See All Videos
DISCOVERY DEMO

5 Ways To Bring B2B Up To Speed With Optimal Site Search

Watch now to learn how prioritizing e-commerce search can help advance your distribution strategy, boost efficiency, and increase revenue


Watch Now

Take a B2B-specific deep dive into our Discovery search solution

We show you how to utilize your data to easily create scalable and revenue-driving experiences. You will see specific use cases that are available for you right from the start of using our Discovery platform.

Hear real examples of like-companies that have driven significant results

We will share specific use cases that have driven success for customers like HD Supply, Sonepar, and Global Industrial!

Q&A with our B2B expert solutions consultant

Tom and Jason answer questions within the session to ensure you are equipped with everything you need to know in order to drive real growth within your organization!

Jason Hein: "Hello, everyone. Welcome to the conversation. We are here today to talk about five ways to bring B2B up to speed with optimal site search. Today is going to be a fun conversation. I have… My good friends here, Tom and Ashley from Bloomreach and what we're gonna do today is we're gonna try to address one of the core issues that B2B has to face, which is to, if you can advance, why is digital so darn hard for B2B? And, you know, ultimately we hear this a lot because eCommerce has come so easily. It seems to be consumer goods, whether you're looking for books on Amazon, you're looking for clothes, it's all very simple. And the logic is that B2B should be able to do that too. But the reality is that in B2B, you've got three big differences from consumer goods. One, you've got very different kinds of products that are created for very different kinds of customers in an industry that has very different problems than B2C does. And the one solution that everybody talks about is like, this is gonna solve all of our problems… AI, yes. But what exactly is AI and how does it actually solve these problems? Because ultimately, what AI, the role of AI when it comes to B digital is, it's a tool, it's a platform. It's not some sort of technological miracle. It's not some sort of cereal but the right technology when used in the right way can actually solve a lot of those problems around those differences. And so today, what we're gonna do is we're gonna show you how the Bloomreach technology is used. A, we're gonna go behind the scenes a little bit to show how it helps customers find complex products for a diverse range of customers, regardless of industry geography, you name it, and to solve this massive set of applications. So who is here today to do this? My name is Jason. Hi, I'm the principal visionary for B2B here at Bloomreach. I've got 25 years of experience working in the industrial distribution, B2B digital, Bob strategy space. And I'm joined by my good friend and I. Tom was, who actually knows what he's talking about? Hello?"

Tom Washek: "I'm essentially Jason's technical counterpart. And I've been searching around for many years and I've seen a thing or two. So I will be here to help Jason tell a really good story today about how Bloomreach search can help you solve some of your B2B search challenges."

Jason Hein: "So before we jump in, I wanna ask a real quick question to everybody. We're gonna open a quick poll. So the poll, well it's really about what is your interaction? What is your relationship with your site search? Now? Are you hearing that it's working? Great? Are you hearing that it's okay? Are you hearing that it's terrible? Are you wondering what site search is? Just give us a real quick question and that'll kinda give us a sense… of kind of where the world's at right now… you know, some of what? Okay, we got… some… alright. Interesting."

Tom Washek: "Do tell Jason."

Jason Hein: "Well, it's interesting because I can't publish the poll while you're presenting something. Didn't know that. All right. We've got somebody who is better than average. That's exciting. Okay. Well, so right now we've got two votes for terrible, five votes for about average, one vote for better than average. So at least everybody has an eCommerce site so far. Great. All right. We're gonna keep rolling here. I'm gonna close that. So AI is something that we've all heard a lot of talk about. We've all read a lot of articles talking about how AI is this wonderful thing. But, Tom, you've got this graphic here that kind of shows what AI is. It's got a lot of different boxes, a lot of pretty colors. What exactly am I looking at? And how does this explain what AI is?"

Tom Washek: "Absolutely. Yes. So you are looking at a bunch of pretty boxes. And here AI is really a thing. We have spent the last 13 plus years, almost 14 years developing our AI model, this machine model that allows us to focus on how to enhance that product Discovery process for any end user. So someone who's coming in first time to your site, someone who's looking for a thing or someone who's looking for a very precise thing like, you know, specific widget I try to depict in this diagram here where Bloomreach touches or why, where AI intersects with that buying journey. So every yellow box is sort of a pit stop if you will be on that journey. So from the moment you interact with the search box or the moment when you land on a listing page, whether it's a search page or category page, or as you go through the checkout process, all these blue boxes indicate some sort of automated intelligent process that enhances that journey. So the True value of blue reach is that there are many of those blue boxes that are wired up with these purple lines because they work together. The technologies out there. There are point AI solutions that fix a thing, right? So you can have somebody come in and enhance your product Attribution, they can enhance your product description, and that's great. But how does that Impact your revenue? How is that gonna help you with Conversions? It's going to help you a little bit. But the trick is to be able to Impact the entire journey in a very intelligent way and that's what we do very well."

