ThoughtSpot: Use Cases for Search Analytics

Transcript
All right, we're good. Let's go ahead and get started. We'd love to do a quick intro about InterWorks. If this is your first time being with us on an InterWorks webinar, welcome. Thank you for joining us. We're super stoked you're here. InterWorks is an analytics consultancy. We're a global team with folks in the US, Europe, and Australia. We specialize in helping people and teams succeed with technology, and that wording is very purposeful for us. The people and the teams always come before the technology. We're a partner of ThoughtSpot. They built a great data product that we really love, and we're going to show you why today. But what InterWorks does alongside our partners and customers is help you build strategies and really problem-solving solutions to succeed with whatever technology you've invested in. We can help you along the entire end-to-end analytics spectrum from data architecture and management to data prep, to cloud migrations, server administration, and really all the way to embedded and more advanced analytics, training, and a whole lot more. We'll be talking about ThoughtSpot today, but I really recommend you also check out the InterWorks blog at interworks.com/blog, and you'll find out a lot more about all the tech we support and really get some great advice from our global team. We're known for our blog. A lot of people come up and say, oh, I know about InterWorks from the blog. So really be sure to check it out if you get a chance. And if you follow us on LinkedIn, that'd be awesome because we send a lot of updates, a lot of content out there, plus alerts on any upcoming webinars like these or events that are going to happen. A few housekeeping reminders: we are recording this session and it will be on our website in a few days, and we'll send you an email to let you know when it's available and you can access it and check it out then. One quick favor we do have to ask is please use the Q&A function in Zoom today to help us keep track of the questions you have. We'll have some dedicated time at the end for questions, but we'll also be keeping an eye on that queue throughout, and we'd love if we can make this really conversational and just interrupt and answer those questions as we go. So that's us, but we'd also love to start off with some introductions before we dive into today's content. My name is Ty Ketchum, and I'm a delivery architect for InterWorks based in Fort Worth, Texas. At InterWorks, I help clients develop and execute data strategies from planning and implementation to people enablement and sustainment. So really the whole spectrum there with tools just like ThoughtSpot. Prior to InterWorks, I spent about a decade in aerospace and energy industry in a variety of analytics roles from analytics lead, data scientist, to data whateverist, you name it. On the data viz side, I actually come from a bit of a Power BI background, which really helps me navigate a lot of conversations around ThoughtSpot, Power BI, and where does everything fit in there. But more than anything, I really love helping people solve technical problems. So that's the core of what we do here and what ThoughtSpot helps us do. So I'm going to pass it over to you, Jack. Yeah, thanks, Ty. Great intro. Yeah, I'm Jack Hulbert. I'm an analytics consultant here at InterWorks. I am based in San Francisco, California, calling out of Nevada today, though. And yeah, similar to Ty, I have a background in data analytics. My focus has primarily been in the SMB and retail e-commerce space prior to joining up at InterWorks. Of course here, now we work with clients across a bunch of different industries. My background is more so skewed towards Tableau and some data prep tooling, SQL engineering, things like that, and even some natural language processing, which is sort of the genesis of how I've come to familiarize myself with ThoughtSpot and why I think it's great because it leverages some of the natural language features that I've always seen as a very powerful way for analytics augmentation and making it approachable for end users. So again, very excited to be sharing this presentation with you today. For today, our agenda. So we're going to cover a couple of things. First, for those of you who maybe are new to ThoughtSpot or maybe you've heard of it, we're going to do just a quick overview of ThoughtSpot, the problems it's solving, and some of its core features and value propositions. Then we're going to go into search analytics and building your first use cases for search analytics and how that can look coming from maybe your traditional BI environment. So we're going to walk through some steps, give you some tips on how to build out, flesh out that first or first few use cases for search analytics. And then lastly, we're going to show what it looks like in the wild. So we have a demo of ThoughtSpot along with some other great embedded analytics tech that we have here at InterWorks to show you how it could look in practice. So really excited to share this with you all. Before we dive into the content, just want to be curious for those who are jumping in today, have you heard of ThoughtSpot or self-service analytics? So, you know, yes or no. And of course, if you have any additional context to give in the Q&A, that would be awesome. Really curious to hear where people are in their ThoughtSpot or self-service analytics journey, if you're a current user or if you're curious or you've heard of it but want to learn more. It's always fascinating to see that there are a few people who are joining in and haven't heard of ThoughtSpot. It's always really exciting because then we get to start from step one. Yeah, awesome. I see in the Q&A, just getting familiar with ThoughtSpot. So you've heard a bit, right, but how do we make the most of it? What do we do next? What's the best way, the right way to move forward? And you're in the perfect spot. That's exactly what we're going to cover today. All right. So yeah, I mean, I would expect, you know, self-service analytics and search analytics and ThoughtSpot, ThoughtSpot's been around for quite a bit. I think with the advent of search analytics and it being such an important feature in the modern BI stack, making it easier and more approachable for your users to leverage and get insights on their own, it's a perfect time to be talking about self-service, building your