Everyone, and welcome to today's webinar. Three ways AI powered search drives up profits. We're excited you join us. My name is Lara Thiesel, and I'll be moderating today's event. Now today, we are going to take a deep dive into how high relevance AI powered search capabilities help retailers achieve their goals. Now before we get started, let's go through some quick tips to make sure your webinar experience is the best it can be. At the bottom of your screen are resizable and movable application widgets Feel free to move them around to make the most out of your desktop space. Now, the slides will advance automatically during the presentation. If you need to enlarge them simply click the enlarge slides button in the top right corner of the presentation window. If you need technical assistance, click the help widget in the bottom left corner at any time. Now as the presentation gets underway, you may submit questions using the a Q and A widget. Now we will answer as many questions as we can at the end of the presentation, but if a more informative answer is needed or If we just run out of time before we answer your question, we will follow-up via email. So please know that we do capture all of those questions, and you're welcome to submit them as you think of them. Now after today's presentation, you will receive a link with on demand access to the webinar so you can watch it again at any time as well as share content with colleagues. When we are thrilled to be joined by our speakers today, Sergio Yacabuki senior director e commerce marketing at Coveo and Vipu Aurora associate direct product management at Walmart. So to kick it off, I'd like to ask each of our speakers to provide a quick introduction of themselves and their work. So, Sergio, and Vivio, you know, could you tell our audience a little bit about your roles and experience? Sergio will start with you. Hi, everyone, Sergio. Very excited to be here today. I'm the senior director of e commerce marketing at Caveo. I've worked in e commerce for about ten years. Done a variety of roles in terms of consulting for customer experience and e commerce strategies through to now marketing SAS products, within this space. So Super excited to talk about this hot topic today. Hey, everyone. This is Bhbora here. I'm currently group product manager on Walmart commerce search product team. My team is responsible for the entire customer journey of search on walmart dot com. Website and Walmart apps in the US market. Prior to this role at Walmart, I was at Facebook where I had launched the portal online store from ground up. I've been working in the product management space for the last ten years. Had started out in the b to b startup space, building out productivity and collaboration tools, for energy industry and then move to e commerce. And since then have really enjoyed solving consumer problems in this space. Well, audience that you can tell we've got two true experts here. So let's get right to it. And to kick us off, you know, what challenges are retailer facing when it comes to meeting their customers online shopping expectations. You know, surgery, Sergio, let's begin with your thoughts here. Yeah. I think this is this is very interesting, right? And it's a great place to start. Consumers expectations are very, very high at the moment. I I argue they continue to increase, time and time again, and it's quite common for, as a retailer to compare yourself against your nearest competitor. So you'll you'll be a certain, say footwear brand, you'll compare yourself to your to your obvious peers. The the reason why that's not always the best thing to do is because consumers expectations are set by the experiences that they have that might be outside of retail. So they could be on spot by, they could be on Pinterest. They could be on Wayfair. They might have, an experience that's highly personalized to them For instance, if you log into Netflix, it only looks like it's for you. That is essentially the product alongside all of the content that they they produce. So when those customers experience that, then when they come to e commerce, because it feels like a, you know, the similar kind of experience is a digital experience. They expect the same. And we recently conducted quite a bit of research into this. It's only a couple of months old. And customers At ninety three percent of customers expect their online shopping experience to be at least the same or better than what they get in store, and I think that's a tall order for for a retailer to hit. At the same time, ninety one percent of that group, it's about four thousand people that we surveyed across the UK and the US. Ninety one percent of that group said that they encounter problems when they shop online. And then even more interestingly, Gen Z, that goes up to ninety six percent. And you might argue you'd expect people that have grown up with technology to not have problems when they're, you know, browsing digitally. But actually, I think their expectations are even higher and so if they see a problem, they'll very quickly, mention it and then potentially move on to the to the next place. So I think it's a very tall order that online retailers or if you're an omnichannel retailer to keep up with those with those expectations. Absolutely. And I cannot agree more to Sergio what you said about people's expectations being driven from what they see in other places. It's fascinating. Actually, there's actually, a law called Jacob's law of user experience which is coined by the Norman Nielsen group, which actually says something very simple, very profound, which is people spend most of their time on other people websites on other other companies' website. So their expectations are actually formed on what they are other, competitors and and other, players. So absolutely true that, you know, there's a set there's a some sense of consistency people tend to expect. So so people's expectations are set from what they see from other what they see on other retailers' website and their expectations are set accordingly. Now, what are some other, challenges which retailers are facing actually, one of the one of the key challenges we see especially in in mass market retail is specifically if you are If you are a mass market retailer, then one of the challenges you, you face when, when you are selling different verticals is that people's expectations of experiences of of different verticals, are shaped by the needs of each of the different So let me take an example. So when someone is shopping, let's say automotive or tires, they need a different sort of experience. To suit their needs as opposed to let's say when someone is shopping electronics or that they they