third party products on your website, you might have suppliers that try and game the system by pummeling in a lot of the same keyword into that description or something so that it's ranked higher in the relevancy.

That's a way of gaming the system. Or you might run into other concerns where, let's say someone is looking for Medication without a certain ingredient, but that is in the keyword. You don't want that particular product that contains the ingredient to come back just because it matched in the query. Like if someone is searching for, uh, painkiller medication without ibuprofen, you don't want ibuprofen containing products to come back in the result set.

So the, the very simple technology that existed in the early days of the search. Could lead to problems in the search experience. And we've come a long way from that. I remember the good old times when I was doing SAO with keyboard stuffing and you were just putting white text on white background with the keywords and it would work.

That was fun times, but that's a long time ago. So. There's a big obviously with AI, there's a big change going on in search in general, the big search engines with large language models. I think 2025 will see a lot of changes coming on that side. How would you explain how AI first search platforms work?

The way to think about AI and the impact of AI on searches is how it makes our lives easier. So all of these advances in technology in 2024 and this year where AI has become super prevalent in the industry is all aimed at making the life of the shopper or the life of the merchant easier in managing that experience and helping people find the products they're looking for.

So there are a couple of ways to think about it from a user experience, from a shopper experience point of view, The AI is understanding the intent of what the user is looking for. And we can go through some examples. An easy one might be something as, as simple as date night dress. That's one that I like to use with a lot of our clients.

If someone is looking for a date night dress, there is this very specific type of product or dress they're looking for. And what you want to make sure does not happen, and could happen perhaps with the legacy technology, like a, like a default search engine that Might still be used is dresses that contain date patterns might be coming back or night dresses might be coming back in the result set.

That's definitely not what the user is looking for. They're looking for something to wear in the evening where we're on a date. So an AI can be used to help translate or understand the intent of what the user is looking for to return the right product on the merchant side of things. We want to be able to create these rich shopping experiences.

at scale without having to go in and change every little detail. And that's where AI can help us as well. It can manage the shopping experience, curate the results in a way that is more relevant for the, for the shopper without having to go in there and tune every little time. So those are, those are ways that how AI is, is really helping us and it can help us in the future as well.

I think a lot of merchants might be interested on how AI is finding out the customer behavior and the intent when a new visitor is coming to your store. What kind of data do you work with to make this happen? Yeah, so there's, there's many data sources out there that can be used. Uh, the first and most important really is AI.

is how users interact in general on your website so you can install tracking pixels they're called which capture user behavior what people are searching for what are they clicking on what are they adding to their basket and what are they buying and once you have that information you can feed it into an AI type platform that is able to, in the aggregate, understand, in general, people who search for genes care about these kinds of products.

So it's, it's, that's the, the general way of, of AI helping individuals, especially when they're new users coming to a website. They don't have any past history on that, on that website. So it's difficult to personalize results for them immediately. But within a few clicks, you can understand the preferences of an individual and then start tailoring the results.

So capturing user behavior is key. A new user that comes to your website, let's say they're looking for jeans, might have a certain preference for gender, or style, or brand, and that can be fed into an AI system that can then start personalizing the results. So if they're looking for jeans, it may be women's jeans or men's jeans.

It might be a certain type of mum's jeans or or skinny jeans. There are things that can be fed into the AI that it can learn from to then return very relevant results. Now we're talking about personalization here and data, personalized data. A lot of people might have concerns when it comes to their privacy.

How do you balance that? Yeah, it's such a key concern. And I think one that everybody should really be aware of is, is what information of theirs is being used. I would say in general, most websites, uh, acting with reasonably good intentions. I would think there are obviously some bad actors out there, but.

In general, the way that these AIs typically work, and the way that we work with our AI, for example, is we only take in non personalized information. So we don't care that the individual is John Smith or Jane Smith, we just want to know what behavior or preferences that particular individual has. So we can tailor a more rich and fulfilling shopping experience.

I know when I go on social media and I'm looking at ads, I prefer my ads to be more personalized because I'm a sucker for that kind of thing. I'll look at the latest Instagram ad and probably make a purchase because it's something that's relevant to the things that I care about. I find that experience more satisfying for me.

It's easier for me to find the products that I'm looking for. It's tailored for me, but it's not necessary to create that experience. only by looking at the personal data of the individual, but we can look at the data in general and understand what kind of person that is, what, build a persona around that individual and, and return very relevant results.

