WTF: What's the Future of Digital Transformation - Search in eCommerce
This is the Dynamicweb Livestream of WTF: What is the future of Digital Transformation.
In this series of Livestreams, we will introduce how you can approach digital transformation.
This Livestream covers search in eCommerce.
We have had the pleasure of interviewing Mark Floisand, SVP Product and Industry Marketing at Coveo. Mark is a 20-year experienced veteran in the IT industry, and we ask him to share his insights on eCommerce search. We cover topics like:
- The impact of search on shopping cart performance
- How data can serve a better search experience
- Mark’s recommendation on when you should consider re-platforming
Eric: Hello, and welcome to this episode of dynamicweb WTF. What's the future of digital transformation. In today's episode, we talk about search in e-commerce and we have Mark Floisand, SVP product and industry marketing at Coveo. Mark is a 20 year experienced veteran in the IT industry and he and I covered the following topics, the impact of search on shopping cart performance, search, and recommendations.
How data can serve a better search experience and Mark's recommendations on when you are considering replatforming, what you need to take away from that. And several more topics, I hope you enjoy today's live stream, and that you can take elements away from it and implement it in your own situation. And with that, I would like to welcome Mark.
Eric: Welcome Mark.
Mark: Hey, Eric. Thanks for having me. It's a pleasure to be here.
Eric: Very good. I very much look forward to this one. I know that we had some pre talks before, and I very excited about the insights that you can give because in e-commerce and websites, implementations, a site search usually gets bossed over as one of those things we will get to when we get to it.
And it doesn't seem to have the same. Priority as personalized triggered emails, shopping cart, abandonment programs, and all that, but taking steps to improve your site search can improve conversion rates and lower shopping cart abandonment, which leads to happier shoppers. So before we really dive into that, Mark, can you share a little bit more on what Coveo is and the role and experience you have?
Mark: Sure. Absolutely. I've been with Coveo about four and a half years. Coveo, is about 15 years old as a company and focuses very much on the whole area of more relevant experiences. So what does that mean? More relevant results? When people look for something that they're trying to find more relevant, product recommendations that help them also add to the value of their cart and frankly, more relevant content that assists.
Then in the process of making decisions, whether those are shoppers, whether those are customers or even employees within an organization. So our business is one of understanding that kind of context of what people are doing. And using AI and machine learning to recommend the most relevant information for them at the point in time that they're at.
So from a commerce point of view, you can specifically think of things like how does one improve the commerce search on a, merchant store so that when I'm looking for a particular thing, the results I find are relevant to me, they're meaningful. They're actually what I want to do because I'm not there to search on there to buy.
And that's fundamentally what we enable and then expanding their kind of frame of reference by giving them the sort of almost Amazon like recommendations that broaden their choices, people like you that saw this may also think about that. And because you bought this, you may want to consider that so improving the average order value as well.
Eric: Cool. Yeah. And that plays very well into what Gardner and many of the analysts are saying as well, and at dynamicweb, what we try to, help with as much as possible that in any e-commerce environment, that when a return visitor is on the website, try to showcase on the homepage, we're ready, the most fitting products.
And I think with search, you can actually take that to the next level as well. And if we talk about improving the site search results, it's way more than just returning results for searching for it right there. There's a lot more depth to it.
Mark: It is. And I think that people underestimate the power of search and the value of search on a commerce site.
We have customers that having deployed search, come back to us and say, Do you know, half our navigation starts in our search box. And the conversions that they see from those are an order of magnitude higher because. People are coming to a particular retailer or, in fact B2B distributor with something in mind.
And so the best path for them is to enable them to get what it is they're looking for and why make that difficult, why they put barriers in the way that's slowed down. Their ability to ultimately transact. We've all grown up as people living on the web now with 20 years of Google. And so we have an expectation that searches instant it's relevant.
It's at your fingertips. We all know how to do it. It's intuitive. And so when someone arrives at an eCommerce store, whether it's B2C or B2B, and that's that on-site, or that on store search, doesn't deliver the same kind of expectation that we take for granted. In Google, in YouTube, in Amazon, we leave.