have a different set of needs for for those particular journeys. So the challenge that becomes for the retailer is how do you have a tailored experience for each of these different verticals, which is unique for the vertical needs, where whereas at the same time, there is a homogeneity of experience so that the experience still feels consistent as the consumer is moving from one vertical to another. So it's it seems like straightforward, but it's actually a very hard problem to solve. The, second big challenge which retailers face is how to pivot quickly to macro events. And we actually saw this happen a lot in COVID. For example, we're actually seeing this happen in the current macro environment where, there's a lot of inflationary pressure like how quickly retailers can pivot to the macro events can actually make or break their ability to exceed for example, like in in in this environment, there's uh-uh there's a lack of positivity in consumer sentiment. So that means is discretionary purchases are going to be put on hold. So how do retailers react and respond to this? That becomes very important. And and this is a challenge which retailers are facing today. Yeah. Great points there. And, you know, just keeping people engaged is so important. And so to that end, you you know, what are some pain points customers experience when it comes to finding real products on e commerce websites, if you? Yeah, absolutely. So, there's there's actually a lot of pain points. Maybe like I can talk about the top two or three. The first big pinpoint as a customer, which I feel when I interact on e commerce website is does the retailer actually understand my tent. And what I mean by that is not not just exactly what I'm saying or spelling or typing like word for word, but what is behind the words which I am actually not saying? So let me take an example. Let's say I'm searching for stand on an e commerce website. So there's actually like different different types of stands. Now what is really crucial in searching for for this product. What actually happened before this, did someone actually search for, let's say, for example, EV before they search for Stan, or they actually search for, let's say, queen bed before the search for stand. And this context is super, super essential because if somebody searched for TV and then they searched for Stan, they probably meant TV stand. If they probably, if they search for, a queen bed and then they're searching for stand, they probably mean a nightstand. So as you can imagine, a TV stand and a night stand are very, very different products and very, very different. So understanding this intent behind what the customer actually searching for is really, really crucial. The other pain point as a customer is especially comes comes in, in in shape, especially in in terms of, like, high consideration purchases is when there's a desire to be educated about the product. So for example, I want to buy a mountain bike but I actually know nothing about mountain bikes or bikes in general. I just woke up one day and I thought I want to explore this hobby. So how can the retailer meet me halfway and and go on this educational journey with me and explain and teach me, like, these are the four, five, six things to consider in a mountain bike. So even before the process of iteration, the process of, active sort of consideration even kicks in. The consumer actually needs to be educated about what they are buying and what is what are the important set of features in that category? So this is, again, a big pain point. Today, what happens mostly is You search for mountain bike, you get a lot of results, and eventually you end up learning what by looking at the results. So sort of, like, do it in a second hand way, but it's not elevated to a proper, like, educational or learning loop. The third big pain point, I think, which is kind of like, a very classic, in in modern e commerce is the decision paradox. The the entire online world is actually optimized for endless versus like what I would like to say in a store, it is more like a friendly and analyst isle is just like you search for, you search for, like, a speaker, for example, which is like a ten dollar, twenty dollar item, and you're bombarded with like thousands and thousands of results. So it is a very, very hard problem to make a decision and actually choose something, from those one thousand results where all you want is like something for like twenty bucks. So, this is actually a very pervasive problem facing consumers today. You know, we did some research again. I think that this was a bit older maybe like six or eight months ago, and we asked people again. It was the same size group, four thousand people across UK US. You know, what's the number one reason why you leave an e commerce store without buying something? And it's because they can't find what they're looking for was the number one thing they put on there. So clearly there's a problem and it's a very broad top And Vubers, thanks for taking us through, like, the detail of that because I think that's that's really important and where intent comes in is, is the way to go for sure There's also some really simple things that people are not necessarily getting. Right. You know, you know, it can be the case that, some searches at the moment don't deal with misspells very well. They don't deal with synonyms very well. They do very literal searches of what you put into, the box. And to Vivi's point earlier, you kind of need to work out based on what that person's been doing, what it is that they mean by those words they put in the box. And in general, in e commerce, people put and be by wonder see the same, I'd imagine you do one to two keywords in the box. They very rarely put loads and loads of stuff in there for you to get in ten from. So it's it's about understanding their their session for here in previous purchases like that. But if we even think about some of the simpler issues again, data quality is a is a big problem in in search, you know, not having a catalog that's fit. And some some businesses will have this sorted and it's not a problem, and some will be struggling every day with the fact that their catalog doesn't truly represent the products and that they have because there's mistakes in there and it's a very, very surprising amount of cat catalogs that are actually human curated and typed in still. So that causes a lot of issues as you can imagine as people have to put things in. You know, there's there's like a three to four percent error rate there, which is pretty high when you've got thousands and thousands products and you change them frequently. The other thing about search, and if we just broaden it up