strategy. And here we're going to talk about what are core tenets or what does a good footing look like for a self-service BI strategy? And how should you evaluate tooling? So we're thinking about it with six core pillars. The first one is governance. Good self-service analytics tools and strategy have governance. So again, clear, easy to implement, and repeatable patterns around administering content. And with that, scalability. So having infrastructure around your BI tooling that can support scaling your use case from one user to ten to ten thousand. So a good tool will support that, and it'll also support internal and external embedded users. And also scale with your use cases. So as the number of scenarios or use cases you want to tackle with your self-service analytics strategy grow, it's good to have a tool that can support your data as it grows or your use cases and your user base. Personalization—one of the most important pieces of a self-service strategy is making sure that the insights we're producing within our BI tools or BI stack are putting the insights in the language and in the tooling that is most suitable and comfortable for the end user so they can act on that information. Flexibility and drill-down. Another common pitfall we see with some BI tools is it can be difficult for users to curate and receive the data in the format that they're looking for, making sure that all the meaningful filters that they need are applied to the data so they can do meaningful analysis. Again, good self-service strategy—this is imperative. And ease of sharing, this is one that's really becoming more important as we're adding more tools to our BI ecosystem, our business applications. We need to make sure we are able to give insights to users in their preferred communication channels, in their preferred format so it's portable, it's shareable, and it can be shared outside of that application. And of course, security—probably, you know, there's no strategy without security. You have to make sure that users are accessing the correct data, and there's strict provisions in place to do that. And with these six tenets, we believe ThoughtSpot sits at the center of this and hits all these criteria. That's why we've really wanted to build this presentation, to kind of understand how you can benefit from building a strategy with all of these different pieces. So going past self-service, I want to go into what is ThoughtSpot for those of you who maybe are new? ThoughtSpot is the platform for self-service search analytics. So ThoughtSpot is a business intelligence platform that allows you to connect your enterprise data models in a cloud data warehouse, and it allows you to actually leverage natural language search and keywords to search your business data, to retrieve answers in a visualized format that can then be shared, visualized, and embedded in other applications. It can be customized to different user needs. You can use keywords and GPT-like search to ask your data questions, essentially. And going further into that, the supported data connections for ThoughtSpot—a lot of cloud data warehouses. ThoughtSpot works with cloud data warehouses. And also there's an on-prem edition, but right now, we're mostly going to be talking about the cloud version that supports data in cloud data warehouses like Snowflake. These are some supported connections. I know there's always more being worked on in the pipeline. And that's something we really love about ThoughtSpot looking back on the past couple of slides—it's just how easy and lightweight of almost like a presentation layer that it sets, it just sits on top of your data and works, right? It lets people actually understand and use all the data you have and that you're storing. I mean, you spin up all these warehouses and you spend a lot of time there. You just want to put something on it that allows you to actually use it. And we see ThoughtSpot really as the tool that's allowing us to do that, do that at scale and do it pretty easily. So just something that has always stood out to me from day one of us using ThoughtSpot. Yeah, you're absolutely right. And just being able to search your data live too is awesome. I think having the live connection to data warehouses allows users to get those point-in-time answers that are not always easy to get through other mechanisms. So I love that it's cloud-first focusing on live search. So now, because ThoughtSpot is sort of the first of its kind in this search analytics space, what are the problems that are being solved? So with search-driven analytics in practice, what this looks like is rather than the traditional BI space where we have dashboards and business objects that are created for users based on a collection of requirements, search-driven analytics allows users to be sort of proactive and ask questions, business questions and queries on the fly. So rather than having to put in a request and scope out what a query might look like and then have a lag time between the ask and the delivery, it's actually put into the end user's hands to go and search their data themselves and retrieve an answer and be able to do what they want with it, visualize it in a way that's meaningful, in a format that is meaningful. Again, the answers to these business questions through search can arrive in the form of visualized data, or they can arrive in the form of a table or an export. So again, having that flexibility of what the end output is is one of the key attributes of search-driven analytics, how it can be powerful. For those who are used to highly visual data, they can build objects, build experiences that are low ink on page, very polished KPIs, things like that. But then if you come from the background of wanting to work in pivot tables or larger data tables, you can also curate that experience for yourself as well. And again, these answers can be drilled into, customized, saved, and shared with other users, getting that personalized experience that we mentioned as kind of a core pillar of good self-service, being able to have it be hyper-personalized and meaningful to you and shareable. And one of the biggest things from a BI operations standpoint is having a search-driven analytics strategy and platform. It nearly eliminates the need for most ad hoc SQL and data pulls. If your data is modeled correctly, you can essentially reduce that ticketing system to a very small subset