I think that's the biggest problem right now in e commerce that people don't want to have the data out there because of privacy concerns. And on the other hand, you just gave the perfect example. You want to have tailored ads. So that's a bit of, um, yeah, chicken egg thing where it's like, that's where Yeah, running on a very small edge.

Now, obviously personalization helps with boosting sales. Can you give me some examples how that works and how merchants can get the most out of it? Yeah. So like I mentioned, understanding and capturing that user behavior is key to driving a personalized experience. We need to know what actions this person has taken in order to create that richer experience.

So, Capturing what they're searching for, their, their, their preferences for brands, for, for styles, for types of products, uh, is key. And then when you create a personalized experience, it's also not just about the products that you return, but also the, the other parts of your website that might drive the shopping experience.

So that could be content. You could have personalized banners, for example, if you know. The group, the cohort of individuals that are coming to your website. You can, to that particular site, you can create, let's say a back to school experience or a, or a, um, a wedding experience for people that have upcoming weddings.

Cohort based personalization has been around for years. And so that's one way of managing it. I think technology now is also at the point where, I mean, you may hear buzzwords like hyper personalization. It's, it's another way of just saying it's personalized to the individual. So it's, I would see a different shopping experience to what you might see.

And that is also done through similar capturing of user behaviors, and then feeding AI platform to create the shopping experience. But really the technology is there and it's up to us as Shopify managers to take advantage of it. The default engine that you might be using, if you haven't touched it, will most likely not be able to take advantage of it.

You will have to look to a more sophisticated technology lab to help you with that. Talk me through how the perfect search page looks like. Obviously, you have your search bar on your store, hopefully right at the top, like Amazon gives a good example, and they know why they do that. But how can you substitute with other information to make the search better, easier for the customer?

It's such a nuanced question, I think. There are many ways to consider a shopping experience. There's actually two ways. Two main paths that users will take when they visit an e commerce site. One is search and one is browse. So when you search, you're searching for a particular query. Date and address was the example we had.

When you browse, you're actually navigating through your taxonomy. Of your catalog to find the right category of products like dresses And so it's important that both shopping experiences are treated With the right level of attention so that they are they are very easy for the user to find the products They're looking for if your search engine sucks Then you're not gonna be able to return the right products if the, if the com, if the user query is complex, or even if it's simple and it's not a very good search engine, you may not find the right products.

If your site layout is, is not great and your users can't navigate to the Genes page or the dress page, then they're also not gonna find the right products. So understanding the, the right UX is gonna be key. There are websites out there, uh, platforms out there that do free audits that, um, will help you with best practices.

And so, and even like you said, looking at Amazon or some of the leaders out there in the industry of how they lay out their search experience, or even their shopping experience, is a good starting point. You want to have, for example, an easy way for your users to see all the different categories of products that you carry.

That's your taxonomy. So that's, that's, that's one thing that has to be there. You mentioned the search box already. Having easy to click on links or, or images or content that lets your users find the right products. Let's say you have a sales page for Valentine's Day coming up in a couple of months already.

Maybe that's a, that's a page that you might want to surface on your homepage through a, through a banner or an image just to help your guide your users to that particular landing page experience. Um, you might want to have. Subcategories for each of your main categories as well, which kind of guide your users down to more specific sets of products.

Dresses or jeans might be very broad. If you sell men's and women's, you might want to have separate pages for each. So these are all ways that through UX you can help guide the users to find the right products. At the end of the day, that's your main responsibility as a merchant is to help your shoppers find the products that you sell.

And any time that they can't find it easily is an opportunity for them to leave and go to one of your competitors or to Amazon. So having a good shopping experience is going to be key to improving your sales. I want to touch on this and just open this a little bit more because we know that the the future of e commerce search is changing.

So now we're talking about voice search, visual search, conversational AI. Talk me through the different things that are already there and the ones that are coming. This is so exciting. I think there's so many different Technology advances that are really helping improve how people can find products. Um, one is image search.

I think image search has been around for a while. We're all familiar with taking a picture of our couch and uploading it to the IKEA website to try and find couches that look similar, something like that, right? But I think where it's getting better and better is one, the technology has improved to make it easier for merchants to use image search.

And then we also want to think about the applications. So let's say for example, we've got somebody, a carpenter out in the field who's doing a renovation. There's a missing screw in, in one of the things that he's setting up. He can take a picture of another example of that screw and try and find that that same screw in the product catalog.

Now you've got some nuances there that you might run into, like What is the dimensions and the size of the screw, you have to find obviously the exact product. But if you can help guide the user down a path that, similarly if they walk into a hardware store and said I need more of these screws, with to an associate, you can have that kind of experience where it's, uh, helping them find the right products.