Jason Hein: "So if I'm here, if I'm reading you, right? Tom, all of these boxes represent sort of stages that either the customer goes through or the code that's underlying the website goes through to manage, you know, basically the experience that the customer gets on your site and… where the algorithms where the machine learning lives in this little cloud. One of the ways to evaluate whether a particular solution… is better or worse more powerful, less powerful is how many of these little blocks can, is the AI connecting to because these arrows actually while they all look like one way arrows, there's actually two ways, right? They're actually learning from the behavior of the users when they are, you know, looking at synonyms, when they're looking at search results, all of the behavior that user is giving is being tracked and recorded by that machine learning algorithm. So that in the future, other users who come in practice those same behaviors. Now, we can better predict what kinds of outputs, what kinds of influences to each of these boxes are going to be most successful for the business."

Tom Washek: "Absolutely. And I think that one of the points that I failed to mention here is that not only that this technology impacts the end user but it also helps the business user in the dashboard to be able to identify and prioritize other opportunities, for example, give you ideas for promotions and things of that sort. So, yes, we can, there's a lot of great things. But I think that as we go through some of the examples in our presentation, I think you're going to be able to see more how this fits into the buying."

Jason Hein: "Exactly."

Tom Washek: "So let's in the interest of time, let's move along."

Jason Hein: "Yeah. So we talked about those three differences, right? So the products obviously in B2B are very different. You know, in consumer products, we all know what a shirt does, we know what a phone does, we know what a blender does, but in B2B, some of these products that people have to search for are not necessarily things that they're familiar with. So there's two kinds of ways that are commonly used to search for products in B2B. One is, you know, you think about the design engineer who is just searching by specification, right? They wanna put in this long, you know, search term, you know, quarter inch, 20 high speed steel… gun tap or something like that, and they may not necessarily know what the part numbers are, but they know what all the attributes are and they wanna put in these really long queries. Well, this is where a semantic search engine comes in really handy because the legacy search technologies that generally come with your platforms are about matching. It's about, you give me a series of words or phrases and I'm gonna try to match as many of them as I can to the data that's in my system with a semantic search. What that query is actually understood by the system. So that's the difference in a keyword search. It's about, it knows what those terms are. But in a semantic search, it knows what those terms mean and that's a big differentiator in the space. Tom, can you dig into that a little bit and show how that works?"

Tom Washek: "I think you just covered it. It's really about understanding the meaning of the terms in the query and interpreting the intent of the query versus keyword matching. So, an example that you were trying to remember is the quarter inch Philips pen which had 10 steel machine screws. It is a mouthful, but it's really a series of attributes that have a meaning and that allows blue reach to match an Ideal set of products for the results. But let me just show you. I'm gonna step out of here for a second and just jump into an example where you can see a version of this in action. So I chose… something that's a little bit more tricky and I think you're gonna be able to see what I'm trying to do here. Searching for a screw machine drill isn't that one of your favorites?"

Jason Hein: "I love it because it's a technical jargon term. It's a keyword term where each of the keywords is a noun itself. And so a keyword based search term is really, it's gonna bring you back screws. It might bring you back any something that has the word machine in it and it's gonna bring you back drills which, but if you use it as the phrase screw machine drill, you know, that this is a short length stub length drill bit and this is an example of where you can see the results here are actually matching that definition because it knows what that term means."

Tom Washek: "So, I'm gonna contrast that with Amazon business, for example, the same query brings back… a lot of."

Jason Hein: "One, everything's just a drill. Yeah, they're just a."

Tom Washek: "Well, you can see that there's a lot of noise and, you know what? And I think that with this example, the fact is that you have the right set of results on top. It's very important. But let's sort of continue and just say and continue to talk about the semantic search because semantic search is a lot more than just understanding those queries, right? The meaning behind those words. There are additional mechanisms within Bloomreach that allow you to further curate those search results. But we have what's called the high precision mode that allows you based on an understanding, allows us to remove… unnecessary results. There's a thing called a QF, right? Which is…"

Jason Hein: "What, what is making you?"

Tom Washek: "Not misspelled. I assure you, it's automatic query filtering that allows us, it's really popular amongst our automotive customers because you can search for your make model. So understanding that in the query implies fitment and returns a smaller, more narrow, more precise set of results, semantic search. Also part of this in our AI journey is we have the ability to automatically generate synonyms, right? So it's the ability for us to understand these nuanced abbreviations, these terms that are industry specific, and we learn in context of every customer. So you may be selling screws and I’m selling the same thing. But my customers are representative of specific industries and they may have specific terms that they use to identify or discover those products. All of that is learned for each customer. And it drives a much better user experience. We have the ability to learn from individual queries. So let me just show you what that is like. So imagine searching for a Korean ranch versus adjustable wrench. I'm gonna bounce back and forth between the two. So you can see that there are 24 products in the results set. But the sort order in which they appear is a little bit different. And that is because Bloomreach learns that for adjustable ranch query, this product outperforms this product which outperforms the third product, unless there are some other rules in play. But that's in a sense, right? The machine learning that allows you to put the most important product that's most likely to convert in front of your potential buyer."