of what it might have been prior to having a search analytics tool that users can approach themselves. So this is a big one for me coming from the traditional BI space, working with BI tools, cloud data warehouses, doing a lot of dashboarding. This piece here definitely resonates with me. Absolutely. So again, I want to ask a couple of other questions here. Pull up a poll. So a couple of questions here, just out of curiosity, as an analytics professional—assuming we're all analytics professionals here, right? We've all worked with data or reporting in some capacity—just wondering where do you find yourself spending most of your time? And also as a follow-up there, where does your data live? It's always so interesting to see that governance is never—it never gets any hits. No love for governance, no love for us governance people. No, but I think it's something we talk about a lot and we all probably recognize that we need to do a bit more there. But certainly, it's hard, right? Yeah, and maybe a lot of the time, in the traditional school of BI, folks are so focused on doing the reporting and the data cleansing, the data modeling that, you know, governance may fall to the wayside. And certainly, we have some strategies and approaches to helping teams with that. But it's always interesting to see how this splits out. Jack, we have a question from the Q&A. If you have an on-prem database, right, does ThoughtSpot just not work then? How are we going to handle that? Yeah, so there is an on-prem offering from ThoughtSpot. I mean, everything we're looking at today is ThoughtSpot Cloud, but actually ThoughtSpot does have an on-prem offering, so you could hook up to your on-prem database. Yep, right. Yeah, I appreciate you putting that in the Q&A. Anything else, just throughout, feel free to throw it in there, and I'll happily interrupt Jack and tell him to be quiet, and we'll answer the questions and then go on. So that's great. Yeah, we definitely see a spread, right? Where do you spend most of your time? Well, I spend it all over the place because I've got to do all of that. So it makes a lot of sense. We definitely have a lot of cloud-first warehousing here, which is a trend that we see everywhere. So yeah, that checks out. That's great to see. Yeah, also one of the things I really love about ThoughtSpot is that it sort of encourages folks, BI teams, to do proper data modeling because of the benefits you get. Having well-modeled data really streamlines the creation of ThoughtSpot objects and search experiences, right? So for the BI folks who are creating these end user experiences, having a good data model to pull from, having good performance, having that security built in, it's all streamlined content creation and ultimately trust. But it's always good to see what the spread is. And we often hear, okay, well, does it require a perfect data model, right? Does ThoughtSpot going to work if everything's just not perfect? Or is it going to work that well? And oftentimes, I'm answering, well, does anything work that well if you don't have a great data model, right? You can be a SQL wizard all you want, but if you don't have it set up in the first place, no matter what you're going to do is going to struggle. So I love that ThoughtSpot kind of pushes best practices natively like you said, Jack, right? It requires you to do the things you should have been doing all along. And there's no other answer where, well, it doesn't work because of this—well, nothing really works if you don't have a great data model set up or at least trying to go better in that direction. So yeah, just kind of inherently forcing some of those best practices and putting the focus where it should be all along, something that we really appreciate about ThoughtSpot. Yeah, absolutely. So in seeing that, we have a lot—and kind of anticipated a lot of folks would respond to ad hoc reporting and data pulls, and then building dashboards and data visualizations as the biggest portion of your time. It'd be good to sort of delineate search analytics versus dashboarding because I think a lot of us are coming at this from a traditional dashboarding background. So kind of comparing them side by side in terms of strategy and tactics, search-driven analytics versus dashboards. In terms of the ownership of content creation, you can think about search-driven analytics—users are the creators and the ones who are fishing for the insights and creating the searches and the queries themselves, leveraging a user-friendly, business-keyword-friendly search interface. So recently, ThoughtSpot launched a GPT integration called ThoughtSpot Sage, which allows you to leverage the power of GPT in the search experience and retrieve business insights directly from a GPT-like search. So again, delineating between—in search-driven analytics, users are creating their own content, whereas in the traditional dashboarding BI realm, BI developers are the ones who are scoping these requests, meeting with stakeholders, and then building and owning that cycle of content creation. And then when you get into the tactics of dashboard design and dashboard development processes, search-driven analytics is designed to maximize flexibility in filtering, building out a drill-down experience, just flexibility and exploration of the data. Whereas dashboards are sort of designed to answer a limited set of very important questions that are outlined and scoped in user interviews, in sort of the fact-finding processes that precede traditional dashboarding. Whereas with search-driven analytics, users can drill into any of those attributes and customize the analyses however they'd like, ideally leveraging anything that's readily available in their data model. And again, with search-driven analytics, content is highly personalized, shareable, which compared to dashboarding—again, a lot of those paths are predefined analyses. There are things that maybe answer your higher-level questions. And if you need to get down any deeper into the weeds, you either have to find an alternative way to explore that data, or export it, or request a new view or experience with that data. And kind of the last thing here, which obviously a little asterisk here because ThoughtSpot does have an on-prem offering, but search-driven analytics is designed to query live data in your warehouse so you get that point-in-time answer off the most recent actionable relevant