Welcome to our third webinar. Really excited to have you. This today's session is really exciting. Something that we work, with I can see some from your name. With a lot of you, usually, that's one of the biggest, you know, pain points that we have. How can we work on our results? Why is this result number two, not number one, etcetera. So we're really excited to share with you, our presentation today, and we're gonna go into demos as well. Anyway, let's give everybody maybe another minute or so before we go ahead and get started. Give me thirty more seconds, and we can get started. I see some people, just joined. Mhmm. Alright. I think we have Goran. Mhmm. Go ahead, Vanessa. Agreed. Alright. So welcome everyone to our third session. As Hamid said, we're really excited, about this one because it's a topic that, comes up quite a bit, and it's really fun to get in the platform and actually, you know, work with the query pipeline rules and look at reporting. And it's a really big, key area of of the Coveo platform. So before we get into the agenda for today, we just wanted to do some quick, introduction. So today, it'll be myself. I'm Vanessa, CSM at Coveo, with Hamid, who's a manager here of customer success, to walk you through, what relevance what relevancy tuning is here at Coveo and how to actually go and investigate some relevancy issues, using some troubleshooting tools. And lastly, Devin Thompson. She's our director of customer experience. She is on vacation, but just thought we'd mention her if ever there's any feedback on the sessions. She's always happy to hear, our customers, feedback. So before we begin, we did wanna ask, all of you a quick question around, you know, experience with, tools, at Coveo. So you should see a poll pop up, and you can go ahead and just take a quick minute or so to answer, these questions. I got k. A lot of people answering. Ten more seconds. Four, three, two, one, and I will end the poll right now. Perfect. Done. Awesome. It's it's actually it's it's interesting. So, the question the question was which of the which of the tools below are you familiar with? We had analytics reports, VISTA Browser, Relevance Inspector, and a debug panel dev tools. VISTA Browser was the one that most people are familiar with. Analytic report was, closed second. Relevance inspector and debug panel were the least, ones you guys are familiar with. And, honestly, that's understandable. Both the visit browser and the analytics, reports are directly in in in the in in the in the system. So, and and we usually most of our courses and most of our sessions, revolve around them. The debug panel, it's it I I wouldn't say it's technical, but it it it involves a bit more expertise from the front end to launch it. And the relevance inspector, it's kinda hybrid between the front end and the admin console, but it's really good to know. So it'll give us a better idea what to what to concentrate on the session, and we'll show you those four for sure. Yes. Alright. Perfect. I'll just stop sharing, the poll. Mhmm. Done. I just wanted to share a few tips, just around, the webinar. So if ever you have any questions, you should have the q and a functionalities. You can always pop them in there, and we'll, address those throughout, the session. And if ever you wanna just chat or add any comments, you can also use, the chat functionality as well. So for today, just to walk you through, high level what we're planning to Coveo, In the first half, I'll be covering kind of the theory around relevancy tuning, what that means, what that looks like at Coveo. I'll share some of the most common types of relevancy issues. So, from Hamid and I's, experience at Coveo, what are the kind of two groups of issues that we see most often with our customers? And then, Hamid will actually walk you through the platform to investigate some of those, relevancy issues using data. And throughout, I'll mention, some of the tools, we can use. And, of course, as Hamid just mentioned, show you some of the ones that were, mentioned in the poll. So let's begin just by talking about what relevancy tuning looks like at Coveo. So if ever you've spoken to, you know, customer success, you've gone on Coveo Connect, to look up some documentation, or you've even taken a course on level up, you'll see these two words appear quite a bit. So you'll see relevancy tuning come up quite a bit at Coveo. They're big buzzwords, and there's good reason for it. So the reason why this is really important to us is because it's a way to actually optimize the search experience and to tweak what the end users will see as results. So, typically, what we will do, to kind of go about tuning is we'll actually look at reporting to understand trends and relevance. So we don't just go through every single, you know, poor results, and fine tune every single one. It's really a process of going in, looking at reporting, looking at some trends, and then, deciding on what ways you wanna go about actually doing the tuning. So, typically, at Coveo, relevancy tuning takes place two places in the journey. The first part, I would say, is usually during the project phase, if ever you start doing any testing or playing around, in your sandbox, you will likely be doing