slightly, just the product discovery in general, is showcasing things that are relevant to them in the right moments in their journey is a very, very important way to how we can engage customers better and and, you know, help them enjoy the experience trying to discover the right products for them. So that can be the search. Yes, of course. But it can also be what you recommend to them on the homepage. It can be what's trending, what's popular, what's popular in their area specifically, but it can also be how you curate the listing page, you know, what facets do you give them to filter with and how do you generate those facets and how do they change based on what category you're looking at, all very key aspects of look, looking through a catalog essentially. And I completely agree with VBoo in the sense that stores are set up in a very friendly way to, understand what products are on offer. E commerce stores actually make it quite difficult to understand what's on offer unless you can understand that person's intent, you know, behind behind the scenes. So the customer is giving you the information. They are usually making multiple searches. And so let's let's take a fashion example, for example, for for for argument's sake. Let's say someone has done three or four searches for long black dress, but then they're getting results still for shorter ones. So they go again and they just put black dress in. We should, at this point, no longer show them the short dresses or at least change the ranking that the the longer ones are up the top, because we should be understanding what they're asking for in those sessions and fulfill those needs. Absolutely. I love I love that point and could not agree more. So that's that was a great example there, Sergio. You know, so let's talk about, you know, what customers are looking for content. What are their pain points there, Sergio? Yes. Again, this is something you can get sometimes missed off because we're all obsessed with the products themselves, and we're trying to make sure that they're showcasing the best way possible. To Vibo's point earlier, Sometimes you don't know exactly what it is that you wanna buy. You know, generally what you wanna buy, but you don't know which one you're gonna go with. In some industries, like if we go to B2B, in general, people know what they wanna buy. They're putting in a product code specifically. They're gonna buy that thing. But in in B2C, there's more of a discovery based experience going on here. So you need to showcase some content alongside that to help customers understand what it is they wanna do. So if you sell barbecues or if you sell decking and materials to build things, have if you do have and you've spent probably a lot of money on creating the content, how to build a decking or how to set up the best, you know, charcoal barbecue showcase that. And, again, use the intent that you have of that customer you know what they're looking for, surface your great content because content increases conversion rates. And, you know, we're talking about how to boost boost that today and boost profits. Content does that, but you need to have a system to index all of that, understand that content and effectively its sentiment and and what it's about and how we can link that to products that are in the catalog as well to know when to surface it. So that's a that's a really, really important part. And the other the other thing is that a lot of, how do I do this sort of questions? Start on a dot com site. Right? Like if if you've just got a product and maybe you've just received it, you go usually to that that, that retailer's store and you you search how do I use this? You know, where's the guide? Where's the manual? How do I set it up? That happens there too. And it's a very, very frustrating experience. If you search that, but you just get products again, then you can't find effectively, the experience you're looking for and the help you'll be looking forward to use that product. Yeah. Great point again there. Thank you for taking on that question. You know, retailers, obviously, want to minimize those frustrations, you know, right? And as they look to do that, what are some of the pillars of great online customer experience that they need to keep in mind. Vip you? Yeah. Of course. So, maybe, like, if I can add a lens of the search journey to to this, question. What are the pillars of of great online customer experience for the search journey? So maybe if we can try to visualize, you know, I'm I'm trying to gonna portrait like a few pillars, or or a few like word skills and a few horizontals, which are very crucial for having a good search experience. So the way the the journey starts in in search, ecosystem is it starts from an expression. So that's the first pillar. How easily when we facilitate the consumer to express themselves. Once they have expressed themselves, the second piece is how, how well can we understand what they are saying? So that's the second pillar. The, the third pillar, once we understood what they are expressing, can we actually go and find that product from our catalog, from our search index, and can we actually rank in a in a meaningful way so that people can find them on the first page at the first half of the of page So that's pillar number three. Pillar number four is when we actually, like, present the results where people can actually see them, can people navigate and refine the search results? And pillar number five, which is the last pillar, is the decision pillar are we providing all information everything in place so they can actually make a decision to their search question? So these are the five pillars. Now, There's also like foundational pieces which support these pillars, and these foundational pieces can actually be applicable to all these pillars So foundational horizontals are, vertical experiences which we just touched upon recently. In different verticals, auto versus beauty versus tires versus apparel, are we actually showing meaningful experiences relevant for that particular vertical. Then foundation number two is also something we touched upon. Is is the journey contextualize, are we actually taking into account what happened before in the journey? Piller number, foundation number three is personalization. Each consumer is different. They have different preferences. All the results personalized to their preferences. Pillar number four is are we aware of seasonality and and what's trending, which is actually quite crucial. As you see, like, a lot of shopping actually happens around, events and and seasonal. So these are actual triggers of driving the shopping behavior. So our algorithms, our systems actually responding to these