data. And dashboards kind of work off a myriad of different file formats and data recency. So just kind of delineating the differences between those two. And next I want to go into, right, how do we come to our first search analytics use case from this traditional dashboarding background, and why does search analytics matter? And Jack, before we jump into that, we do have a good question. And I'll tee it up for the Tableau side—doesn't Tableau have Ask Data function for the end users? And they might be getting to this, but what are the pros and cons of ThoughtSpot search versus Tableau or any other one of the myriad of tools? Why do we like ThoughtSpot search relative to all these other offerings? Yeah, it's a great question and a common one, right? I often hear these features compared side by side. Personally, what I've really enjoyed about using the ThoughtSpot search functionality is it's just very intuitive. The keywords—I've recently been playing a lot around with the Sage integration, and it's been awesome just to get that natural language experience, right? So that has been one of the main distinguishing factors for me. The other thing is search is built throughout all aspects of the application, which is a little bit different from Ask Data. Ask Data has always been its own experience within Tableau. It doesn't really sit within the dashboard experience. It's sort of a separate entity or experience for users, whereas in ThoughtSpot, search is at the core of everything you're doing. Every page ostensibly that you're on in the application, you have a search experience within one click, and you can line up your keywords and essentially query your data within one click of anywhere in the application. So that's how I'd answer. Ty, I'd be interested to hear your thoughts there too. Yeah, just my personal perspective on it. Like you said, it felt like search was an add-on that I didn't ever really use in some other tools, and other tools are great, right? There are so many great tools. Tableau's awesome. Power BI is great. ThoughtSpot is wonderful. ThoughtSpot's whole thing is it is search. That's what they do. That's their whole mission. This is the core of their product, and it shows, right? When you use the products, if you use them all, I think you find yourself actually using ThoughtSpot search, and I don't find myself using search in the other tools as much at the moment. And that's what sets it apart for me. I get a lot of requests like, okay, well, I just want to see the number. I just want to figure it out on my own. And putting that self-service back into the hands of the end users, the business users, it's what I've been asked for for a long time by a lot of people. And this really fills that gap and that niche perfectly. So that's what ThoughtSpot does. That's their core mission there. Yep. I would say, yeah, I mean, the TLDR, the short response is that it sits at the center of all of the different screens and features of the application, and it can also be embedded in an embedded analytics experience as well. Certainly. And that's kind of leads into another one that I'll say we're going to cover at the end—can you integrate ThoughtSpot with Tableau, right? A lot of people don't just have one tool, they have many tools or maybe they just have one, whichever it may be. We're going to talk about some of the integration and how we see the embedding piece of it work and where does Tableau fit in with ThoughtSpot or with Power BI or what have you. So we're definitely going to hit on that at the end. Yeah, definitely. But first, we want to address why do we need search analytics in the first place, coming from this background in dashboarding or sort of traditional BI building out business objects? So as dashboard developers, our process sort of looks like this. We have a few stages that come from ideation to execution and creating a data product. This define stage is defining your data requirements. Maybe you have a data model in place, but you need to assess if that data model is fit for new analyses. Building out visual wireframes, working with designers to make sure that whatever's being created is high fidelity, polished. Then also doing some additional data discovery. Development looks like this: you're building out a V1 or a V1 through V3 of a dashboard before even considering launching it to production. You have this laborious process of validating business logic—maybe some of that is net new, maybe it's existing or being modified—going through that design and UX process to make sure things are pixel perfect, and of course, data source development, which can be one of the largest time-consuming processes of dashboard and BI development. We have this refine process. This is where we're gathering user feedback, doing QA on both our visuals and our data. So engaging sort of the front line of users, right, getting that initial feedback, all leading up to a launch period where we have training, documentation, socialization, sharing of this visual or highly visual dashboard that's gone through all these stages. And then into the maintenance phase. So after launching, yeah, this process of managing change requests. Obviously, dashboards are not designed to be one size fits all, so you're always going to have these sort of offshoots, these change requests, triaging bugs, ad hoc reporting, sort of offshoots of what your dashboard is designed to do. So this is how we typically look at the dashboard development process. Again, spend a lot of time in this loop validating logic, pixel perfect visual experiences, and ultimately coming up with what are the priorities of this dashboard. It's hard to design a one-size-fits-all approach, especially when you have all these nuanced business questions that folks want to explore and dive into. And maybe you have to cut corners sometimes in BI development because you're not able to design one tool that's flexible to tackle all the questions. So you eventually spend more time in this phase as well towards the end because in your maintenance process, you're modifying your dashboard to accommodate these unique needs because they're important. I mean, and you can only modify a dashboard to be so flexible. So again, your analysts' inboxes can start to look like this. And if not managed well, this can really increase the training requirements and the amount of effort that analysts put into making sure users