what we would call relevancy tuning. So you'll be playing in the query pipelines, adding some featured results, adding boosting source entries. That's really, really typical during that phase of your Coveo project. And, secondly, post go live. So, Coveo a process that will, you know, it's a process that will, you know, take place once you launch Coveo. Of course, machine learning, if you have it enabled, will also come into play. But post go live, once it starts to learn, we can kind of do some additional tweaking if needed to make sure that, you know, you have really, really relevant relevant results. And just something to keep in mind is that what relevancy is today to your end users might change over time as well. So it's you know, just keep in mind that your end user will kind of dictate what is relevant. So what I might expect going on to your website today might change, tomorrow, and, it's really interesting to look at what you think may have been relevant versus what's actually relevant, to the end users. And how we go about this, quite simple, is actually using the Coveo platform features. So, again, using reporting, our troubleshooting tools, understanding machine learning, and, of course, going into the query pipeline to add in any, featured results, boosting rules, etcetera. Let me Vanessa, we have a yeah. We have a question. Tiro, you raise you raised your hand. Let me allow you to talk. Go for it. See if you have any questions. You're on mute if you're trying to speak. Oh, maybe you can just oh, here you go. I think that's good. Yeah. No. I think I pressed the button. Oh, it's okay. My business. No worries. No worries. No pressure. No pressure. Okay. In in any case, if you, we can just ask the question in q and a as well. But, yeah, no worries. Go ahead, Vanessa. No worries. All good. Alright. So I think, you know, this kinda sums up, hopefully, what relevancy tuning looks like for us at Coveo, and what you can expect in terms of, like, what features we will be focusing on, to actually do the tuning. So, in our experience with, you know, across, working with a lot of customers, something that we've, kind of observed is that relevancy issues tend to fall into those two, into two categories. So the first is, really around unexpected results. So, this has to do with any queries that are kind of deemed as low relevance. So, for example, maybe I go on to, a sporting gear site, and I'm looking for, I don't know, boating equipment, and some other sporting kind of, equipment will actually show up in the results. So even though, you know, it's not that there's missing results, it's just that they're very irrelevant to me as the end user. And so the way that I would go about actually fine tuning and making sure that there are relevant results is by looking at some of these specific features. So maybe we have to boost, content related to the query. Maybe we have to add some featured results, or maybe it's just a matter of, the way, that I'm spelling something, and we need to expand, the search term by using the source entries. So unexpected results are just results that are maybe, less relevant to me as as the user. And on the flip side, we have missing results. So this isn't necessarily low relevance, but it's actually when a query, generates a no results, page. So the way that we will go about, diagnosing this will look a little bit different than the unexpected results as this often has to often has to do with missing content or, perhaps exclusion filters, item permissions, that are maybe, you know, maybe it actually is an opportunity to add content that the end users are looking for, or maybe it's just a filter that's blocking, the content from actually showing on the page and so on. So, really, this is here just to illustrate that the way that we'll go about it will look a little bit different depending on the issue, and the tools that we we will use might look a little bit different. So that said, and we'll show this with some examples in just a few moments. But, for example, if we're looking at those missing results, so queries that are generating no results, typically, the way that we'll go about diagnosing this is first by using some out of the box reporting. So we have some really good reports on content gaps where you can actually see those queries that are leading, to no results. We have our content browser where you can actually test, what that looks like. So if you're in a specific query pipeline, we could see the content that's showing, versus another pipeline, for example. We could look at exclusion filters in the pipeline or on the page. And lastly, we could use the debug tool on the actual UI to see, what's happening there as well. So these are the tools that we would recommend, kind of looking at if ever you experience some missing results. And in terms of unexpected results, similarly, we'll also look at out of the box Coveo reporting. So instead of content gaps, we'll be focusing on queries that are, deemed as low relevance, and we'll most likely be using the relevance inspector here to see why that content was actually boosted or shown in the first place. So it's a really interesting way to see, why certain content is maybe, appearing on the page versus, other. So