And so as a climber navigation vehicle, as a primary mechanism for, servicing people in the moment, it's really important that it's up up to the kind of expectation level that modern consumers and buyers have. And so a large part of how to improve it is not just the. What I would call the, I, it's almost that kind of cliched whack-a-mole of where you have merchandisers spending a lot of time and energy constantly chasing their tail, trying to put new merchandising rules in and if then rules and thesauruses sources, because they've seen someone looking for something you just cannot scale like that.
And nor can you again, deliver the experiences that people have come to take for granted. With AI driven machine learning capabilities. It's no secret that Amazon and Google and these companies have been investing literally billions of dollars in data scientists in scalable cloud platforms in algorithmic intelligence to drive automated.
Recommendations automated responses. And really our part in this from a coveo standpoint is to provide that same capability to other companies so that they don't need to be an Amazon to provide the meaningful, relevant recommendations that someone's looking for when they simply go to a website and search a lot of, it is also recognizing that many companies, frankly, the vast majority of retailers are never.
Going to get the volume of foot traffic, digital foot traffic to their store, that someone like an Amazon does. Amazon knows more about you and I than we probably ever will. And certainly more than any other retailer ever will, because the reality of it is that for most online retailers the, math is that 70% of people that show up do not create an account or register.
They shop anonymously and they may be there two tops, three times a year. So the reality for most retailers. Is that they're not gathering like screeds of personal information to deliver these highly intuitive, personalized experiences. There isn't the, volume of information for them to do that. There isn't the frequency of shopping to do that.
However, what they do have is a lot of customers who individually may have gone and bought a pair of shoes and a pair of trainers and a pair of ski boots and collectively. Collectively across the retailers catalog. And in this example, I'm using shoes as an example, but across their product catalog, they've actually got a lot of data that they've been gathering about the kinds of things people buy.
And so one of the things that we advocate and really this is. Very fundamental for solving. What's often called the cold start problem of how do I get this machine learning to make these recommendations? When I haven't got all this data? Guess what? You actually have a lot of data already in the web analytics and the product interaction that people have had.
And so we have models that start to be able to make recommendations to anonymous first time visitors, because we see patterns of behavior. That other people have exhibited when interacting with product catalogs, you can almost think of your product catalog as like a. Almost a graph of the characteristics of product, a white shirt for a man with long sleeves has a certain set of characteristics that might be quite similar to another product.
And so just by being able to analyze the data of which products people have interacted with without even knowing who they are is still the basis for beginning to personalize and deliver more relevant and meaningful results. Even to people who are occasional shoppers or anonymous shoppers. So I guess my point is that site search has come a very, long way from the days when it was just about keywords.
That's the 20 year old capability. What would you,
Eric: I like that you go into that because also with the dynamicweb product information management solution you, get all that product data. But in our case, because we are that platform builds from the ground up, we can do to, to what you're. Referring to is get our web traffic, behavioral data interactions, and combine that and use the product data element as well to then serve the best possible products on, whatever page is probably ready for.
Going to interact with. So I think that's actually quite close to what you were just sharing.
Exactly. And so the whole point is that every retailer and every distributor has actually got a lot of meaningful data that they should be tapping. And so search in an of itself has very often been seen historically is this silo, the one does actually, you said Eric, sometimes it's an afterthought and I'm here to say one, it shouldn't be an afterthought because it's a primary navigation mechanism. And two, it doesn't have to be your granddad's search from way back when, where everyone did manual keywords.
And then there are machine learning models. Now under the hood that automatically do things like. Yes. We said, understanding product interaction things like automated query suggestions, the ability for, example, when you and I go to Google and we start typing, those suggestions are being put to us.
And guess what? They're spot on, right? No, one's doing those by hand. There aren't a team of ELLs in the background. Like we're going to write in them all out. It is all algorithmically driven and, this capability is not unique to an Amazon or a Google. Yeah. The Coveo relevance platform does exactly this, which is to provide these machine learning capabilities, to automate those kinds of things, because that's scalable.
It doesn't require manual maintenance. It grows like crazy and continues to evolve and adapt based on. Actual user input.