are comfortable with the tool, making sure it's useful, making sure that the information they're getting from it is accurate and what they expect compared to a source system or what have you. Yeah, and again, this is where tools can have sort of a pitfall in terms of users are creating workarounds to get the information they need because they can't get it natively out of the tool, right? So this is one of the common challenges of dashboarding that we've seen in a myriad of tools is that they're not always flexible enough to capture the more nuanced questions that often require a data analyst or someone who's extremely data savvy or some sort of feature to tackle that. And this can take a lot of time and kind of take us BI folk away from some of that strategic work, some of those initiatives that we maybe want to get after that are not tackling these tedious, having your email bombarded with bug notices. And I think, in my experience, I've probably shared with you all too—a lot of times I build a lot of really great dashboards, and the minute I create them, right, they're obsolete, or they're something else, right? There's something we've got to change, and maybe we can't change it. And maybe it's a completely different ask. Because if we change it and add too much on, we get lost in translation and what's the point. So I love ThoughtSpot because it really puts again the power back in the consumer's hands. Whatever they need, that one-off type thing, they don't have to go—I don't have to bug Jack, hey, give me an insane SQL query to give me my answer, and I'll ask for another one tomorrow. It allows me to do it in my own language and get the answer on my own and do it on a repeated basis. So I just love the whole dichotomy and really the whole shift in thinking from—dashboards are great, and they still are. And honestly, ThoughtSpot has liveboards, which are a lot like dashboards, and they're great and they have their purpose. And when used well, they're fantastic. It also has this wonderful way to get around all these problems that we're all used to having and dealing with, and it can really help lighten our load and maybe change our prerogative as a data analyst, a data whateverist. It allows us to do more of that strategic work that we probably all want to get to. Yeah, you're absolutely absolutely right, Ty. And these common challenges of BI, especially tackling these edge cases, these questions that dashboards aren't always flexible enough to answer—that's where ThoughtSpot search can be the superhero here. Again, a few of the core features and the value propositions of ThoughtSpot in tackling this problem. First off, right, users can create their own ad hoc queries to leverage that natural language. This is a demo using ThoughtSpot Sage, where you can actually use natural language to ask your data a question, and it'll return—you have the option to see visualized data. Or you can toggle really quickly and get a table and then have that scheduled out to your team or your Slack channel, right, where it can be acted upon within seconds. And you didn't need to write a line of SQL. You didn't need to build your own chart using a BI tool. You just used the thoughts that are streaming from your brain, which is pretty cool. So yeah, again, puts that power in the user's hands. Big one for me is drilling down on your data is super easy and intuitive. You essentially have the capability to drill down into any dimension in your data model in one click. Not even a right click, just one click on a chart. You can activate and filter in on that particular point and cut it by any dimension in your visible data model, which is super powerful. It's really easy to get that lineage and monitoring and keeping track of your flow of analysis as you're doing some data exploration here. Yeah, it's not just drilling down from month to week or week to day, right? It's like Jack said, you don't even have to remember that you need to right-click this thing because you don't. You just click it. You just select it, and then you can drill down into anything else in your dataset, which is really cool, and it brings back to why we love it so much. Easy and intuitive, right? It makes you want to use these features. It goes back to the similar like, well, what's different about the search and ThoughtSpot? What's different about the drill-downs? Well, they're just easy to use, and we find and we've learned that people actually do use these features, which again, there's a lot of cool stuff out there. The coolest thing is the thing you're going to use, and we see these features getting used on a repeated basis because they're easy to do, because it makes sense. Yep. Yeah, on top of that, highly personalized KPI and landing page experience. So when you first log in to ThoughtSpot, you can curate what that first landing point looks like. And if you're not sure where to start, ThoughtSpot has built-in recommendations. So it shows you, using usage-based ranking, you can see what other users in your environment are looking at, what's trending, what's popular. If you have security set up right, you can see that on a group basis, like people in your group, or organization-wide, which I think is awesome, especially for those who may not know where to start, but you want to see what other people are viewing. Very flexible in how you can customize this landing page experience. And again, one of the core tenants of good self-service is being able to get the insights out of the application. Another great thing about ThoughtSpot is you can essentially take the information, the insights, or the visualizations, and send it to other business applications in just a few clicks. You even have the capabilities to integrate it with data write-back and actually create very highly customized integrations with the data you're getting out of ThoughtSpot. So writing it back to Salesforce or a marketing automation system, right? There are ways you could do that within ThoughtSpot as well. And as Ty mentioned, ThoughtSpot, you can create liveboards to group your KPIs and your visualizations. So if you want that more executive-level overview, but also provide the benefits of having a flexible drill-down search experience, you can create a liveboard to have that serve as the starting point for your users. So I know we've been going for a little bit. I'm going to run through this next poll really quick, but