Jason Hein: "Of those customers. So that the key there is that you're no longer dependent just on the product data to drive your results. You can actually now incorporate the buyer behavior into like people who are searching for a product using one term, might prefer certain products over others because that's maybe a regional thing or a brand Loyalty thing. And that all can be accounted for automatically now because of AI and machine learning."

Tom Washek: "Right. But again, that's another aspect of machine learning. So one thing doesn't really solve a problem. It could solve a nice problem, but it doesn't solve all of the problems. So working together, it definitely drives a whole lot of value. That theme learning from that click stream from those Conversions is also applied to these different to the sort order of facets. So, one query and again, this is up to customers… prerogative if they want to allow machine learning to reflect the order of facets. Great business users may say, no, I always want products that are in stock to be, I want that to be the first filter. I want categories to be second. And then maybe the rest, I will allow for some Dynamic assortment here. And I notice here that they're also boosting this particular value, I guess or maybe not. But client, you also have flexibility to determine the most Ideal order for those facets, right? So you can allow the algorithm to recommend the optimal placement. But there's also a way for a business user to come in and say, hey, maybe I want to highlight a particular brand and I'll show you how we can do that in the dashboard later on. But there are a number of different ways you can configure the system to tell the best story for your customers in terms of search."

Jason Hein: "Yeah, it's hugely powerful. I mean, just having the, having worked with companies in the past to try to figure out like what is the Ideal order for us to display our facets and doing all the research because we know that once it's done, you know, it's set forever, but instead of being able to say like, all right, well, let's just see what our customers want and let's let the order flex depending on that is hugely powerful. So let's flip to the other side here. So there are, yes, there are those customers who may not know the exact product and they need to search by specifications, but then there's some that just get a requisition and it's got a part number. They don't know what that part number is."

Tom Washek: "And this is very unique to B, right? You don't go to your favorite store, you don't go to Amazon, you don't search for a serial number or manufacturer number of."

Jason Hein: "Program search by upc code?"

Tom Washek: "Exactly."

Tom Washek: "That is very unique… to be, right? And to your point, there are manufacturers, part numbers, there are distributor numbers, there are industry specifications or standard numbers, there are custom part numbers. There are partial numbers. So that whole numeric world is extremely, it sounds like it's more of a bit of an exercise for a database, but no, it is actually a search problem, right? So you're not going to have a disconnected search experience where part numbers are driven by one technology, and then complement that with search technology. So you wanna be able to have, you want your search engine ideally to be able to handle both, right? So I'm gonna show you a quick example of that. And this number that I searched for is a specific customer specific ID for this particular hammer. So when I dive into the PDP, that number is nowhere to be found in the product specification and product information. It's not an alternative. There's a web ID, but there's a different number. So again, being able to provide a unique experience for a customer to be able to find the same product called differently numerically is super important. I can…"

Jason Hein: "There could be a different customer who had a different number for that hammer. I set it up so that each of them can use their own number."

Tom Washek: "Right. But there's also this partial matching. So again, you can be super precise, but there was a 51 - there was another number that allows me to find this particular product. So I'm gonna start by typing a 51 dash and going with that, that's going to bring back a whole lot of noise, right? Because this is a partial match. Again, it does apply to certain businesses. If you want to enable the kind of experience you're more than welcome to do that. It's the ability to then sort of start reeling those results back and kind of reducing and bringing back less and less noise. So sometimes you have those applications where you're quite not certain what that partial number is, but the ability to discover and get yourself to the, you know, smaller number of products is extremely important in B2B."

Jason Hein: "Right. So sometimes you wanna see like, hey, how many CNMG 432 do I have to choose from? And I don't want to have to put in the entire industry standard."

Tom Washek: "Exactly. So we talked a little bit about the difference in products, right? How those differ from B2C and all the things that we as private citizens are used to. So let's talk about the difference in customers."

Jason Hein: "Yeah, exactly. So the really different thing in B2B is that you're selling to a business and in businesses, there's a lot more structure to those relationships between the businesses that we are and we represent in the business that we buy from Amazon doesn't want to put any sort of constraints or restrictions on what I can buy as a consumer. But in B2B that's not exactly the same. So there's this concept of what we call custom catalogs which solves the critical problem of, hey, I have different customers who I'm only allowed to sell different things to and guess what they've got different prices from each other. I can walk you through."