Yeah, absolutely. And I think that is also very close to what you're referring to is the ability to take all that kind of data. And then basically not, so much stare, but trying to give the best possible search results back in, in with the goal to have the person find exactly what they're looking for.
Exactly. Do you maybe see in search in combination with e-commerce? Do you see significant differences between B2C and B2B?
Mark: It's a good question. Yes and no. And that's and I'll that's like a non-answer I'll give you I'll do the both sides. So, where there's commonality is the fact that.
Yeah. Enterprises, any frankly, any business that's looking to acquire products from a distributor or a manufacturer directly is not doing so in a, as some abstract entity behind that business as a buyer. And that buyer has a pulse and heartbeat and flesh and blood, I E people. So what's common between BMC and B2B is ultimately, the fact is we are dealing with people and those people have in and of themselves brought to the workplace, the expectations from their personal lives just because I'm a business buyer looking to buy, I dunno, plumbing parts.
Yeah, it doesn't take away from the fact that I've got 20 years of experience living with Google and Amazon and Facebook. And I bring those consumer expectations to my job. And so, from a B2B point of view, the pressure is on frankly, to deliver great experiences that very much parallel and are on a par with the B to C consumer experiences, which have been much more focused on brand experiences, much more focused on consumer orientated, digital commerce, generally a further ahead, but the B2B sites are having to catch up because that's how they're going to survive.
What, where I think the differences come down to are things like The nature of the company is that very often people do business with, in a B2B environment. It's less of the, as I said, the occasional things that we might see in the shoe store analogy that I gave earlier much more. One of what are my repeat purchase orders.
Are there automatic replenishment rules? Is this what, are the pricing permissions? What are the product entitlements that might be specific? To a particular business contractual arrangement. And so very often, if I'm a distributor or a manufacturer, I'm going to have all sorts of different pricing, entitlements based on different tiers of channel.
Maybe I've got distributors and top retailers and independent mom and pop shops, and they've all got different discount levels and you want to make sure that those prices are dynamically shown, honoring those contracts. You've got things like. It might be more similarity with the retail here is things like being able to do lookups for product availability.
The difference is that product availability may be governed by contractual relationships that require you to always have available a certain lead time or a an automatic trigger to replenish for business to business. So there's a lot of business rules that perhaps aren't necessarily, or deemed required in the B to C world that are essential.
In the B2B world. And so particularly you think of things like the the, price tiering, the entitlement, and how do you make sure that search on those business rules? Because if I'm a, I don't know, and mid-sized engineering shop and I buy all of my my car, my truck replenishment parts from a distributor who does those.
I want to see my prices. I certainly don't want to see retail because I'm not selling it retail. And I also, it would pain me to see prices that are not entitled to, because that would just annoy me. If there's somebody that's buying it cheaper than me. So things like pricing, integrity availability the ability to also tap into things like product information management system, to understand the rules that go with that from a recommendation point of view, because I've ordered a particular component there's also for example a replacement lubrication kit that needs to go with it.
How do we automatically cross reference and cross sell those as well? So those are the kinds of differences, but centered on the commonality is the fact that these are people. Two
And I can completely agree with that one. And I could see that as well with if your e-commerce system is able to recognize the visitor either.
Implicitly or explicitly that you do want to be able in a B2B environment to do customer specific pricing based on ERP integrations other available data sources ranging from CRM to ERP, basically. And I could also imagine that adds a fun complexity. To the search element as well, because when I'm in a B2B environment, trying to two, and the system is able to recognize me, or I actually locked in to my own customer portal, I should be able to still use the search functionality.
But as you said, it should not start returning other customer deals pricing deals and all that, because that will be indeed a pain. But I could imagine from a search perspective that's a fun complexity to actually have your ERP data connect with your e-commerce system, connecting with your search system to then actually combine still all the right data.
And those were doing the wrong stuff.
Mark: I've had, I'll give you a worked example of where this lives in practice. One of our customers coming to court keep pride, and this is exactly what they do. They have a network of direct customers. So feed pride in the, trucks component, spare parts business.
