Transcript
Welcome everyone to today's session on ThoughtSpot use cases for search analytics, and I think we're right on time to kick this off.
First of all, a small background info on InterWorks, who we are and what we do. We're a global data consultancy. We have teams all over the world in Europe, we're based in the UK as well, the US, Australia. What we do is we specialize in helping people and various teams succeed with technology. And I really like how purposeful we are with that wording—that people always come before technology at InterWorks.
We are partners with ThoughtSpot, so we love the product from the first time we tried it out, and we're very excited to show you and tell you more about it today.
But what InterWorks does beyond partnerships like we have with ThoughtSpot is really problem solving, and it's always a fun part. As a consultant, I need to explain what is it that I do for a living. So we help you build data solutions, solve business problems, help you with whatever technology you're currently working with, something you've invested in.
And we can help you on the entire end-to-end analytics spectrum from data architecture and management to data prep, to cloud migrations, server administration, and towards the more advanced end of the spectrum like embedded analytics, more advanced analysis, but also training of your teams and a whole lot more.
And today we're focusing on ThoughtSpot, but if you're interested in any of the other technologies and tools that we love, I would only recommend that you head out to our webpage shown here on the slide. And especially our blog, interworks.com/blog.
We are well known for our blog. Vicky can confirm. And when I talk to our clients, they usually bring this up, say how our blog helped them, how much they've learned because of that, and all the new things they can try out.
And yeah, if you'd like, you can also follow us on LinkedIn, stay informed about things that we do and also upcoming events like this one.
And Vicky has been so kind to offer you a couple of housekeeping reminders. Yes, this session will be recorded, and please use the Q&A function for all the questions in Zoom today. And like we mentioned, it will be a bit more engaging and interactive, we hope, and we'll have a couple of polls.
So with that, a round of introductions.
First of all, my name is Darinka. A warm welcome to everyone who has joined me today. I'm an analytics consultant at InterWorks, and I'm based in the very cool, figuratively and not figuratively speaking, Berlin, Germany.
And at InterWorks, I help clients wrangle, get their data into shape, and visualize it and transform it into manageable and useful business insights with tools like Tableau, but also with tools like ThoughtSpot.
And I got interested in ThoughtSpot actually very early, based on my previous interest in natural language processing, NLP. So once I heard, you know, there is this new shiny tool and you can ask it your data questions directly, I was hooked and I knew I had to try it out.
So very excited then to show you ThoughtSpot.
And this is what we have on the agenda for today. We're going to do a quick intro to this tool. And I think it's always exciting for people joining in who have never heard about ThoughtSpot before and are just here to get that first glimpse.
Then we're going to move on to use cases for search analytics in ThoughtSpot, you know, like how to build the first use case for search-driven analytics. How can that look like for teams coming from traditional BI background or environment, and we'll give you some guidance, some tips on how to roll out these first use cases.
And lastly, we're going to show you how this looks like in the wild, and we're going to quickly introduce the platform and what it looks like.
Now before we dive into the actual content, I was curious to know how many of our guests today have heard of ThoughtSpot, like who is that brand new audience. So Vicky, we could pull that up. Thank you so much.
So it's a very simple poll, yes or no. I think it's really interesting. And, you know, it doesn't have to end there. Maybe you could give us some context in the Q&A or even chat. Don't be shy. You know, like, where are you in your ThoughtSpot journey, so to speak, or self-service analytics journey if it's a different tool. Have you heard of it before? Do you want to learn more?
And some first answers are coming in. I see yes is actually much wider, the yes bar than the no bar, which I'm pleased to see.
I'm also very pleased to see that, Darinka, considering that we've got a bit of a bug in the system, and we've got an untitled question. So it looks like the poll was such a good question, it's pulled it through twice. So it does look like for those who have answered the question, I'm going to end poll and share the results.
We've got out of five people who kindly answered, we've got four people who have said yes, they are aware of ThoughtSpot and self-service, and only one who is not. So this could be educational for everyone, I think.
Yes, definitely. But really again, very pleasantly surprised that we have audience who have heard of ThoughtSpot and are interested to learn more. Thank you, Vicky, for that.