I'll just take a quick pause and see if there's any questions. Hamid, did you have anything? We're good so far. Awesome. Alright. So, with that said, I'll pass it over to you, and we can actually get into the platform. Sounds good. Let me share my screen. Alright. Can you see my screen? Good. So, gonna talk about couple of things. The the theory that Vanessa told us would try to to to give you practical examples. Obviously, we're gonna use, we can't choose customer, you know, an actual org. So we're gonna use our demo org, which is Barca, and, our demo front end. And I'm gonna also use something that you guys are familiar with in the last sessions, Coveo double we call it covey double o seven. That's an internal org that we use, and we, you know, we crawl bunch of unrelated stuff, but it gives us a better idea of of the tools we can use. So I'm gonna start off here. You guys said most of you are familiar with our analytics, which is great. In our analytics, section, there is a really, really good report, which all of you should have in your org. If you don't, you can you can add it because it's a template. It's basically the search, details report or search reference. So to add it, you're gonna go to add and dashboard from template, and then you're gonna scroll down to, search performance. Right? We put it under advanced, and I'll tell you why in a second. So, I'm gonna select the template, and I'm gonna you will choose which, UI you are looking for, which is origin level one or interface. If you don't wanna choose a UI, you wanna see the whole thing, just add a report, and here you go. Once you save it, I'm gonna call it this here, webinar, it will be in in your reports. The reason I like this because it has all the different, tabs that are needed for to troubleshoot issues and to see how relevant it will see the query performance, content gaps, etcetera. So since this is demo and we we we we look at this basically, we have several different sub, sites. I wanna focus on one, which I have it open here. We call it the engineering page of Wicca. So right now, I wanna see only the the the analytics from that, from from from that specific page. So what I'm gonna do, I'm gonna make sure my query pipeline is engineering, query pipeline or or search UI exactly like we we explained to you in last session. Alright. So let's look at maybe look at all June and and July. You can also, you know, check, modify the exact dates you guys want to want to look at. So I will jump right we did our analytics session last time, so I won't go into all of these. I will jump right into our content gap and query performance. So one of the, troubleshooting relevancy issues, which is really the first one we should always look look for as content gaps. Meaning, somebody searched for something and they got no results at all. Some of those would be expected. The query and even machine learning or our did you mean feature was not able to pick it up. This is expected. I won't worry too much about it. On the other hand, some queries seem legit, and I still don't have any content. Maybe I wanna explore this more. So, pro tip pro tip. Always look at the, the queries that have a significant number of visits and a significant number of search events. Because if one person try to search for something twenty times, I'm not gonna worry about it as much as if seven people, unique visits visits or people, search for this specific, query fifteen times. Again, we are in a demo environment. So in your case, you're gonna see maybe thirty thirty unique visits and seventy searches. These are the ones that I wanna pay more attention to. So perfect. I undify GPS tester as a Compton Gap in in our, in our in our report. What do I do next? Number one, what you really need to do is start to replicate it to see if it's something you know? Is it actually a content gap or maybe this is, you know, it got fixed and, you know, this was probably historical. So you're gonna go to your front end, and I'm gonna literally search for the exact same thing. Copy paste GPS tester. Uh-huh. Makes sense. We could not find anything for GPS tester. Right? Barca engineering doesn't have anything related to GPS tester. What we did for Barca, in order not to keep a blank page, we added popular right now, which is a machine learning model. What's popular, what's by popular, it's what's been purchased, most or clicked on the most. So that's a module you can add. Or, for example, maybe you wanna bring results that are just GPS because GPS alone, it has a lot of results. Right? So how do you go about this? This is where where we can use, tools such as the, thesaurus entries. Right? Where I can say, anybody who's searching for GPS tester, give me results for GPS. Right? And then you're gonna have results. But it really as, as Vanessa said at the beginning, it really depends on your end users. Some end users know. In our case, it's engineering. Those folks know exactly what they're looking for. If I show them only GPS results for GPS assessor, they're not looking for GPS. GPS assessor specifically, this is why we did not add a resource entry here. Right? What we did, we we acknowledge that we don't have GPS testers. Right? And we we might have other popular items for