triggers which are happening all around us. And the last foundation horizontal, I believe, is the brand consciousness. What that means is if you're a value retailer, then hopefully your algorithms are all working in a way that you are promoting value If you're a up market retailer, then hopefully, you are promoting highly rated products. So we see that these are it it will be crucial for your algorithm to respond with products which are consistent and harmonized with your brand. So these are what I would say are the five pillars and five horizontals of good customer experience. I think I would add a couple just from from Ireland, but there's there's huge overlap here, obviously, because, I think we're both talking very in a very similar similar vein, but from a product discovery point of view and a purchase and decision and payment point of view just to, like, simplify the the journey. Product discovery, we've touched on it a lot already. We need to make sure it's easy for a customer to find the product and the content that they need in order to to discover the catalog properly and and decide whether or not they wanna buy, buy a product. Now that can be multiple different ways. It can be the search it can be looking through the content, but it can also be through elements of guided selling. You know, we have, a large distributor of of truck parts and they have a huge huge catalog, but you need to kind of whittle that catalog down to find the right things for what that person by. And so you have the option to kind of guide that person through by asking essentially a set of questions or creating a set of filters where you can find your way to the products that will meet your needs. And so we need to make sure we're always curating the experience that they would expect and VB said this earlier, it is dependent on the vertical that you're in. People have these kind of expectations. If I'm in this auto world, I expect it to work like this, but then I'm in a more fashion world, I expect to work with this. We need to meet those needs. Then when it comes to purchase and payment, there are things that even if they've had a great search experience, it's been, based on their intent. It's been highly personalized. It's brought up relevant results. Help them find the product. You still need to meet the customer where they are and where their expectations are. They might not want to pay for delivery. They might have thought they're gonna get free, and then when they get to the checkout, it's no longer free. They might not like the returns policy. They might not like the fact that they can click and collect or that they can't click and collect. So there's there's a lot of, things that could still scupper the conversion. I need to make sure that we understand what the customer's expectations are and the e commerce stores meet those expectations. But the last part I would actually add, and I think it's gonna become more and more part of commerce's, point of view and where something they have to deal with is this element of customer service. I I touched on it before, but people will start on the dot com side of dot co uk or wherever you are in the world to start their service experience if something has gone wrong. Now if you turn that service experience around, and you give them great customer service. There's lots of research to prove that you can actually make that person go from or would be a huge detractor into a big advocate of that brand by just giving them a great experience in dealing with their with their issue properly. And I think it's time for e commerce start thinking about what does that feel like for my customer? Because if you maintain your loyalty there, it will make acquiring them again and converting them again dimpler that it would if you had to go and find a brand new customer. So I think that's that's really key, and it's how we get these customers to spend every dollar pound that they have with the brand. Yeah. Absolutely. You know, driving up conversion rates is another way to improve profits You know, what are some company, you know, what are some ways companies can do that? And how can AI powered search play a role here, Sergio? So there's there's lots of angles to how AI powered search helps, drive up conversion and drive up profits. One of the parts is ensuring it's personalized. This experience has to be personalized. And in store experience almost is because that customer chooses what they look at and they could scan the store really quickly and they can figure out what catches their eye and doesn't. You can look at, like, a hundred products in a couple of seconds in store, you can't do that online. So we need to take these little bread crumbs and nuggets that people leave us about what their intent is and personalize. To that is an extremely important part to driving up conversion rates. You do need good foundational search in place and a good product catalog and clean data first. To make this work, don't get me wrong. But if you've got that in place, the next phase is making sure, it's personalized. And The key reason why we need to be even more sensitive to this in e commerce, then maybe we would be in an offline experience is that a competitor is just a tab away. They can literally be, like, a few centimeters in the screen and you click it and they're there. And we we need to be cognizant of that when we're thinking about how we're curating these customer experiences that we that we do for them all the time. Getting in your car and moving to another store is a is a is a bigger ask. So maybe they will persevere and try and find that product. But online, it's way too easy to make the switch so when it comes to personalization and you need to make sure you put the customer at the center of the experience, it should be all about them. And how you can meet and hopefully exceed their expectations, and make their purchase go smoothly so that they come back another time. So if somebody is making that tenth search or even that tenth purchase, again, we've we're kind of hammering this home, but make sure it is personalized, make sure it's based on their intent, and everything that they've done in session. It's super important. It's the expectation of today. That will drive up the conversion rates for those customers. And it it's it's helping them live in that world of their own context. It's very, very important. Then I think is is the other angles that we haven't really spoken about in terms of how AI can help this, it's thinking about how do you make sense of the catalog? How do you understand where these products fit in that catalog based on the other products that you have on offer based on what people view together, but also