And let's move on. And what I wanted to kind of run through is, you know, what are the pillars of a good self-service analytics strategy? And, you know, where can you start? Where should you start as a data organization looking into these tools?
First of all, definitely data governance. When we mentioned the poll, Vicky, I believe we've done a similar poll on LinkedIn recently asking our clients, you know, what is it that you're struggling with most at the moment when it comes to data? And actually, data governance was the top answer.
So it's always something that we kind of, oh, yeah, we know we have to think about it, but it also gets a bit sidetracked and not something that we devote enough attention to. So definitely, before you even start thinking about search-driven analytics, think about your data governance. Is it clear? Is it easy to implement? What are the repeatable patterns around content usage and user administration in your company?
Then definitely scalability. So having a BI infrastructure alone is step number one. But, you know, you might start with one ThoughtSpot user, with two, five, and, you know, scaling from five to ten to really like one thousand users. And as this number grows, and the number of use cases grows as well, you need to think about a good scalability approach.
Personalization, so one of the most important aspects of a self-service strategy is personalization, you know, making sure that insights that we're producing align well with our BI stack or current BI tool that we're using. But also, we need to think about the end user. Is it presenting in the language that they're comfortable working in? What are the preferences that this user has? Is this tool suitable to use? Is it comfortable to use?
And we want to set up the end users for success. Can they completely use the tool? Can they act on any information that they've gathered, and especially through tools like ThoughtSpot?
Then we've got another initial pillar which is flexibility and drill-down. Another common pitfall to watch out for in any search analytics tools is flexibility and drill-downs. We're buying into this idea, oh, I will be able to search through my data and drill down to everything.
First of all, it can be difficult for new users to get the correct format of the information they're looking for from the very start. So on your end, what you need to make sure is, are all the relevant filters applied to this data before the end user actually starts working with it for further meaningful analysis? Is this dataset properly managed? Is data security in place? And we'll mention this also in a bit. For a good self-service strategy, this is really important. So not just thinking, yes, there's flexibility, there's drill-down, but also how you go about it.
Then definitely ease of sharing. This is becoming more and more important as your stakeholder base grows, as your company grows. But it also might be the case, you know, you're adding more BI tools to your ecosystem. And we want to make sure that we're sharing meaningful insights and communicating with that stakeholder base in their preferred communication channels. And also in a preferred format. It differs from team to team, from organization to organization.
But definitely, these insights or whatever you produce in your analytics platform should be portable, should be shareable, and definitely shared outside of that native application. So my colleague, my end user doesn't have to, you know, consult the entire business object that I've created.
And like I mentioned, definitely security. There's no BI strategy without security, and you have to make sure that correct users are accessing the correct data, and we need strict provisions in place for self-service analytics to work.
And we believe that ThoughtSpot kind of hits all of these six tenets or targets, if you will, and you can build your self-service analytics strategy with all of these pieces.
Now a bit of background on, you know, what is ThoughtSpot?
I like to remind people that ThoughtSpot is actually not a new tool. It's been around for quite some time now. It was actually, if you want to say, started or invented in 2012, by people who used to work at Google, Oracle, and other Silicon Valley companies. And you will certainly get this connection once you start using the tool that it's very similar to what you have in Google, Amazon search from your day-to-day.
And if you ask me or my colleagues, at the moment, ThoughtSpot is the platform for self-service and search analytics. It allows me, my team members, and all data analysts to connect to my enterprise data model. And this data model is connected to a cloud data warehouse.
ThoughtSpot does have an on-premise solution, but it is a cloud-first BI platform, just something to mention from the start.
And it definitely allows me to leverage that natural language search that I mentioned. So to start typing in words and keywords to search my business data directly. I can retrieve all of these answers in a visualized format in the form of charts, even crosstabs. I can adjust that, and that can be further shared.
Visualizations can be very easily changed, also embedded in a different application if you do have one, and other apps if you have something connected to our BI stack.
So the overall take here is that ThoughtSpot can be very heavily customized, if you will, to different user needs. You can use, like I mentioned, keywords, adjust synonyms, and even type, you know, GPT-like prompts to ask your data questions. I know ChatGPT has become quite popular recently, but ThoughtSpot is there as well. Recently, they've integrated GPT-like search into their product, and it's called ThoughtSpot Sage.