now look at. So that's really important to know your end users. You don't want to basically impose stuff on them just to brush off a content gap. And, ideally, I would actually check with my product team. Do we even carry any GPS testers? Can we maybe redirect the the our folks to one of our, you know, other companies, alliance companies that sell GPS testers. Right? So always think from a business perspective, not just, you know, solving for for the actual content gap itself. Alright. So this is one. One the first way to actually go about this is to try it out. Number two, what I would also do, I would I would also check our next tool, which also most of you are are, are familiar with. It's called the visit browser. What is the visit browser? This is the raw data that's coming from the UI, into Coveo. It gives you the the visit, everything about the visit that we know, and every single event that was done during during the visit. Why is it important? It might give me insights into what did the customer after searching for GPS tester, what did they try to do next? And it might give me an idea what's their intent. Right? So let's just, first of all, let's filter on exactly the the the, user query, which is GPS Tester. So I'm gonna say user query is this. Also, very, very important, whatever you'll whatever date you looked at in the reports, make sure you look at the same date here or else it's it's not gonna make sense. Right? So I did the user query as GPS tester and from this date to that date. Right? So, if I look for example, this, event. Let's look at that. That's very interesting. So the person is from Quebec, Canada. Obviously, it's our demo. And then they search for a m f AMFM. They search for GPS. Then they search for GPS tester. And then guess what? They actually search for GPS. So they kinda did whatever I was doing as well. And then they search for handheld GPS. Okay. That's interesting. So they're probably looking for, you know, handheld GPS or GPS in general. So this is telling me what they're doing next. Obviously, this is, again, demo, and this is getting into a specific visit. So I would not jump into conclusions based on one person, but I would look at other people as well. Let's look at this person. What did they do? They really insisted on checking, GPS tester. This person as well, five hundred and ninety nine, ways they really looked for tester. And then, you know, we had the generated answer there, etcetera. And then after tester, they looked for antennas, etcetera. So this is another place to better understand, the the the content gap for sure. Now another place as well is the content browser. Content browser gives you almost the same thing you see here. But in this page, I'm only I'm filtering out special products which are engineering. What I would like to see, do I have in my index, even if it's not I'm I'm not exposing it in the front end, do I have any content that's GPS tester? No. Perfect. That's my answer. But on the other hand, do I have content that's tester? Alright. I have a lot of stuff, not just engineering. Perfect. Do I want to expose any of those products that are testers, not GPS, in my engineering page? Not sure. We have to, obviously, check with our product team and check with our content management team, but that's another tool for us to to to make sure we are, exploring, and we know exactly our our relevant issues. This is what I wanted to to talk about from the content gap perspective. I wanna talk about another aspect right now, but I'll take a pause and see if there's any questions, any comments, Vanessa? Not so far. Nope. I think you're doing good. Beautiful. Alright. So now we talked about, content gaps, but if I go back to the same report that I uploaded, what about those results that essentially have, have results, but the results are not great. How do I know that the results are not great? Right? If I open again my same report that I just created and I go into, query performance and I scroll down a bit, we have another report which is called queries withdrawal relevance. So what does what does this mean? We have what we call the relevance index. Relevance index is a number between zero and one. The closer to one means the the the the query has high relevance. How is this calculated? It's calculated based on the number, average number of results, average number of clicks. How many people did this specific search? What is the rank, average rank of the clicked result when when this this query is searched. So it's it's a bunch of different criteria that's bringing up this, index. Anything low, less than point forty five, point five, this is deemed as a query with lower relevance. Anything higher than that, it's okay. What we could do here, we can sort by queries with, high from higher to lower. And I can say I can say, okay. This is interesting. This is point sixty eight, not too bad, and the sixties. Those are not too bad. So what are those telling me? I can learn from these queries for the queries that have lower relevance. Right? What's difference between handheld portable devices, which has results, but there's really no clicks at all, and AIM, which has high high results. So those would be I would really focus when you're doing this