Tom Washek: "Sure. So, so custom catalogs again, they apply to certain B2B customers because they, you may cater to a specific industry, right? So you may want to have restricted product assortment from customer to customer from contract to contract, right? So it doesn't not all B2B customers have that requirement but it is a very common requirement from what we've seen. And the way we handle those restrictions is through our custom catalog feature. And that custom catalog feature could be used in a couple of different ways. It's mainly used to restrict the assortment, but also could be used for filtering and boosting particular products. So let me show you how staples advantage is doing this. So they were gracious enough and they granted us a little… a test account here that I am able to, I have some products on contract. So a basic query of printer paper brings back almost 1,300 different options for me to purchase from, right? I have a filter where I can, where I can go from 30, nearly 1,300 to about 32. I believe products that are on my contract, right? But you can see that number has drastically changed. So again, if I, if you choose to restrict that view and never show anyone that there are 1,200 other options for you to choose from. Fine. But if you want to be able to allow them to discover those products and potentially add them to their contract, great, you can also, I can clear this and then you can see that you can use this as a short value. So I can push all of my contracted items to the top of my results versus showing you things by price, low to high to low. So again mileage does vary and you have that flexibility to determine what is the most Ideal scenario to handle by or how to cater to your customers."

Jason Hein: "Yeah, it's such an important part of B2B that, you know, having the ability to just handle that complexity is really important. And now we're going to talk about my favorite topic which is the concept of personalization, which you cannot throw a stick at in the industry trade literature without hearing people talk about how important personalization is in B2B, right? The whole B2C experience for B2B. But the problem is that in B2B, we're not personalizing it for people. The customer isn't a person, the customer is a business, or it might even be like a department. So, how do you create a personalized experience when you need to account for more than just one person?"

Tom Washek: "And Jason, before we even get into personalization, what does that really mean? So when we speak to our prospective clients and with our customers, when they think of personalization, sometimes they really mean Recommendations, which is, you know, again not to abuse the Amazon example, but it's like people who bought this, also bought that it's a personalization. Is it really good, if they think it is great? It is. But personalization to your point in our world especially in B2B is a little bit different, right? We have some data. We have some stats that really help us sort of drive and continue to grow the personalization efforts for B2B customers specifically relevant by Segment, super huge, super important. Again, it doesn't always apply to all of our customers because some of them don't even think of segments, but Segment really could be anything, right? You could be catering to a specific industry like you don't want me to buy a laser machine for medical use, right? Because there's regulatory Requirements. So there's you have to have some guardrails, I shouldn't be able to find that stuff on, you know, certain websites, you know, I shouldn't be allowed to ship plutonium across state lines, right? Because not that I've tried, but so there's a lot of, there's a lot of really interesting things that segments could do, right? So let me kind of show you what that is in the bloom rage world. When someone goes in and searches for gloves, for example, there is, we produce a clean set of products. We'll show you all the gloves that you currently are selling that are available. We will then prioritize them based on that Conversion. So product in position one always sells better than product in position two. Unless there's some sort of promotion or merchandising rule and play. But that said everyone gets the same thing. Doesn't matter who I am. So if you apply a Segment to this, then whether I come in from a doctor's office, I get a different set of products. If I represent the food industry, food and industry, I have a different set of products. If I am a mechanic or a machine shop, right? I want to be a machinist, I'm going to be able to see different types of products. So by passing that Segment information with every search request, blue reach will learn the popularity, which products are going to be most likely sold or purchased by those individuals representing those industries. Have another example, adjustable ranch, same thing if you're representing a specific Segment, you should be able to find exactly… the accurate set of products that relates to your line of business."

Jason Hein: "Yeah. This is really about defining people like you, right? So, instead of it just being everybody in my customer base, right? People like you bought these night trial gloves, people like you bought this black ox, I coded ranch. It's like, well, let's look at what people like you. Are they other medical buyers? Are there other shops for Foreman? Are they other design engineers? And it looks at just the behavior of people in that Segment to better recommend or better adjust search result relevance to match the needs of people in that particular Segment, whether that's a geographical Segment, a regulatory Segment, and industry Segment, however you want."