just out of curiosity, what are your greatest challenges with self-service analytics today? If you do have a self-service analytics strategy, or if you'd like to start one, where do you think your greatest challenges would be? Yeah, and I mean, even if you don't have that, right, you've got data, you've got analytics, you've got stuff, where's your biggest headache out of these four? From data quality and trust, like, is that the number one? Are we sure? To user adoption skill gap—is it the people, the hearts and the minds? Is it security and compliance? We don't know how to set things up. It's too complicated. We don't trust that we're not showing the wrong stuff to the wrong people. Or an integration piece—we have stuff all over the place. I've got to go to different systems and local files, and I just don't know how to make it all work very easily. Yeah, some of these results are—I mean, it's pretty even compared to what I thought the results would be. Again, the complex data integration, not a surprising number of people. Again, Jack gets to hang out in Lake Tahoe and doesn't have complex data integration problems. So just congratulations on your life, Jack. For the rest of us, totally agree, right? That's a big hiccup I see. And of course, the user adoption, skill gap, the hearts and the minds, the training—a tool is only good if someone uses it. We certainly see that. And the good news is we can do a lot about that. Yeah, and that kind of launches us into building our first use case for search. We kind of went through why we need search analytics. We have all these edge cases, these cases where our BI objects, our business objects are not meeting all of the complex, nuanced, specific needs of our users. And that's where building your first use case for search comes into play. So again, just let's go through—let's imagine we're a BI team. You have a dashboard. It meets all of our requirements, clear KPIs that we've scoped out with our users, visual best practices, and predefined analysis paths that have been determined valuable by our user base. So we have all the right dimensions in there for the analyses that we've identified as relevant and useful for our user base and ostensibly answering all the important questions. So we spend all this time going through this process, and then come launch, we get these sort of additional questions, right? These KPIs—for instance, let's just say in this example, you're a restaurant with lots of franchises, and you'll always have folks who are going to ask, well, I want to see this one level deeper, or I only want to see data in this time period, right? So in a traditional BI ecosystem, right, we might be able to accommodate two or three requests, write a quick query or export the data and land it in a way that can meet these or create a calculation that can create a flexible date analysis period. But then with that, users are always going to have more questions, right? They're going to build upon this. Users will always want to drill down and understand more. And then maybe they're serving the data to other users in presentations, reports, and they're getting asked these questions. And then the chain between source data and questions, the gap widens, right? And this is ultimately where the search analytics, your first search analytics use case starts—thinking about all of these more nuanced specific questions that cannot be answered, or your tool cannot be designed to capture all these questions. So you need an alternative approach that can and can meet users technically where they are and also give them answers to these more nuanced questions. And I think it's great because, you know, what's the point, right? Well, the point is the answer. The point is the end user, us, getting what we need. It's not, I love looking at beautiful things, but that's not necessarily the point. The point is that I can get an answer, I can get it now, and I can make a decision and move forward and do it again throughout the day and throughout the year. So if we focus on what we're really trying to do here, that's where we want to get into and, okay, how does ThoughtSpot help me get there, right? So a few steps into sort of fleshing out your first use case or identifying and building that out. First step is to understand. So looking at all of these unanswered business questions and grouping them and understanding for all these different questions that are being asked of you, what is the size of the user base? What is the overall business cost, which is tough to assess, but usually in terms of time, that's the biggest one. What are the costs of not having the answers to these questions? And what is the risk overall of not having these answers? Those are important questions to pose when you're understanding what you want to focus on for your first search use case. And ultimately, how users want to act on this data, right—is a dashboard or a liveboard or a visual form of data the right experience, or do they need something that arrives to them in a business application elsewhere? So these are all important questions to ask because maybe your users have the right information, but they can't get it in the format or the location that they need it for it to be effective. Second step is to prioritize and organize. Again, the best way to go about this and the easiest way to go about this would be to organize these common business questions by data attributes, sort of build out and draw guardrails around your use case, right? Are they using similar dimensions, sales, or is the audience similar, right, by business or departmental alignment? And also being able to prioritize and organize by user base. So maybe you have a use case where there's thirty to seventy users annually who need to get this one query and there's no current solution for that versus a user base of five hundred or five hundred to one hundred users monthly. Either one of those is a good place to start, but I recommend starting where you're going to have the highest user impact with the lowest effort. So identifying those areas where the data is ready and the user base is large, that is like a perfect setup for your first search analytics use case. And again, that activation method, how they're going to act on the insights is also imperative to understanding that. Step three is to build. Again, assuming that you have your use cases defined, making sure that you have