exercise in your org, I would focus on the lowest ten, quotes, unquote, bad, bad searches or searches with the raw relevancy. And I would explore them one by one in the visit browser and in the front end and see what kind of results we're getting. Do I need to add, for instance, a thesaurus entry? Do I need to boost? Maybe I have a result, but it's buried in the third page for some reason because it's not tagged properly. Right? So concentrate on the lowest stand and go from there. Most prop most of the times, when we dynamically solve for some issues example, let's say computers. Right? And then I simply boost all the categories that has PCs in them or computers. It will solve for other problems. Right? So in instead of solving pushing one result, I can push a category of results as well using our ranking expressions, which is something that we we we went over last time, but we can send you also some documentation about it. So that's really important to to look at queries with low relevance. I'll take another pause and see if there there's any questions. Yes. There is a question from Karen. Sure. More around the content gaps. So what is the typical amount of content gaps, that most customers experience, and when should we start to consider it, like, too high? Great question. If the content gap is zero, that's ideal. Right? Because I really don't want any content gaps at all there. But it's impossible to have that because always gonna have someone who's gonna do something like this, right, where it JPS doesn't make sense. Even machine learning cannot, cannot, cannot solve for it. We the best in class is less than five percent of your content has no content gap. Right? And the lower, the better, obviously. We can also it's it's also industry specific and use case specific. Example. In an Internet in our Internet, our content gap is very, very low because, usually, who whoever is the persona that's searching is somebody who is at Coveo, and they know what they're looking for, essentially. Right? While on a public site, there are bots. There are other aspects. You know? Somebody who is who who wants to buy a Barca boat, but they actually went to Barca engineering by mistake through Google, so you're gonna find it a bit more, but I would say less than five percent is our, our goal for sure. Perfect. We actually do have another question. I don't know if you wanna take it now. Sure. Sure. Let's do that. Awesome. So, this one is, what would you say that we should prioritize queries with low relevance index depending on their number of unique visits? Sorry. Would you say that we should prioritize? For sure. Yeah. Really, really good question. So here I have, for in for instance, very similar to content gaps. This is significant. Hundred and fifty one unique visits, hundred and fifty unique user IDs. But this one, Fishfinder, not that significant. Even though it has, you know, very low, very very low index, it's only three people. I would definitely not prioritize this. That's a good question. I said prioritize the top ten or the lowest ten, but, of course, always take into consideration the volume of searches that we're getting, around that specific way. That's a really good question. Yeah. Perfect. Thanks. Awesome. Moving on. I want to, move on to a New York because it it has different stuff. So this is our Coveo seven that I that I talked about. And what I'm gonna do here, I wanna show you something that we always get, and this is one of the specific questions we got offline. Right? So I I have access to to to the front end. Right? But I really don't wanna change stuff, add it to source entry, or, like, test things out from the front end so I don't influence the analytics and the numbers. What can I do? I do not have access to a non prod front end. So number one, we could use the content browser in the of your org. That that's that's one. Number two, we even have something more interesting. Under under, search, we have what you call search pages. These are hosted search pages are not you it's not published when you create it in the front end. Nobody can see it except you when you access the Coveo org. We call it hosted search pages. So you can create a a mimic of your front end, but only in Coveo. And it what's really good is it's in your prod. Like, and you go into your prod org and create it here. You don't have to create it in your non prod conveyor. Right? Once you go and add search page, you have, two options. You have the simple builder built through headless and atomic in the future. You have the advanced classic interface editor using JS UI. If you're more familiar with JS UI, you wanna have access to the actual Coveo. If you're if you're, you know, more on the developer side, you can use the classic interface editor, or you can keep it simple and really go through the simple filter. So let's create one together. We have, I I found somebody crawled something ready to dog Brigitte, so I'm gonna create a page. I'm gonna call it docs, and I'm gonna add a search page. Let me create this. It will ask you to log in. And while doing that, I think something popped up in, in chat if there's any questions. Yes. So there's a good one here. What would you recommend and maybe you're gonna answer