Tom Washek: "Yeah, absolutely. So segmentation is one thing, right? And segmentation could be very automatic. You define your segments because you know who you're selling to and you unleash the machines and then everything else will tweak itself over time. Great. But then for those customers who actually have merchandisers, and they want to be able to be a little bit more prescriptive, you can personalize experiences for their targeted customers by applying targeting rules and I'll show that shortly how you can come in and you can change the way products are set up within a search results or a given category, and then be able to target those results to a Segment. Right? Then there's also pathways and Recommendations. So what are Recommendations? I think everybody knows, people who bought this, also bought that. But then there are also pathways. So let me just jump out of here and show you what those, what I mean. So again, this could apply to some of you out there in B2B where people who are looking at this particular valve… they view other products or they purchase other products, right? So in some cases, that works again may not necessarily apply to all B2B customers. But there we can, we have algorithms for highlighting past purchases. We have experience lead Recommendations which are more real time. So even though you are a Segment, even though you are an individual, you can, because you've spent and showed affinity towards particular products and categories, we can build these Recommendations of products that you might that convert better than the products that you looked at within those search queries or within those category results and pre present those to you. We can show you similar products. We can show your trending products, but those are automatic things. Those are machine driven things. There's also a pathway option which is an effectively manual configuration. So our friends are using pathways to highlight instant savings. So if you want to feature particular products, if we are in this gift giving season, if you want to highlight products that are under a specific amount, if you want to shop new products, you can show featured products. Again. You can do more things. You can recognize whether this person is a new visitor to your site or returning customer. That's kind of more of a B2B or B2C trick, but you can, based on certain trends or behaviors, you can dynamically change them out and display different recommendation strategies."

Jason Hein: "So, Tom, is it safe to say that like the Recommendations function is really kind of that, you know, bulk behavior driven customers who bought also bought customers who viewed also viewed, but, you know, there are other kinds of merchandising that companies want to do and whether B2B folks are experienced or familiar with that or not, there are other options where the person can kind of move the needle or create something a little bit more curated around that particular promotion, whether that's to hit a certain rebate number or to hit a, to try to emphasize private label lines, having the ability to create these other kinds of recommendation experiences that are a little bit more that are still automated and still scalable, but also can be tuned a little bit more precisely, can add a lot of…"

Tom Washek: "Absolutely. And again, it's a lot of… the underpinnings to be able to enable some of those recommendation strategies. Some of it is based on the actual data itself. So we would need certain connections to be, to exist within the product data, right? So we wanna know which product fits with another product or goes with another product or is considered an accessory of another product. And then we can apply the performance elements to this and then recommend those. And then you can, yes. And then you can manually, also, you can manually define things like, hey, I have a high inventory of these particular products in my warehouse X. Therefore maybe for those customers in that Segment, that shop in that particular location, I wanna be able to apply that. So, yes, absolutely. Again… some of those things really are great for more mature customers. But for those who, you know, are just in their infant stages or they're taking their first journey with search technology, those things, knowing that those things are possible and that you can leverage those strategies. I think it's super important. Okay?"

Jason Hein: "Such a wonderful transition. It's almost as though that was intentional. So looking at all of these examples here, you can kind of see the power of what relevance segment can do, and that basically sets you up to solve."

Tom Washek: "One."

Jason Hein: "Of the biggest issues in B2B right now when it comes to digital is that B2B companies because of how they have gone to market over the last 150 years or so simply, you know, those interactions when it comes to sales, when it comes to marketing, a lot of those have just been conversations that sales people have had with individual customers, merchandising as a concept, unless you worked for a catalog distributor like the rangers, the master cars, the FAST all, if you work for those kinds of companies, maybe you had some print merchandising that happens, but you didn't really have the backstory of and the capability to do the kinds of what we call web merchandising or digital merchandising that a lot of the B2C companies have already been demonstrating experience with. When you hear people talk about bringing the B2C experience to B, that's kind of what they mean, they want to have products served up. Have you considered these that are actually relevant to what the, to, what the customer is trying to find? So, this idea of B2B merchandising is one that is very new to this industry and that's what we're gonna talk a little bit here today about how technology and AI can help companies that may not have a merchandising department, may not have merchandising experience to get themselves ramped up? Like this is actually one of the powerful things I think about Bloomreach is it has the capability to support companies that are both at an early stage in their merchandising experience because the AI and the algorithms can do a lot of the lifting around behavior. But as the people who are doing the work, get more experience, get more comfortable, they're able to start to layer their work on top of the algorithm and be more strategic about driving what is in the head of that business. Tom, can you kinda walkthrough that a little bit more with a little bit more intelligent?"

Tom Washek: "Sure."