So they have customers who are entire fleets of, long haul. Trucks, they have independent truck repair centers and, indeed they have distributors who in turn resell their components in regional bases and whatever. So this complexity is, the world they live within.
So UN honoring those rules honoring and understanding where the, components are. If there's a trucker, like out on the highway and. He or she who's driving this thing needs to pull over because there's some component that's gone wrong. How do they source the right compatible thing at the price that their feet is entitled to?
And where is it? Yeah, I E if I'm in the middle of Nebraska and this thing's broken down, is there a part that I can get my hands on to get my haulage moving again? So these are the realities that they deal with and Yeah. So, that tight integration is really, important. I think one of the other things to bear in mind is that as we, think particularly about, about B2B there is this need to be able to take advantage of, as you say, data from other systems that makes the experience more relevant.
I'll give you an example. One of our customers are Salesforce. So Salesforce, of course we know them, everyone knows them. And the app exchange is the place that you as a Salesforce user go. If you're a Salesforce administrator to add more components and capabilities to your Salesforce instance, and there's, I dunno, 600,000 I've lost count, maybe a million different apps up there now.
And. Of course the sheer variety of Salesforce customers and the number of different things that they're doing, the different industries that they live within. It's complex. What we able are able to do under the hood is effectively say, when you go to the app exchange and now you've logged in, because you're a an authenticated Salesforce user, we know what apps you've got installed.
And we know that if you've got these four apps, there are people like you that have got those four and this fifth one, that's correct. Based on observations of what people like you need and have found useful. And again, this stuff is all. Dynamically created. It is all using different data in this case data about an installed base and being able to put that to work, to make meaningful recommendations to those customers.
So those are the kinds of things where you start tapping into kind of non-intuitive pockets of data that can prove the relevance of the recommendations that have been made to the customer in a very kind of honest Netflix like fashion Here's what you might find useful because we've seen you do this and you do that.
And the data says that you're probably going to find this useful. Let's read, let's predict that and recommend it.
Eric: Yep. Makes total sense. Absolutely. And I think that it's also based on the webinar we've done a little while ago with one of our implementing partners on, search in e-commerce and direct results in conversion rates.
I can imagine that you have some. Implementation examples there as well.
Mark: We do I hope we can do a year. One of the things we find is that we tend to deal with very large accounts. So Coveo was very much an enterprise level organization. We focus on companies a billion dollars North and and the results that they get, they tend to not want us to talk about.
So I'm not going to give them names or numbers, but I will tell you that we have, for example a, leading global electronics retailer who deployed Coveo on their e-commerce store and saw a 7% uplift in conversion now. At the scale they're running there. That is,
Eric: yeah. That makes a lot of input.
Mark: And, I think it's testimony to the fact that not only does this technology work but that it scales because of course every every year I'm based in the U S every year in the U S there is the seasonal holiday, a holiday shopping process that starts. And November around black Friday and pretty much runs all the way through to the, Christmas holiday.
And there's a big spikes in shopping. And so a large part of what the the retailers themselves want to make sure is that. Nothing has introduced the war cause that black Friday shopping boosts to break. So we'll put us through rigorous testing, lots of like heavy, load testing. And of course, because it's a multi-tenant cloud platform, it just scales with no problem.
So these are the kinds of things that when you think architecturally about how do companies deploy search? We've seen a lot of people who are using maybe technology that. Doesn't scale like this and generates these kinds of problems because you can't count for planning for that capacity, but when you move to a multi-tenant cloud platform, all of this stuff is literally on demand and enables that let's call it trust in knowing that those conversion rates are going to hold true over the most important shopping periods of the year.
Eric: Absolutely. That's, what we see as well with the cloud solution that we have as well. That it's so easy these days, not only to get up and running, but also to scale whenever it's really necessary. It's still good of course, to try and smoke test and, give it some, testing rounds shortly around the November days of, shopping.
But I certainly hear you that. Yeah, the cloud has brought a lot of benefits when it's also true slash the right cloud environment, because there are some differences to be seen there as well. So maybe also as, I think you really gave some, really cool examples already may be technology agnostic.