a business data model that supports all these questions to be answered. Oftentimes in some BI tools, you'll have people pulling data and have complex data integration to support your use cases. You want to make sure that all of your data that will help users answer these core questions is supported by a central, well-governed data model with security built in. And again, the analyst job in ThoughtSpot is to build in the business-friendly keywords in making the data model searchable so users can leverage natural language to get their answers and leverage that search experience. Step four, test and validate. So as you build out your searchable worksheets and your visualizations, bringing together a task force with a few core power users, you can test the outcomes, making sure they're matching what the users expect, what's useful. So again, in BI, we have that sort of test and iterate process. We're always going to have that, but what's great about ThoughtSpot and a search analytics tool is you really tighten that loop and you shorten the time from creation to testing, and again, subsequently testing the insights. So you can gather that group of test users, making sure the search experience is optimized and that you have your keyword synonyms and that your data is consistent and secure so it can be launched out to a larger user group. And lastly, launch and monitor, making sure that your end users are trained on search. They know all of the different business keywords. Making sure end users are trained on search and they understand what the full suite of capabilities and keywords they can use are. And then yeah, just creating tutorials and resources as you would with any other BI launch to support the unique needs or the questions that these users might have. And last thing, I want to inject before there. I want to just say my personal anecdote for anyone that cares. The cases that I've seen work really well, they spend a lot of time nailing that data model, which is great, again, just best practice. So they spend a lot of time beefing up and building that data model, and they spend a lot of time training their people, right? They don't just train the power users. They don't just train the data teams that are going to be monitoring it or administering it. They get their core users in a group on a consistent basis, right, and they train, and they talk about it, and they share. So like any other tool, ThoughtSpot is going to be something that if you water it, it will grow, but it's really easy and people can get behind it and understand it because it's all natural language, and they don't have to go read a science book to understand. So from my perspective, it's that build and then the training, right, the launch and monitoring. That's where we see a lot of benefit, right, that exponential growth. Yep. Yeah, and again, this piece here, ThoughtSpot has a very detailed, granular way of tracking adoption across your ThoughtSpot ecosystem so you can target the users who you expect to be power users who maybe are not adopting as fast. So it gives you a very precise way of seeing where your efforts are working and maybe where you want to focus or recalibrate your efforts and revisit this process of creating your search experience. And lastly, alignment. Making sure that your search interface is aligned with other tools, right? So with that, what we're showing here is a way that our initial dashboard that answers a few of our questions, right, our overarching big numbers that we may need to visit on a quarterly or monthly basis can be supported by a search experience. So users can get the benefit of visual analytics and creating their own self-service search analytics in the same experience. And this is all being done in InterWorks' product Curator, which is an embedded analytics ecosystem, which we're going to show you really quick. I want to make sure we save a couple of minutes for any questions towards the end of this, but I just want to quickly show you what this could look like in practice because I know we have some new folks who are seeing ThoughtSpot live. Awesome. Yeah, so I'm going to quickly show—so imagine in this experience we have a dashboard where that has our overarching numbers, and then we have our search experience here where we can leverage our natural language search. So in ThoughtSpot, we can select a data source and very quickly begin searching. So let's just say we want to see the top three—we'll use the franchisee example—franchises for each state by sales. So notice how as I'm typing things in, it's giving me recommendations. So if I don't know where to start my search, I can get some recommendations of what other users are searching on, and it can help me start my journey. And then if you have a more ambiguous field as well, like sales, for instance, it'll give you all—it'll give you multiple options for different versions or related fields to sales, right? It'll give you other measures that you can search by. And you can see here, you can use things that require complex calculations in other tools. You can just type that right into the search bar here, and it'll allow you to fire up a search in seconds. Yeah, it does. One thing to note on this too, right, it's obviously ThoughtSpot's engine under the hood. It's converting your words, your natural words that kind of make sense to you into a query that goes and gets the job done. You also hit on it very briefly, Jack, but it'll use usage-based ranking to say what other people care about in your organization. I think that's super cool because it gives someone the ability to get started on their own and diving into things that people actually care about, people actually view in your organization. I think that speaks a lot to the culture thing and advancing that. And self-starters—we've just seen a lot of people go farther, and it makes it less intimidating using the product because they build it to kind of learn on itself and grow with you as you grow in your journey with it. Yeah, absolutely. Yeah, so in the same situation or in the same example, right, you have the ability to toggle back and forth between visualizations and an actual table, which is very familiar for users of Excel or those who are used to consuming sort of raw underlying source data. What's great about this is in either format, you can actually drill down into any data point and expose any of the other dimensions