Tom Washek: "I don't know. I'm gonna follow that. So, yes, merchandising is a fairly new concept for the B2B prospects that I've been having conversations with and some of them are, they have very small teams. So they, the key to rely on some of those AI elements is super important. Those things that we try to cover earlier in our presentation. But for those that have some merchandising strategies that they want to deploy and continue to manage things like I want to highlight a particular brand, I wanna be able to have a promotion or I have a sponsored product that I wanna showcase. Sure there are many ways you can manage those experiences within Bloomreach. So I'm gonna show you a couple of things, very real examples that could be very familiar to our audiences today. And again, this isn't going to be rocket science, but it's going to be something that you will definitely use within the system. So, well, let me just jump out of here and just go into our dashboard. So I'm gonna tell you a little bit about what you're seeing here. There are a number of different tools in this dashboard. When you log in, there are different modules. So they're all permission based. So if you don't want your merchandisers to mess around with… let's say insights which is our intelligent actionable workflow that talks about all about the analytics. You don't want them to have access. Great. If you don't want them to have access to setup and stuff like that, you can manage that experience very easily. What I'm gonna show you is how you can actually do something very simple. You know, the good practice is and Jason always keeps me honest. Here is when you have your product taxonomy, it is very important for that product to have its home. It has to have, it has to live in a specific category. It's important for a lot of reasons. You wanna be able to be able to navigate to that product. You wanna be able to from our perspective, you wanna be able to attribute the conversion to the right category and stuff like that. So you can determine the dominant categories where products are featured, but merchandisers can actually break that model for merchandising purposes for a short period of time. So I'll show you I'm going to change my ranking rule for a query of light switch. I'm searching for light switches and what you will see here is you're gonna get a bunch of light switches that in the real environment will be prioritized based on those Conversions. Signals that we capture from every buying journey and those will be representative here. They'll show you, we'll show you what the cart rates are, Conversion rates are revenue per visit numbers are, so that you can see why certain products are in particular order. You can easily just move them around if you want. You can pin them into specific locations. You can engage with any and all of the attributes that are available about this product. So if you wanna be able to boost products of a particular brand by let's say nine point eight points, you can easily do so, and those products will be changed up. But what I wanna do is I wanna be able to, in addition to all the things that I wanna do is I wanna be able to maybe add a particular product. So I wanna feature a switch box… that with all of my switches, why? Because I'm a merchandiser, I really want to be able to see if I can, if this actually makes sense. Maybe I have some, the signals from my analytics that would indicate that frequently products that are brought together are… would, yeah to be able to just say, hey, let me just show this to anyone who comes in and searches for a light switch. Let me show him a switch box and see what happens. You can apply this particular configuration to a specific audience, right? So, if I have people that shop in my warehouse in New Jersey, I wanna test, I wanna just promote this for them. Great. If I have people coming in from an email campaign or coming in from fill in a Segment, you can target those audiences with those changes. You can also apply any date range. So if I want this to happen next week for 12 hours on a Thursday. Great. I can do that as well. I can preview. So I can see the exact changes because, you know, people are using my search technology all the time. You can see, this is my current state of search results and the changes that I'm making will sort of move this product by one, move it up by one. This is a new product and everything else. This product was bumped by two places and everything else is sort of by one, right? So you get to see how, visually, how you're impacting the experience today. If I save this, then everybody gets this. If I save this as a variant, this is one of the simplest ways to start an AB test again. No, it is just business testing this particular merchandising tactic and being able to say, hey, if, let me just start this test and see what else you can do here. So, right from within here, you get to determine if you're selling a lot of light switches and that query generates a whole lot of traffic. You may not want everyone to be part of that test, but you can easily scale this down to maybe 50 percent of the query. Traffic. Bloomreach will do all the heavy lifting. We will split the traffic. We will notify you when the test has reached a critical significance, so that you can see what those add to cart rates and different Av deltas are. So you can determine if this was a good strategy to deploy."

Jason Hein: "And most unfortunately, Thoma, as you pointed out, I noticed you didn't have to. I didn't see you doing any JavaScript, I didn't see you cutting a ticket to it like you as I as a regular human being look like I could actually do this, which is really exciting."

Tom Washek: "Yeah, absolutely. So this is all it's designed for business users and merchandisers without any intervention here. I'm gonna show another example. I've mentioned brand spotlighting, right? So if for some of you that you have your own brand, you have your own label products and you want to be able, you can boost them. You can highlight them in a pathway for every category or search results, but you can also do something very simple or very little. Where mad lines for example, for any query, you can search for paper towels for drapes for gloves. They will always feature their brand first, right? So that's their way to say, hey, if you're not displaying the number of products that are shown for each individual brand, but they give you a very quick way to be able to get to their pro first. Again, that's their strategy, that's their business. So let me show you another thing which is very real. I spoke to some of our, to a customer, one of a, so of our brands and they had a since I'm sure you guys can identify with this as well. The product content is not always authored by you. So you don't always, you get it from your supplier. You get it from somewhere and that the way people describe certain things is really different. Again, this is where machine learning kinda streamlines that information. But also what's important is that you may, you know, they sell wire, right? And they, the wire could be yellow and black could be black and yellow could be yellowish, dash black those exactly. So what you can do is not only that you can prioritize, but also that you can determine which facets you want to show and the order in which you want them to appear. You may, this is an example where, you know, I can boost the 3M brand. Let's say I can… boost this particular brand and I can save it. So 3M will always show up as the first brand for me, in the brand facet, but I wanted to, maybe I'm gonna show you how to… merge certain color values, right? So I can take, you know, my black and magenta and gold. And this will not make sense for the people that are not color blind, just gonna change this to something. So I can give it any name I want, right? And maybe I want to push that color to the top, right? And maybe I want to boost it. So again, you have a lot of flexibility, you know, this could really be a quick fix. You don't have to wait for your IT team to be able to go and do some data transformation and then feed that. To the search engine and then wait for the results. It's a very quick way. You can be very Agile. If you need to remove a product from results, right? One or another customer comes to mind when COVID was hitting, they blocked mass."