this, but Sure. What would you recommend if we have a public website that employees are searching in and creating content gaps? Is there a way to guide them away? Good question. I wasn't gonna cover it, so great question. So in Coveo, we are able to create what we call is internal filter. Very easy couple of clicks based on the IP. Right? So we can have a range of IP addresses, Assuming they're not logged in, by the way, from the, the public side, we have a range of IP addresses which we can upload into the system. It will very important. It will not stop those queries from happening, and it will not stop them from coming to Coveo, but it will add a filter for us in the in on reports to be able to distinguish the internal versus external. Right? And it's really one simple for two. We can send you documentation offline. That's a good that's a really good point. But sometimes even from the even if it's internal, candidates doing stuff from the front end, It's important to also learn from those folks as well. Right? So I wanna know because they have different types of expertise, and I wanna tailor to them as well. Long story short, we can create a a filter use, what we called is is internal IP range. Also, great question. Thanks. Awesome. I created this page as you could see in couple of clicks. This page right now, if you see here, the sources, blah blah blah, it has everything, which I don't want. I said I wanna create a page specifically to dogs. Right? So the page name is dogs. I have it created. Now how do I make sure it's only Going back to last session, what's extremely important is to have a query pipeline pointing to the front end. So pointing to our say our our our, origin one or search hub. So how can I do that? You will never need to create a pipeline from scratch, to be honest. So always go for example, if you're creating a page and you already have a live query pipeline, duplicate it and then point it to this, search app. How do you point it to the search app? Very easy. Double click into the query pipeline, the duplicate, the one you have. I just created this one before the call, to be honest. I kept it, you know, simple. And then what you're gonna do is we have to add a condition. What is the condition? It's literally the search hub is box. Add a condition and save. So now this pipeline has a filter on only dog breeds as site. So if I go here here you go. Completely changed. What happened here? Because I pointed to my, to my search hub. Right? And and now I have only I I can I can I can dissect specifically on the searches you want and the search hub you want? That's step number one, and here you can test really whatever you want. So if I do Perfect. That's a great example as well. What happened here, I completely misspelled Chihuahua. Right? So it corrected it, for me. That's that's one, one example we could do. Alright. So if I want to add, aetisaurus for I wanted to give you an example about your, specifically, but I don't know how to spell it. Let me just do shepherd. If I misspell shepherd with an a, right, I still have some content with shepherd with an a. Right? But I know for a fact that's not how it's it's written. What can I do about it? Again, relevancy perspective. And during UAT during your UAT or if you see it in your reports, a lot of folks are searching for shepherd with completely misspelled not completely, but a instead of e. I'm gonna go here. I'm gonna add it to source rule. I'm gonna say whenever anybody searches for shepherd with an e, sorry, with an a, make sure we change it to an adderall, and then it will be literally applied right away in in the system. So if I go here now and I do this, look at that. Even though I search for chapter with an a, it gave me all the results with an e. Very low hanging fruit, really easy to affect, to affect relevancy on the spots in in two seconds. Now if I if I check for something, let's say this is your dog breed site, and I want, for example, to careers. Right? What can I I wanna apply for a job at your company? There is a lot of careers results here. Right? But it's not what I want. We have an actual career page or that that I want to direct the customers to or tell them, listen. You are not in the right page. You are in the dog breeds description page. If you want to apply, go to our corporate site. And this, again, based on our based on our reporting, what does our reporting tell us telling us, we go there. If it's one person asking it, I really don't want don't not don't mind, but it's not as significant if I have a lot of people trying to find carriers there. So what I would do, we can use also for relevance to, tuning reasons, we can add a trigger. A trigger could be a notification trigger where I'm saying something to our end users, and the trigger would be on top of the results. I can execute a a JavaScript function, so add a pop up, for example, or I can redirect. Let's try two, things right now. First of all, let's try the redirect first. So if anybody searches for carriers, let me redirect them to Coveo carriers, for instance. I will take this URL. Right? And I'm gonna go here back to my triggers. I'm gonna say, take people to this URL, not all the time because I don't want them to redirect it away from my