in your underlying data model. So in here, if you wanted to drill down—I'm in California, so I'm going to drill down on California. We have this ability to one-click drill down, and we can cut this by any other available dimension in our data model here. So let's just say we wanted to see types of restaurants or restaurant type. Okay, maybe it was a bad example because it's null. But it was easy to get there. And that's kind of the point, right, is like, hey, we'll spend the time getting the data model set up. The tool makes it easy to get to what you're looking for. And that's the name of the game. Make it usable, right? So we have some questions too in the Q&A, just briefly, Jack. Maybe we're running out of time, but maybe do you have anything else to add about the Tableau Ask Data comparison? I know you touched on it a little bit earlier, but anything you want to add onto that and how it kind of compares and is different to ThoughtSpot? Yeah, so kind of going back, I think the primary difference for me is that the search experience within ThoughtSpot is embedded in every part of the platform. And searchability is accessible within one click of any part of ThoughtSpot. And it's part of the content creation cycle as well. So I think that that's probably the biggest distinguishing factor for me is that your heuristic of ThoughtSpot is that you start with search, and that's used to create your data visualizations, your liveboards, that then turn into these other business objects. So you're essentially taking these streams of consciousness, you can build them, act on them, and then pin them, and then that can help create content. Whereas a lot of times when I'm thinking more strategically, I see Ask Data as supporting dashboards rather than using the process of creating data objects. It's usually used as sort of an ad hoc tool on top of or to supplement the dashboard, whereas in the search analytics experience, you can actually use it as a part of your development and your content creation cycle. Yeah, definitely a lot of other differences, but I think that's the one that people need to be aware of for sure, users who have experience with both or are evaluating both. It puts search and it puts answers as the focal point, right, not as the side point in addition to it. It's what you're coming to ThoughtSpot to figure out, right, that answer, the search in your head. That's why you're here and that's why ThoughtSpot exists. A couple other things that I just wanted to share before we close it out. We talked about earlier cloud, on-prem, right, that sort of conversation. I know we demoed a lot of cloud stuff and we talk about a lot of cloud things, but ThoughtSpot does have Embrace connections. Embrace is kind of their terminology to connect on-prem sources, so SQL Server, what have you, all the things. So it definitely has the ability to do it there. And the other thing we always like to talk about is, okay, I want to get started with ThoughtSpot. Think star schema, think good data modeling, right, modern, good, best practice data modeling. That's the way to go, that's the way to get started, kind of rapid start, and go run, have fun, ask us questions. We love to poke around and bug hunt and do all the things. So yeah, Jack, any final words from you there? No, I don't have any. Those are all great questions though. And of course, if you have any other questions that we didn't address, you can hit us up directly, and we're happy to take those offline and answer those. But as we're closing out, best practices—this is all recorded, so you'll be able to access this recording up on our YouTube at some point in the next couple of weeks, right? That's right. Links, good stuff. You'll have it. You'll check out the webinar. You'll see the links. Just check your inbox for it. We do have a poll coming up that we would love to understand what you're coming here for, which we're hoping to get out of it. And again, like Jack says, if we didn't get to all the questions or if you just have another one, there's actually going to be a survey that'll automatically pop up when the webinar ends in your browser. And if you have twelve seconds there, I'd really appreciate it for your feedback so we can get some of those questions and follow up with you personally afterwards, give you those answers, and give you a bit more context around it, right? Not just the answer, but we'd love to have a conversation. So if you loved our session, remember my name's Ty. If you hated it, Jack—his name is Jack, right? So but really, if you have any other questions, feel free to drop them in that survey. I promise you we'll get back to you. We'll scan the chat. And like I said, we don't just want to get back to you with an answer. We want to get back to you with the why behind what we're talking about here and what we believe in because ThoughtSpot certainly makes it easy for us to be believers and advocates. So thank you all for joining us. We're right at 2 PM, so I hope you all have a wonderful afternoon, morning, day, wherever you're at, and we'll talk to you next time. Yeah. Thanks, y'all.

In this InterWorks webinar, Ty Ketchum and Jack Hulbert explored ThoughtSpot’s search analytics capabilities and building effective self-service analytics strategies. They covered the six pillars of successful self-service analytics: governance, scalability, personalization, flexibility, ease of sharing and security. The presenters compared search-driven analytics with traditional dashboarding approaches, demonstrating how ThoughtSpot’s natural language search and GPT-powered Sage feature empower users to query data independently. They provided a framework for identifying and building first use cases, emphasizing proper data modeling, user training and alignment with existing tools. The session concluded with a live demo showcasing ThoughtSpot’s drill-down capabilities and integration with InterWorks’ Curator platform for embedded analytics experiences.

InterWorks uses cookies to allow us to better understand how the site is used. By continuing to use this site, you consent to this policy. Review Policy OK

×

Interworks GmbH
Ratinger Straße 9
40213 Düsseldorf
Germany
Geschäftsführer: Mel Stephenson

Kontaktaufnahme: markus@interworks.eu
Telefon: +49 (0)211 5408 5301

Amtsgericht Düsseldorf HRB 79752
UstldNr: DE 313 353 072

×

Love our blog? You should see our emails. Sign up for our newsletter!