Tom Washek: "From search results, for traffic coming in from specific regions, right? So you have, you can, there's a lot of examples that you can use this technology to be super quick with… making any sort of changes to your environment?"

Jason Hein: "Yeah, great."

Tom Washek: "Okay. I know that we're almost at time. So we talked about the different types of products. We talked about the different customers. We talked about the different roles we're trying to solve. So just let's all this down. Let's bring that home before we open."

Jason Hein: "There's three things there's three things to take away from this conversation. One is that because of all these problems, if you're going to be, if you're going to create a good experience in B2B online, you need search that is more sophisticated. You need, you're not going to be able to manually handle everything manually just to write all the rules. Just means your people are doing a lot of time."

Tom Washek: "It doesn't scale. You have to have a…"

Jason Hein: "So your technology has to be able to adapt to them. And that's really where, you know, AI becomes practical in terms of being able to manage these rules across large catalogs with large numbers of products, large numbers of customers with different use cases and catalogs, and different applications. And finally, all of this technology is only as good as the people that you have, who can use it. And so, if you are trying to build an eCommerce team and you have to make a decision about, well, do I to go out and hire a bunch of people who know how to code, who know how to write scripts in order to, in order to do merchandising or would it be better for me to take people that I already have, maybe who are on the inside sales team? Maybe we're in customer service who know the products already and give them some technology that they can use to become digital. Then they're that much better off. So… let's…"

Tom Washek: "I'm gonna stop sharing and let's just open this?"

Jason Hein: "Go to the name. All right?"

Ashley Giese: "All right. Hi guys. So I'm just jumping in as short of a silent listener until now. However I wanna just help facilitate the Q and a portion of the webinar. So, Jason and Tom, thank you so much for, you know, such incredible insights and overview here. We do have a couple of questions coming in. So I'll start with one from Connor. He asks what are some tips to try and get some of our existing customers to move over to eCommerce when it is common or it is more common for people to buy from people directly?"

Jason Hein: "So when I get this question, the first thing I say is, well, you gotta ask yourself, why are people buying from our people, right? And it's because the people that they're working with right now, you're outside sales folks are actually adding value to those conversations. They're being helpful. They recognize. When a question gets asked by a buyer who says, hey, do you have this thing in this color or this size that your sales people know the answer to? And it doesn't take a whole lot of time or effort for that buyer to get that answer. So when it comes to the simpler questions, the less complicated, the stuff that doesn't need an engineer type of solution, a consultative kind of answer, if you, if your search can handle those questions, if it can recognize the term that user is using and actually put together here's, a collection of products that I think matches that term. And then I can sort them in an order that makes sense for what I need to see in my industry… that simple ability to recognize what they're saying, show them the stuff that's actually what they intend and then put it in the right order. It takes a huge load off of your salespeople. And in fact, those types of questions are the things that your sales people don't wanna do anyway, they want to be doing relationship building, they want to be doing consultative selling. So if the, if your search can start to answer the kinds of questions and not be a source of frustration for your users that in my experience, that's what I've seen as being a big blocker is that I go to the website, I search for things, but it just doesn't work."

Tom Washek: "And I think Jason, I can just add a couple of things. You know, this is a common problem that we've seen before. I think it could be, it could lead to technology problems, right? So your search may not necessarily match with the people that you're selling to and our search may not work. There. Are you there's more questions, right? Do you, are your products available? Are they behind the log in or in front of a login? What are your seo stats rate is, you know, how are those products being discovered? There's a number of ways to potentially help with this, right? I mean we were from Bloomreach perspective, we can help assess some of those problems we can help identify and how those problems could be solved. We can prove out how our technology could actually make more changes or how can address some of those challenges you may have through like a simple poc process?"

Jason Hein: "Well, I help how we help them make more money. I mean, that's ultimately that's really what we're trying to solve the problem here is, I mean, and, you know, certainly one of the challenges with digital and any sort of investment and technology, right? Do you gotta sell this up stairs?"