search page, only when the user query contains carriers. Add add a condition, and let's see if that works out. I'm gonna add the rule, and let's try this out. Well, it did not work. So let's how do we how do I troubleshoot this? Right? If I change my condition, I wanna see if the problem with my query itself, right, or something else. So I'm gonna change from contains to is. Carriers. And then add a condition. Save. And let me again, carriers, it did not work. So I have a problem. What do I do in this case? I'm pretty much gonna open a support ticket with Coveo. Right? This is not intentional, by the way, to tell you go to our support. It usually works, but it did not. So we need to probably troubleshoot it a bit more. And this is where I'm gonna have to rely on the expertise of our support team. How come my trigger rule is not working? Right? Now, if I try another trigger rule to see if the problem is with triggers in general or specifically redirects, I can say something like this. So I can use the same condition, which is carriers, query as carriers. Sorry. And I'm gonna say whenever the query is carriers, add this message on top of the results. For open jobs, please visit our corporate page. Let's see. Right? So that's a string. It's really just a sentence on top of the results. I'm gonna add the rule. Rule was created successfully. And let me delete the other one so they don't because they're pretty much looking at the same thing. And I'm gonna go more and I'm gonna delete this. Perfect. Now, let's go back here and search for queries to see if the problem is okay. So problem is probably the the redirect. That's a notification trigger. Right? I literally just created some text, and I can put it here. This text could be changed. It could be altered, using, what we call the atomic handlers on the front end. So you can add pictures. You can even add, links, hyperlinks to the corporate page directly. That obviously requires a bit more, a bit more, you know, technical expertise, from the front end, but it's very doable. And really important, what what I did here, I did not need to code anything. I did it all from the front end, and it was applied directly right away, which really makes makes you guys super powerful when it comes to dissecting our issues and influencing results on the spots or influencing relevancy on the spots. This is what I wanted to talk about. Do we have any questions, Vanessa, or any comments? Or Nope. Not at the moment. You're good. Okay. Beautiful. So I this is what I wanted to cover. So for next steps, what I recommend you doing, whatever we worked through today, try to do the same in your org. Step number one, as a summary, go to the reports, look at content gaps, look at the queries with low relevancy. Right? Then choose you the ones you wanna focus on, try to replicate them, use the visit browser, use the content, the content browser as well, and then check what are the steps that we can do. The CSM could help you out for sure from from our end. So if you have any questions, send it to CSM. We can we can dissect them together for sure. And we have a lot of resources that that you can also use for self serve, obviously. Vanessa is showing us now on the on the, on the screen. Yeah. So we definitely have some further courses that will get even deeper into this topic. So, investing unexpected queries, how to use the relevance inspector, measuring search result ranking. We'll go a little bit deeper into those, topics. And do we wanna launch our last poll? Yes. That's our last last poll, please. Alright. So you should see one last poll. And it's really just, to get some more insight on what you guys wanna dive a little bit deeper into. So I'll just give it a minute or so. Sure. Nice. Results are coming in. Still have couple of people who did not answer. It's gonna take ten seconds. Thank you. Thank you. Awesome. I think we have everybody, answered. Yeah. Great. So summary, the question was, I'll give you also a bit of, results. The question was, which subjects do you want to, dive deeper? Interesting. So fields and metadata machine learning models are the top, with sixty four percent each. So sixty four percent of you guys wanna deep dive there. Close next for with forty five percent is index configuration, and then troubleshooting, dev tools got the least, which is thirty six percent. So what do we do with these polls? Based all those sessions are really all tailored around the questions that we that we're getting for you and the offline submissions you you send us. We're gonna take this, and we're gonna tailor a future session based on it. Some of those, like, fees and metadata or machine learning models, they can, you know, deeper into that one might require bringing maybe an architect or somebody from our, you know, machine learning team to to talk about it, but we'll make sure we tailor it accordingly. As a summary, thank you guys so much for for joining. We will be sending, recordings. If you have any questions after the fact, please send them to us, and we will look forward for the, next session taking into consideration your feedback. Hopefully, this was helpful a bit for all of you as well. Yeah. Thanks, everyone. Thanks, Amin. Thank you. Have a good day. Bye.