are these products similar or not? How can we work that out? And you can use things like product vectors, etcetera, to work out how similar these products are. And in what space is someone currently shopping. So that's a really important way that we make sure that, we drive up conversion rates with with it. Type out session? There's actually, a lot of levers. The AI power surge has to improve conversion. And what is conversion really from a search perspective? From a search perspective, it is actually a decision. And what is that decision? That decision is I'm going to put this product into my basket, which is an add to cart. So that is conversion from a search perspective. So some of the commonly known facts about search ecosystems is, sessions which have, which have engagement with auto complete, usually convert better. Also, sessions which have engagement with, filters, also convert better. So if you put like two and two together, then, you could you could sort of make a case that if we drive adoption of the auto complete, which is a search enabler. And also, let's say drive adoption of another search enabler, which is like filters and facets as Sergio was mentioning earlier, then, automatically, conversion will increase. So this is where AI can come in really, really handy is in helping drive adoption. So the more improved, the more efficient, the more smart, these search enablers are, they will actually invite people to engage with them, which in turn will drive it option. So, this is, this is, like, one straightforward way to, to sort of, like, improve conversion using So we need to think about how retailers actually define profit and what their goals are because not everyone is looking to boost profits all the time. Sometimes it will be the case of they want to grow at the cost of profits. So they'll they'll reduce their profits, but they'll gain market share in order to attain that. Or they might have a warehouse full of stock, and that's costing them a significant amount of money. And the e commerce channel can be a great way to alleviate those costs by making sure that those products show up very frequently for customers in the hope that they buy them more frequently as well. But it is about balancing this out. You need to make sure that the customer's needs are met alongside the business needs. And that is possible. You can ensure relevance and good outcomes for for both parties, obviously. But it can also be the case of high relevance and high margin. Go hand in hand, or it can be high relevance, and we really need to get rid of this product, go hand in hand. And so, if we can curate these experiences, we always should, and the the search should be capable of enabling that based on whatever lever it is at that retailer. Wants to wants to pull at the time. Not everyone can do that, but, you know, in this in this day and age, we we hope that they they would. And we've just been discussing elements of how AI can impact AI search can impact commerce. I think we can't really have this conversation, today without talking about GenAI and its application to commerce. Is obviously exploded. ChatGPT has made that world very open for everyone. You can go and try it. You can see what it's like. It seems to blow everyone's mind. It certainly does still blow mine. So Vivo, I wanted to actually ask you, you know, what have you been thinking about and how it could be applied to search? Yeah. Absolutely. And and you're right, Sergio. Like, everyone's head is just exploding with sort of, like, the use cases that can be slide to to this technology. There's actually, if you look at the search stack, you know, the the the technology stack, like each layer of the stack can actually, benefit from using some shape in form of LLM. But purely from an experience standpoint, there's there's many, many use cases which, which come to mind. And I think I can pick three, which I like One is, gifting. Gifting is really, really broken right now on on internet. Or even in general, even offline and in internet combined. Like, it's very, very hard to find something to give someone. And if you really think about it, it's a very emotional. It's a very, it's a very, like, connection oriented experience, which which we still haven't been able to solve. So, so enter LLM. Example, you sort of like, asked this question. Hey, help me find, a fourth great graduation gift for my nephew. Who lives in Seattle and likes skateboards, but already has one. So you already created this really nice prompt and and hopefully, the the LLM model will actually come up with, like, a lot of, like, good suggestions for this. Another use case, which actually comes to mind, is like an enthusiast project. Let's say I'm a avid biker. My bike broke down. Or let's say I'm into, like, fish tanks and I collect fish, at home and my fish tank needs cleaning. So I can actually go ahead, to the retailers website and and I can actually pose a problem. So I don't even know what product I want. Right now, I'm having a crisis. I'm having a problem. So I go about, asking something in a natural language way, how should I clean my fish tank? I need specific details on handling algae, and I don't want to use any chemicals that can hurt my goldfish, for example. So that's sort of like, you know, removes a very sort of like mechanical process of searching keywords to a more humane conversational experience. The the third use case, which comes to mind, is, a little bit like what we touched upon earlier, is about, like, high consider and nuanced purchases, purchases, which people take a long time to make up their mind, product which are complex and complicated and have a lot of features. So it takes a lot of time to, like, understand and make a decision about these products. Example, you could use a a LLM experience, for a high consideration purchase, like Hey, I'm beginning to develop a liking for dirt biking. How should I go about starting to pursue this hobby? Should I get initiated into a mountain should I get initiated into mountain biking first? Can I try a few dirt bikes nearby? Are there dirt biking spot near me. So you frame all of this in like one prompt. And, hopefully, if you if you've done well and and integrated your experience with with the LMM, you can actually, you know, get meaningful responses, for for this use case as well. Yeah. I agree. And I I think this is such an exciting and interesting space. And everyone needs to be careful not to get caught up in the hype of it and try and implement things here and there. It needs to serve a purpose and solve an actual pain point. And it's quite interesting because at Cavea, we're we're thinking about how do we apply this to all these different angles