That's with images. Even another example of images might be if you take a picture of a social media post, you've seen Taylor Swift or the Golden Globes, you want to buy the same dress. So you can upload either upload that image to a website or take it into a store and show an associate and they can do that on their site and try and find a dress that looks like that.

That is an example of where we can use this kind of technology, AI technology, to identify what is the product in question, and then do a search based on just the image without having to take in any other information. That's pretty cool. I think we've got a lot of applications there that, that can help people, uh, create a better shopping experience.

On the conversational side, similarly, let's, we talked about how the types of queries have become more complex since the early days of search. We've gone from dresses to maybe something like date night dress. What if we took that a step further and said, I'm doing a renovation for my bathroom, what do I need?

That's a very complex question that a search engine, a basic search engine can't answer. So you would have that type of conversation with somebody in a store, but not an online search engine. But technology now is getting to the point where you can type that in as a prompt. If you're familiar with prompts and GPT, you can type that in as a prompt to us, to an AI and say, I'm doing a bathroom renovation or kitchen renovation.

Tell me what I need to buy. And then it will list you all the products and it'll give you a, an article on how to do your renovation. And it'll, it'll provide all the information you need to complete that activity. And we can now look at search experiences moving in that direction as well. So imagine if you could go to.

The website of your hardware store, type that in and it'll be able to present you exactly those things, a buying guide for all the equipment, a list of all the materials you need, and a step by step instruction on how to renovate your kitchen counter. Like that would be a very interesting shopping experience.

Technology is there now. We can actually do things like that. So, um, that's something I think we can, we can look forward to as well as, as, as users. I know me as a consumer, I'm very much looking forward to the time where I can do something like that. Um, and, and you mentioned voice search. I think voice searches is.

Again, pretty standard technology that has been out there for a while. The way that I think about that is, is a basic voice to text conversion where you're just taking what the person is saying, transcribing it, and then submitting that to the search engine. To me, that's pretty table stakes, I think, and most platforms should be able to do that.

I think from my perspective, voice search, visual search, conversational AI will merge together. It will be just one sort of AI agent in the store helping you with everything. And then we're basically the same situation in a brick and mortar store where you would talk to a person. Now you provide a agent with all the information you have.

Mind the pictures, talk to him, um, to find the right product. Now, why are you doing exactly that and helping merchants with solutions for that? Um, talk to me about your perfect customer, who comes to you and how does it work? Yeah. I mean, I think the perfect customer for us is one that hasn't had much luck with their search engine.

They're aware of The experience being poor and they're looking to make a change. I think those are the ones that will see the biggest impact. If, if they are coming to us with a perfectly tuned engine, they've already gone a step in the right direction. So I'm very happy with that. But for us that the customers that we help the most are the ones that.

have not tuned their engines, are unaware of how poor the shopping experience is at, or maybe they're aware of it, but they don't know how to solve the problem. And so they need help. A question I like to ask is, when was the last time you did a review of the search experience on your Shopify store? If the answer was yesterday or last week, then you're a step of the way there.

But if it was, it's been a while or it's never, then that's definitely a problem and something that we can help with. So that's really where we like to start is. Let's do an audit of your site search experience. Um, what do the basic searches look like on your website? What are the top searches on your website?

If you don't have this kind of information, if you don't have even simple analytics to tell you what your customers are doing, that's step one. We need to understand how do your customers interact with your website. Okay, most customers go through search. Well, actually most customers go through browse. So then we can look at where we target and improve that experience.

Um, once we understand the basic layout and the, and the state of, of the search experience, then we can find things to improve. An example would be, okay, you're currently using the default Shopify search. It is okay if you're, if you have a very basic catalog and if your customers are fairly standard, but if anything goes outside the lines of the norm, then you're running into situations where there's problems.

So, Let's look at how do we improve your search engine itself. That might be one step. Looking at analytics, do you understand what your users are doing on your website? Let's start capturing data. And you don't even have to do anything with it apart from just looking at it from time to time and understanding For example, are there any products that are out of stock that I need to need to buy it so that my customers can actually purchase those products on my website instead of going somewhere else.

A very common problem we've got with some of our prospects coming in is they have inventory issues and they're not even aware of those inventory issues. So if they don't know that they've run out of Hello Kitty backpacks because there was a viral post on social media last week. And they need to buy some more backpacks.

That is a problem for the merchant. They're not, they're missing out on those sales. So how do we help use analytics to help guide some of our decision making? Uh, we like to provide those kind of actionable insights that, that users can take. If we can help merchants create campaigns, like I mentioned Valentine's Day coming up or, or any seasonal type events.