Would you already have certain best practices that you would recommend any e-commerce environment that is really going to focus on equal search?
Mark: Yeah, I think a couple of things I would point to first and foremost, don't do this last we talked to them during the conversation about the fact that very often search is seen as a phase two thing.
It should not be, it should be right. Bang in the middle at the beginning of the project and designed in from the beginning, because it's such an important part. And I'll elaborate on that a little bit more as well from a design standpoint, because. There are there, there are things to take into account right at the beginning.
And the other piece of it is that we talked about scale. We talked about the need to make sure that you're using a platform that can scale like crazy. I think that's really important because as we said, if we've got people who are seeing half their traffic come through search, and if that breaks.
And half your shoppers go away because of a scaling problem or a lack of ability to deliver rapid results in real time you just, you're just pushing customers out the door. So making sure the right underlying architecture is there to support it. And that starts to speak to yeah, recognizing that we're seeing more and more organisations think about their architecture, not as one huge monolith. But utilizing elements that make sense for the right engagements. As you say, worry, there's maybe there's product information management, there's e-commerce checkout capability. There may be a mirror content management components, and we're starting to see a little bit of mix and match where people are saying.
Let me bring in almost a best of breed component, that supplements what I have fits within what I have but takes away perhaps a dependency or improves the way that this particular function might be being delivered. And that's very much a direction searches going in. We're seeing a lot of companies say, do I really need to build this again from scratch?
Mark: If I may let me go back to the point about the design thing though.
And the reason I bring it up is that you're very often, we think about search as the results we think about it, maybe as the recommendations that might be adjacent to them. And, but it's deeper than that. What, a modern kind of relevance and search platform does is not just deliver the right content or the right product recommendations, but also take account of what is going to help improve the actual buying experience itself for the shopper.
Let me be explicit. So for example, at Covera, we have a feature. We called a dynamic navigation. And what that does is if you think of your search results, you go to a store again, let's go to the shoe scenario. You ha you looking for a pair of shoes and you'll get a whole set of results. And probably with a set of facets down the side, that you can filter and drill into like shoe size and color and men or women or whatever the sort of navigational facets.
Those things are invariably hard-coded built by hand and the largest stores on the planet will put a lot of development, energy into actually writing and tailoring those navigational things. They're all done to date. They're invariably wrong. The content very often doesn't match the facets that you picked.
I'll give you an anecdote. My wife and I recently re reorganized our lounge and we've had this we've had this mat in there for the longest time. It's like it's rectangular and it just doesn't fit the room anymore. So my wife says I'm going to go buy a square rug, so fine. We went to a website that will remain nameless and of course look for rugs and they're on the facets.
It's got sizes and you take square and we get all these results, including one that was eight foot by five foot. You need junior school math to know that's not a square.
Eric: So that one's always amazing.
Mark: So there's, this broken experience that crops up time and time again, this disconnect between what the shop is trying to buy and the navigational experience that frankly is just getting in the way and delivering the wrong stuff.
So dynamic navigation is about changing those things. Only putting up facets that are relevant and taking away any of the distraction that may actually slow that shopper down. You don't want to do this by hand. You don't want to do it using rules. You want to take advantage of machine learning that sees the kind of things that people click on.
I’ll give you a very real example. One of our customers is in the consumer electronics business. And so for example, if you're shopping for something like a monitor we see in the sort of search terms that people look for, that there's a lot of commonality with things like 4k monitors. I'm looking for a 4k monitor.
I'm expressing intent of buying a monitor with a resolution that is 4k capable. So, there's a whole lot of facets that suddenly become irrelevant in terms of things like screen size and so on. So let's push those down because I've already expressed my intent. In that very simple search term.
And so it's using machine learning to start gaining this intent from the data and putting that into automated practice to change, not just the results, not just the recommendations, but the entire shopping experience to be tailored, to make it easier for that buyer to buy.
Eric: Yeah, fully hear you on that.
And I think that is the kind of I would call it, personalization that, to bring it back into some of the content elements of it, that where the strength is. Because as you see here, if you are searching for 4k monitors, you're probably not going to find that at a 19 inch. If those things still exist monitor.