in Phoebe, like you said, you can put it. It can be non customer facing can solve so many things in the stack, and it can also do loads for the customer experience at the same time. Just a couple of things that, you know, I've been thinking about personally are, it's this is great for the customer and actually great for any teams that need to create content, but generation of comparison table when I've used any of these large language models and I've asked them to do some comparisons for me, they are very, very capable at drawing up a table very quickly and pulling in all of the aspects that seem important about those products and showing me the stats so I can compare them. It's quite difficult for teams to go away and build out all those tables and and do them on product pages, but we could use LLMs to to do that and effectively surface elements of the catalog that are really important to to that that customer, that time. There's also things around, you know, we could create content on the fly. You know, if somebody wants to know how to use x, y, and z in this specific way, utilizing the content that's been indexed, you can pull the information from that to then create three hundred words, five hundred words on how to use the that product in a particular way and how that's done in general, which is very, very exciting, and it's effectively personalized and content creation. And I was thinking about how, normally people put one or two search keywords keywords in their search. So How does it apply? But I think sometimes you realize you wish you could have put more in, but you know you can't you know there's no point putting it And example that I was thinking of is, like, I want two monitors for my Macbook. I need it to be compatible. My Macbook is a twenty twenty version. It's got USBC, and I want the screens to be twenty four inches in size at least and two hundred pounds. That'd be a thing to be able to put in sometimes because you just need to find, some products and start whittling that stuff down without having to go into each page and seeing it as USB if it doesn't have the facet and all that kind of thing. It'd be a great way to to do a search, but then also for teams internally. So customer experience makes a massive difference. But for teams internally, like, being able to draw insights from data and just by asking a question effectively to that data within LLM, would be an amazing way to start using it more, not having to wait for analysts, not having to wait for reports to make decisions on data before you move into into the next phase. So super excited about how this space, turns out. And at cabello, we we're excited to move forward with this. It's gonna be game changing. Great points there from both of you. A great discussion going on too. So so knowing all this, you know, how can retailers design search strategies around business goals, driving up profit, being a potential one here, Vifu? Yeah. Love the question. Actually, there's, many, many, livers to pull to drive profits. You know, maybe like to summarize a few It's always good if the retailer understands. What are the what are the hotspots of profitability they have in their catalog. So, we know this that usually in in e commerce, for for most like mass market, like retail categories, the margins are thin. There are certain categories where margins are good, but talking about, like, e commerce in general. Also at the same time, services as opposed to products have very high margin so let me take an example. So let's say you are buying a TV. Flat screen TV is seventy inch TV. It's actually gonna be surprising for most people to learn the margin on, like, a three hundred, four hundred, five hundred dollars TV, they are wafer thin. But here's the kicker. There's actually value added services that the retailer can sell with the TV, which have very high margin. So, for example, TV installation or extended warranties at the time of purchase of that TV. They have very, very high margin. So, some sort of like to strategy to to club these together, can help the retailer. Another example could be you're buying pet food. Again, a very competitive vertical and and margins will be challenged in this vertical. So again, if you pair it up, with, let's say, pet grooming services or dog walking services. So I think the basic idea is identify, your hotspots of profitability and services as opposed to products usually have, significantly margin. So I would say number one, sort of like a very, like, no brainer, simple, to execute, strategy is, shift a little bit toward services, and you will automatically see your contributing margin go up The second, direction to pursue, could be to actually optimize your search ranking algorithm to take into account contribution profit as a data point for each item. Now, this actually sounds like really straightforward and simple. But, based on experience I can say that, calculating contribution profit for a particular item. And, you know, we need to keep in mind that the usual size of catalog at a mass market retailer is like five hundred, six hundred million items. This is a very, very daunting task to actually have the the contributing profit recorded and stored for each product. And in part because, accounting for like cost structure is really, really hard. It's very hard to, get to the final landed cost of a product. So, anyway, like, not, not, without going into, like, the, the, nitty gritty of it, and as difficult as it might be. But if actually a retailer can pull it off, actually they can record in store their their CP data for for each product in their catalog, then they can actually tweak their search algorithm in such a way that it optimizes for a profit. A third way, a third direction again, it's a it's a very straightforward direction to pursue for retailers that that that are actually sitting in this is to go after private labels or in house brands. Now, we all know that private label in house brands, they actually are priced very competitively. They have a good connect with a consumer. And usually, they they have a higher margin. So that's another way to sort of boost these products into surgical development to drive more profit. Yeah. Great examples there, and it makes a lot of sense. And we'll move to our q and a here soon. But before we do, you know, are there any other things you want to touch on or summarize about how AI powered search capabilities can drive up profits, Vivu? Yeah. There's actually, like, we can keep going, on this. There's actually more, techniques and tactics to, drive a profits using high powered church. Just to like, you know, name a few more. Retailers can actually, craft strategies to increase their average order