Ashley Giese: "Right."

Jason Hein: "You have, you have to convince the C suite folks who may not necessarily be digitally savvy, that this is gonna matter and you have to boil everything into is either going to get more revenue. It's gonna cut costs or it's gonna make my customers more happy. And the fact that this technology exists right on top of the transaction… makes that a lot easier because we can run AB tests and show that, hey, making this change using this technology from these algorithms has caused our average order value to go up by 50 dollars in order or our Conversion rate to go up by five percent. Those things all help you translate your digital business into the dollars, whether it's revenue or costs and customer satisfaction because you start to get better traffic, more engagement… and."

Ashley Giese: "That's great. So, I hope kinda that answers your question. But of course, we can dig into that deeper at any time. If you'd like… of course, please feel free to continue asking questions as we do have a couple more minutes here, Jason, I had a question for you that you might build the answer to, of course chimein, but as it relates to AI going back to sort of the beginning portion of the conversation, how long does it take for the AI to learn enough about the user behavior before it starts improving that experience?"

Tom Washek: "Great question. So, I'll start Jason, you can fill in. Typically, what happens is it takes about 30 days and it's so we relaunch the learning portion of the solution first about 30 days prior to powering the search experience, the search box and search results and categories. So it takes roughly about 30 days, but then that, it's a rolling 30 days. So it keeps learning over time and the learning is all about the products. It's about the queries, it's about the engagement with various elements within the experience. It's about the Recommendations. So the simple answer is 30 days."

Jason Hein: "And it only gets better over time, right? I mean, I think it's that it's not looking at just, you know, 30 days and that's it, there's also some, you know, some fundamentals that are in the software now in the platform now, right? A lot of what powers semantic understanding is to some extent the history of what the technology has seen over the 13 years that the algorithms have been tuning, right? I mean, it's not isolated just to that one customer. There's a baseline as."

Ashley Giese: "I think we fit or we're pretty passionate about here at bloom rage is our ability to do relevance by segments and just how powerful that can be for the end user. But I'm curious, what if… a company doesn't have systems of segments for the customer base yet? How might they be able to take advantage of something like that? Or how should they be considering something like that?"

Ashley Giese: "Sorry?"

Jason Hein: "Forgot."

Jason Hein: "Sorry… I was looking at one of the new questions from Chris, I got distracted by Ashley."

Ashley Giese: "Apologies, we can definitely jump and answer Chris's question."

Jason Hein: "Yeah. So he asked, you know, we have the example of the quarter inch Philips ahead Sale machines crew. And he was asking, could we outline where you see this domain knowledge comes from to do this as your search, do this level of object extraction to generate subsequent faceted searches. If so, can you get at the domain parsing to then fire at a different search engine? So for instance, the ontology in the box where you recognize that a quarter inch was a unit and so on is returned."

Tom Washek: "So compounded question?"

Jason Hein: "This is a, this is a very…"

Tom Washek: "No."

Jason Hein: "Detailed question."

Tom Washek: "Yes, Chris. Thank you. It's so the domain knowledge comes… from the model that we've built for semantic search. Be able to understand the pattern of the, how this Philips stainless deal machine screw is described. It's typically some type of attribute followed by an attribute by an attribute, by an attribute that ultimately leads to a product. So, understanding where that product is sort of part of the secret source, how we do that. And then we match that against some of the attributes that are defined in the product feed. So in the product feed, you will say that this product is a screw. It lives in this category. It's part of this taxonomy. It has a brand of, it has a material of, it has a color of, so all of that matching takes place when we serve up those search results. The… is the second part of the question more about sending that toward another search engine. Is that what you're? I'm not sure if I'm understanding the second part of the question. Yeah. So, okay, great. So, that would not net. So that's more of a federated search approach and that's something that we wouldn't necessarily do. So we have to build the index from your product catalog. We, the part of the response from the search engine, includes all of the results. So… you wouldn't necessarily take that response and send it to another search engine."

Ashley Giese: "Sake of time, I know we're wrapping up here in just a minute. So I wanna make sure that if there are any other questions, you can definitely follow up with us, reach out directly. We'll be sending some post communications with this deck and this recording. So by all means you should have all the means to be able to connect with us here after. But Jason, any quick closing words to?"

Tom Washek: "Thank you for everyone. Thank you for your attendance. And I hope you enjoyed it."

Ashley Giese: "Yeah, thanks so much guys."

Jason Hein: "It's a hugely important topic and we're here. If you have other questions, don't hesitate."

Ashley Giese: "Yeah. Any time? All right. Well, we'll be in touch soon and thank you, Tom. Thank you, Jason for taking the time to walk everyone through this."

Tom Washek: "Thank you guys."