So yeah, push those things out completely. And then start with the most likely, and now there's maybe a little bit of a double goal here, right? Because you can have to go off the after manufacturer. Of the seller, but you can also have to go off course of the individual shopping for it. And there's a balance between those well
Mark: let's, drill into that.
That's a really important point because we think about personalization almost always a reflex reaction is do right by the consumer. The best experience, how do we delight them? And if you think of it I'm looking at my desk here. So Hey running running GPS, right? So I'm shopping for a new running GPS cause I've had this one eight years and I need another one best possible experience I could have would be going to a a sports equipment site that happens to know me because of my prior purchases knows that I own this thing would make a recommendation and the best experience could be.
It's on its way to you and you know what Mark, it's free of charge. That'd be awesome. That's probably the short of pinning money in the bag as well. Just to, that's probably the best possible experience I could have that company would go out of business. Because that doesn't make economic sense for them.
Exactly. So it's this balance personalization doesn't have to be at all costs you've got to do right by your shopper, but it's okay to do right by your own business. And so let's play it out. Let's say that I would be happy with two very similar brands and one the retailer next 50% margin on.
They only make 20% margin on, but the retail price is about the same perfectly. Okay. For that retailers recommend the one, they make more money on first. And then secondly, the one that they only make 20 foot points on it, it's their business as well within their role. And if I'm indifferent between those two, they both really meet my needs from a personalization point of view.
Why not optimize for the business as well. And so what's really important within this search capability in a modern search platform is not just to understand the products and the shopper's intent, but to also be able to bring in business rules and data that affect things like what's the order you might want to list products in based on factors like product margin.
Based on factors like return rates, product quality. If, again, let's paint a picture here. Let's say it's Thursday evening. A promotion is put up, live to go live Friday morning to sell some more of these things at 50% discount. Great. And they fly out the door and we have this wonderful promotion, but literally when they're arriving on Monday, Tuesday and the showing up in the mail or showing up in the career half of them don't work.
Yeah. And what's happening on the backend is that people are getting hold of the customer service people saying, I need to send this thing back. It's garbage. If that promotions running at the front end, selling lots and lots of these at half price, and half of them are coming back in and the other end because they don't actually work.
And there's no round trip within the business that says we have a problem here. Stop merchandising. Yeah, those rules need to connect up because siloed businesses can self-destruct when they don't take account of these things. I'm old enough to remember. I grew up in the UK, in South Africa and the UK and in the UK.
At one point there was a, promotion by Hoover. Yeah, I can clean up people if you remember it. But this promotion was basically by a Hoover for, the Christmas period. And you get two flights return flights to New York from London. Guess what? People bought a Hoover because they wanted the two flags from one and there was no time.
There was no limit on this thing. It made bankrupted Hoover. Because they were on the hook for all these double a double flights to the new to New York. People didn't even want to Hoover. It was just
Eric: but tickets. Yeah. Makes total sense.
Mark: Exactly. So, my point is that these things need to be born in mind.
When we think about search, it's not just search box type in some results. It is, are those results relevant, are the recommendations around them relevant to improve the average order value, and indeed is the experience being tailored, irrelevant to make it easy for that person to buy. And those things need to be driven, factoring in things like other business parameters to affect the merchandising itself.
Eric: Yep. No I fully, hear you on that, that it is based on all the available data that you have to try and accommodate the visitor to shopper as much as possible while also not losing your own goals out of sight and the ability to, to, sterile that. Yep. Fully hear you on that one. Yeah. So maybe also to, ask you a question on with your experience on marketing technology and being a veteran in the industry when a person or organization is ready to replatform it, feel free to take that from a search perspective, take it from a website online environment perspective.
What's ideally from an online e-commerce of course, what would be your biggest recommendations?
Mark: I would look for the ability to be agile, to harness data in automated ways and trying to eliminate as much manual processing as possible because otherwise you can't compete because that's what the, kind of tech giants have already been doing for the last decade.
Mark: that's running at scale and using AI and machine learning to predict what people actually want to buy. So if you're not doing that, you're just falling further and further behind.