value, which is like another way to say that they should promote more basket building. So the bigger the baskets, higher the chances that profits will crew. Another one which we discussed, briefly ago was more in the context of, products and services. But the the strategy here is to promote bundling, which is, not it it just it just does not need to be product and services. It could also be product and products. So let me take an example. And this might not be like a real example, but just to sort of like drive the point home. Let's say you're selling a queen bed and the margin on green bed is really, really low. But what if you bundle this with a mattress and if the margin on mattress is really high? So your overall net margining is So bundling is a very, very powerful tool. And third, come the third, liver, I would say, which actually comes in a really handy if, especially if the retailer has an omni channel presence, is around optimizing your fulfillment methods. And fulfillment methods can actually have a significant impact on the cost structure, of fulfillment for that order which ultimately affect the bottom lines. For example, the same product, the same t shirt, the the same trouser, yes, the retailer can fulfill it from the fulfillment center, which is like two state lines across from me, which is gonna be very expensive, cost structure for that order. What if the store ten miles away from me actually has the same product And the retailer can somehow entice me to to do the job of fulfillment for them, which means, like, have me go to the store and pick it up. So that will dramatically reduce the cost of fulfillment for that order for the retailer And it also has a benefit for the customer instead of waiting for, like, two days or three days for the order to come in, can actually drive down to the store right now, pick it up and have the, the t shirt or the trouser on me in thirty minutes. So another way to, like, create win win, and then drive profitability. I think to summarize, like, this has been a really a good search algorithm is is is highly complex. And it has a lot of parameters keeps an account. One of the primary goals for most search systems as surgery was also alluding to earlier was that, a search engine should output results, which are highly relevant. And, showing relevant and meaningful results is crucial for building trust and long term relationship and retention. So, relevance just to, like, describe in simple terms is If I'm searching for water, then show me water. If I'm searching for chips, show me chips. If I'm searching for t shirts, show me t shirts. So just, you know, like, understand I'm searching for and show me something which is relevant and meaningful. Now having said that, within the confines and within the guardrails of showing relevant results, there's actually a ton of opportunities to micro optimize around business goals, which I think we touch upon, quite a lot in this, discussion. And what are those micro business goals? So they could be, for example, revenue or conversion or profitability. So I think that's the that's the key sort of crux or distillation here is How can we have a search engine which is relevant? Of course, because if if it is not relevant, then there will not be any retention and people are not going to come back. But at the same time, it's like advancing, the business goals. And for example, like profitability, Now, the slight complexity here which arises is, the overall metrics for the business, which is if you think about, revenue, relevance, profitability, conversion, and etcetera, etcetera, they all need to move in a harmonized direction. So, what this means is if a retailer can actually improve one of these metrics, like, let's say, profitability without actually sabotaging other metrics, then that could be a really, really powerful technique to advance business goals using a empowered church. I think that's such an elegant way to, to summarize a VP. Thank you very much. And your point on sometimes, yes, I just wanna go to the store and pick it up myself. I know I can be there in ten minutes and I don't wanna wait until tomorrow is how I think a lot of customers feel today as well. That that actually resonated with me too. You know, just from from my perspective, you know, we've covered a a lot today and it's been a really great discussion. And I think that there's a lot of ways you can deploy certain search strategies based on business goals, but you need to meet the them to where they are. And so showcasing the right content, the right product, in the right moment, you can't do that through a load of manual rule if you start to build up that sort of approach where you've got, you know, hundreds and we've had examples of people having in the thousands of rules, it's just like a house of cards that could fall down. You need to make sure you're scaling these efforts with AI. And that's not only to help you with the scale, but it's also about freeing the people that you have up to do the creative tasks that humans are really good at, thinking about new ways and new strategies they could deploy instead of, you know, they're trying to work out why this rule is conflicting with this rule and how they can override it with maybe another one is not a good use of of of human time. So that's a really important, point I just wanted to make to to finish. We have to scale with AI and, you know, gen AI being a huge part of that now as we as we move forward. Alright. Great points from both of you. And thank you so much. And we've just got a few minutes left here. So I wanna jump right into some questions from our audience. And our first question here is what are your thoughts on lowering customer acquisition costs And can AI powered search help with that? Who'd like to take that question on or start it off for us? I can have a go. Depending on the acquisition strategy that the retailer has, so that, you know, a little bit of a caveat there. One way in which again, if I just, like, bring it up to product discovery again, rather than just AI powered search, I think I can answer that question better. If people are coming in through paid product list thing. Example is you've got your product listed on Google, so when someone searches for it, they click on it and they land on your website. Quite a lot of the time, it can be the case they land there. And if they're looking for apparel, those size they want is not in So the immediate reaction of that customer is to bounce back, go to Google and find, you know, someone else from that list. One way in which you can help alleviate that. And, obviously, that's just the cost. You just that's an acquisition cost. You paid for it. No. We're gonna