If we know, we know these days, these are, these are the same every year. We can help create rich shopping experiences for those specific merchants. Sales periods to help get and help create a richer shopping experience that at the end of the day, the goal is, how do we find the right product and give that to the right person at the right time?

That's what we try and help our customers with. I think it's a very important point to say that you need to have a set of fresh eyes to look on your store because as a merchant, you probably don't see the force because of the trees because there's so much in the topic and yours. Search, um, on your own store probably will work very, very different than to someone who comes to your store for the very, very first time.

And I know a lot of merchants who do not look even into their search terms, what people are searching for on their website. So I think having someone who's going and really auditing and analyzing your store is a massive help. Walk me through the typical onboarding process of a new customers of yours.

So with Shopify, actually, we've made it very easy. We have a Shopify app that can be installed, just a one click installation, and it'll connect to your Shopify catalog. It might take a little bit of time, depending how big your big, your catalog is, I'm talking an hour or two to upload your catalog into the search engine.

And then you're ready to run with it. If, and if you use the default theme, it's as simple as that. If you have obviously any customization to your Shopify theme, there might be some more connections that you have to take advantage of with Groupon, but the, the most basic or default themes work out of the box with, with the app and.

And it'll take over your search bar and you now can take advantage of group by search. The customizations, if you do run into any, any situations where you need to customize the theme a little bit, we have a team of people available to work with you as well to work on those customizations. But it should be as simple as a one click installation.

Talk me through the results you saw with your clients. What kind of KPIs do they need to look at and what kind of growth do they see? Yeah, I am. It's actually a great question because it's going to be a topic for the next conversation we have in a little while. I, I won. Won't share too many details about that customer because I wanna save the, tease the audience and then give them a reason to listen to next time.

But really, we, we've seen over the default search engine, significant uplifts in key metrics like revenue per search and overall revenue. Really, that's what the engine is designed to do is, is improve revenue. Like at the end of the day, we want to help merchants. And that's what we've built our engine to do and we work with partners like Google and others to actually improve the AI on that engine.

And that's what merchants can expect. Once they switch over to something like this, it really does have a significant impact on revenue and revenue per search. Revenue per search, if you're not familiar with that particular data point looks at how effective your search engine is. So how much revenue are you making for every search that happens on your website?

And that's what a search engine is meant to do is be more efficient in that key metric. Other data points used to be important in legacy or early days of search. People used to care about things like the number of times no results were shown to the user because it then Stopped the user from being able to buy anything or convert on the website Nowadays you have other ways of cross selling or upselling if you end up in a situation where you don't carry Helikitty backpacks for whatever reason you might want to direct them to other products that are similar so other kinds of backpacks or other kinds of categories of products that might help them You they may be interested in as well.

So, so there are ways of, of improving some of these other metrics that used to be important that are not so much anymore. But to me, if the number one metric that merchants should be keeping an eye on at all times is, is revenue per search, when they're looking at their search engine. We will dive into this in the next episode.

So we'll leave that open for today. Is there anything that you want to share with our listeners that we haven't touched on? Yeah. I mean, I think it's the call to action that we, we kind of, What I want to help people think about, which is when was the last time you looked at the search experience on your website?

And I think that's, that's where you all merchants should be starting from when they think about searches. Let's go and review it. Let's see how it, how it works. Let's see if you run into any situations where there's bad user experience, like it may not be an issue with the search engine. It might be that the layout of your page needs to be optimized or that you're not taking advantage of some of these best practices that we've talked about.

So I would say that should be step one is do an audit or a review of your website. You can always do it yourself. I think there, there are guides out there to help you consider what is best practice. And if you need help, there's help out there as well. No, I think that's the very first step. And if you're looking to optimize your search, you definitely should start there.

Where can people go and find out more about you guys? Yeah, check out our website on www. groupbyinc. com That's g r o u p b y i n c dot com Uh, and we'll post a link in the show notes, I'm sure And, uh, yeah, feel free to reach out We'd love to chat to you if there's something that you think We could help you to improve the shopping experience on your website We'd love to have the chat.

Excellent. I will put the links in the show notes as always. Then you just want to click away. Uh, thanks so much for this very first episode. Again, there will be two more episodes coming up. So stay tuned. We will dive more and more into search to make your store search better and to bring your growth through search.

I think it's one of the most important thing, because as I said, that's where shopping normally starts with. Thanks so much for your time today and talk soon. Thanks for having me.