Eric: I think that is a very fitting point. And I think that indeed that when replatforming, it is always in my opinion about.
Think about scaling, think where you are today, but also build out where you want to be in a couple of years from now eliminate manual work, as much as you can and get all the insights, all the data to your disposal. While first thinking of that data strategy before starting to build, because it's one of my beliefs that I do ventilate regularly, is that integration has become.
Extremely easy these days. I would even say sometimes too easy and because it became so easy, everything is slammed on, top of each other. And you still end up with a data jumbo. Maybe you have some thoughts on that one as well.
Mark: It's true. One of the things I mentioned earlier, we tend to focus on larger organizations who by almost by definition have got.
Lots of disparate systems, lots of messy data. And one of the challenges, many organizations invariably run into is do we extract, transfer and load? Do we ETL all this stuff from one place to another? Do we try and get a perfect homogenous data set? And our answer to that is no. Leave it where it is.
People talk about a system of record and, or system of information. And, to some degree I, think of it more as an ecosystem where there's just there's stuff in all sorts of different places. Yeah. Because of our search roots we started in search and we actually started an enterprise search way back in the day.
We have a deep understanding of how to connect to all of these disparate systems, these weird databases whether they're modern cloud hosted data stacks or some legacy thing that's was hand-built by a guy 15 years ago. No one touches it because no one wants to break it. That our CTO Laurent has has a saying, which is, we'd never met a content source we couldn't index. And he's basically right. So how do you take advantage of the information that's in their systems to unify it all together and, do two things unified the content so that it's discoverable, whether that's product information in a pin, whether that's pricing information, whether that's catalog information, whether that's content related to those products, like videos and fact sheets. How do you index that content to put it all to work where soever it lives? How do you unify and understand the interaction data that people have had with all of those different elements of content, whether that's through the store or a site or portal or whatever, because that's the, raw material, the data raw material, which AI uses to learn what people found useful to analyze those patterns, to discover the content that influenced the sale and therefore be able to recommend it, knowing that you will drive an outcome.
Eric: Cool. Very cool. Mark, I think that I don't have any further questions. Would there be anything else you would like to add?
Mark: Sure. I think one of the things that I would encourage your, listeners and your viewers to think about when they go through these things is to to, look at how to harness the best capabilities out on the market. And, not be let's call it. Let me be blunt. Not, Just be discerning about who you do business with, because there's an awful lot of vendors out there that perhaps are saying, Oh yeah, tick, we do that.
But when you actually peel it back, they might do something very shallow. We see, for example, on things like, as you mentioned, platform a re-platforming exercise exercises that search might be a checkbox that is to grossly underestimate. The requirements that your company probably has, and you need to double in drill in and expand that out to understand what truly responsive search is to understand how best to deliver automated recommendations and to change fundamentally the experience of those results to make them work.
It's not just about popping up the search box with an index behind it. Those days are gone.
Eric: I fully agree on that. I think that search is the numbers off of performance already support that search is such an integrated in yeah. Integrated part of any e-commerce environment. There are numbers out there when you have even an unsuccessful search happening, you already get a higher conversion than those search at all.
Which always amazes me. But the numbers, even double incursion rates when the search is successful. So I think that emphasizes how important it is to have a smart search functionality that is fed all the necessary data to, to do its job really. And I also think that just having a standard that indexed shirts functionality does not always cover it anymore.
And. The ability to really have that. I think it actually isn't has benefits offsite, because you're able to take that in syndication offers and even other channels to just make all that data available. So I think with that, yeah, great point as well.
Mark: Good. Thank you very much. Really appreciate it, Eric. This was a lot of
It was absolutely long and fun. Thank you, Mark. Thank you for sharing your experience and insights on this. I think we've had several key takeaways from a have all the necessary data feed into your e-commerce and sorry, the search functionality. It will drive a better performance when done.
And it is about offering the goals that the user is looking for, but also a little bit, of course, your own goals there as well. So as well, thank you very much. Thank you for tuning in, and I wish everyone a great day and a half, two very successful e-commerce solutions.