end in One way you can try and get around that is to use all of these amazing product discovery technologies that we have around search, the understanding of the catalog recommendations, except to showcase very similar products that are in the size that they want. It's that last attempt to try and get them to stay on-site and make, that per is rather than losing them because then that acquisition cost is it's just mounting up. So that's just one example I can think of. Yeah. It's a great example. And maybe I can add, like, one more thing here is, there's something to be said about, product growth. And there's actually a lot of, like, think tanks and and discussions and forums and credits and whatnot on on PLG on the internet. It's actually an really effective strategy to lower acquisition costs where the the crux of it is the product itself is is so amazing, and so delightful. I think the the main thing is the word delight. If you can actually create a moment of delight for someone, then as much as you even want to stop that person from talking about you, they will actually not stop talking about you. So, having that sort of product led growth mindset and building those moments of delight into your product will actually kick off the the referral and acquisition and word-of-mouth Flywheel, which will have a dramatic effect on your customer acquisition costs. Great response from both of you and love that example from somebody who's a petite extra small. I can tell you, and I I I I can't stand being shown a bunch of things that are not in my size after I just put it in there. So great, great example. Okay. I really like this next question. So I wanna get this one in before our time ends And that's should businesses focus on recommending less expensive items right now since consumers are fighting high prices. I really like this one. What do you guys think here? I I think this is this is a, you know, a bit of, obviously, a hot topic. Again, you know, in this research that we've done recently is the last, you know, a couple of months old, we are seeing a shift to you know, private label, own label products because of the rising costs. So there's there's no doubt about that. There is that shift. I think this probably comes back to our message all along, which is understand the intent and the and understand that customer as best as possible. If they are browsing a lot of high priced products. I think you can and you can take in lots of other signals as well. You can make a pretty good assumption that I can continue to recommend higher priced products to these cost to this customer. If they are not and if they're maybe using filters and facets to to lower the price or sorting low to high, on the search results page, I think you can infer pretty quickly that it would make sense to recommend the price products in that situation, but that it's not gonna be one size fits all. It's gonna be dependent on that customer's need during that session, and we have to into account. Yeah. Absolutely. Plus one to that, but there's no one size fits all over here. Should should the retailer focus on, value, products only. I think it's a it's a very deep question. And especially like something we think about a lot at Walmart, my current company. And I think it really depends on on how you tackle this really depends on how you understand yourself as a company. So if if, you sort of, like, relate as a value brand or or or a or a brand which, you know, people who come to for everyday low prices then, absolutely, then you have to be mindful of what is happening around you. And, you know, customers are really looking up to you not just as, like, a business or enterprise, but at at some at some level, like, as a part of your community, to sort of, like, come and support you. So, in that sense, it becomes really, really, crucial for the retailer to meet the customer where they want to, meet them, and sort of like double down on on providing value. We actually ran you know, some of some of us might recall, Walmart actually ran this campaign, last Thanksgiving. Where in general, the, the food prices, all over the board were, like, thirty, forty percent high. Like inflation was really, really high in December, November, December. And, ad campaign, which, which was run by Walmart, was actually the entire grocery, Thanksgiving shopping list, for, for last year compared to the year before, was actually exactly the same. So it, I mean, it, it is, it's kind of like unbelievable, how it was pulled off. But I think the main point over here is, understanding who you are as a retailer, what is your soul as a retailer, is your brand as a retailer? And, and, you know, doing everything in sort of like cognizance of that brand. If you are a sort of like a up market, a more discretionary, sort of retailer, like, for example, like food delivery apps, right? I think we can sort of, like, argue that, most value conscious people are probably not, like, you know, going crazy over like food delivery apps. So then maybe it's okay. It's okay to sort of pass on the the inflationary pressure to the consumer. They're looking for convenience, and they're willing to pay for convenience. So it's okay to pass on. If you are a high end luxury, fashion brand, then also it's okay. You can sort of like, you know, your your niche is not necessarily value. So I think it all really depends. And again, like, a really good point that Sergio called out is the same customer might actually have different price and value expectations from different verticals. They might expect something very, very aggressive on or very, very conservative on food and consumables. Because it's kind of like their life depends on it. But they might be okay, for some sort of, like, you know, higher pricing on more discretionary verticals like to say toys or or or, you know, auto, for example, or or bikes, for example. So I think it it all depends on, like, the the different sort of verticals as well. Like for the same consumer, the patents can actually, change. But I think, the main thing to keep in mind is to respond back, with something which is true to your brand and which, deepens your relationship with the customer and takes it forward. Great points there. Audience, I told you we had two experts here. I know I have gained so many valuable insights today. So Vivhue and Sergio, thank you for just a great discussion today. You know, we appreciate you sharing your insights and expertise with us. And I'd also like to thank our audience for joining us today. Today's webinar will be available on demand. You will soon receive an email with a link so you can rewatch or share with your colleagues. We hope to see you again soon. Have a great day.