So definitely natural language search as something that's a default in ThoughtSpot.
Now since I mentioned that this is a cloud-first solution, I wanted to go over some of the supported data connections in ThoughtSpot. Here are some of them like Snowflake, Google BigQuery, Databricks, Oracle. And these are continually updated, and I think new ones are coming in as well.
A common question that I get from our clients is, you know, ThoughtSpot works best as a cloud solution or when combined with a cloud data warehouse, but what if I have on-premise data? Yes, the answer is yes, so it does work as an on-premise solution as well. But it's much more powerful as a cloud version. It's almost built as a cloud-first version, and it leverages the power of these warehouses like Snowflake, like Teradata, even SAP HANA and so on.
And like I mentioned, there are really more and these are continually developed and new ones are introduced.
The main thing to remember here is that many organizations have maybe invested in these data warehouses, and what ThoughtSpot allows you is to have a mask set on top of your data and let your team start working immediately with the underlying data.
And first of all, it's been a small effort setting up a cloud data warehouse, and with ThoughtSpot on top of it, you can really make the most out of your investment and actually use every bit of that data. And we see ThoughtSpot as the tool that allows us to do that easily and at scale.
And I'm just going to quickly turn this off because it's getting in the way of my screen.
So coming back here to the initial question, what is search-driven analytics? What is it? What are we talking about today?
So first of all, we need to think about the key business questions and queries that can be asked in this search-like experience. So if you think about traditional BI, we've got some dashboards, maybe some embedded business objects that we created for the end user, and we base that on a collection of requirements.
So it's usually a pool of business stakeholders who reach out to a data analyst or even a middle person, and this is given to a team of BI developers, so there's a whole cycle of how this works.
What search-driven analytics allows us is, you know, as users, we can be more proactive and ask questions and business queries ourselves and on the fly. So rather than putting a request with your data analyst, with your analytics team, and, you know, wait for this amount of time where they need to scope out how much effort it takes, also lag time between your initial ask and final delivery in your inbox.
What search analytics does is actually it puts data into your own hands. So you can ask questions directly, retrieve any specific answers that you're interested in that day or that month, do whatever you want with it, visualize it, share it further. And, again, in a format that's meaningful to you.
As a second bullet point here, answers to key business questions arrive in the form of visualized data, which means once I type in my question, like, I want to ask data something, the answer will arrive in the form of a chart, a crosstab. I can export this data, this chunk of data. So again, I'm having that flexibility in how the output will look like, and I think that's one of the key features of search-driven analytics.
For those users coming from a strong data visualization background, these answers can come, you know, in the chart form, your well-known KPIs or something very polished. Not as much text—visuals are the most important thing. But then, if you come from a background where you worked with a lot of pivot tables, maybe you were a heavy Excel user, and you appreciate larger tables, you can also curate that, so that's also not an issue.
As number three, we've got answers that can be drilled into, customized, saved, and shared. So it's the important user experience that we mentioned. It's almost a core pillar of self-service analytics. So the option to have it hyper-personalized and meaningful to you, but also shareable to the rest of your team.
And then number four, one of the biggest points from a business intelligence standpoint is having a search analytics strategy and platform. So it eliminates the need for most of those ad hoc SQL requests and data pulls. So if your data is modeled correctly, you can reduce that overload on the data infrastructure team. And I'm very well aware that this is a big one for me coming from a BI space as an analyst, you know, working with large datasets, building dashboards on top of them, and I don't want to overwhelm the data engineering team with too many requests like for this one specific dashboard, could you please do X, Y, and Z. So you can really reduce that overhead for the data engineering team as well.
Now for our second poll today, I was interested to learn, you know, where do other data professionals like me spend most of their time? So if I'm spending time bothering data engineering teams, what is it that you do?
Let's have a look. This is also a very interesting poll to see. Yes, cleaning and modeling data, definitely. I empathize with that. Building dashboards, definitely.
Again, interesting that data governance doesn't get as many hits. It's the one that's always forgotten. I kind of feel sad for data governance.
So far, cleaning and modeling data and then building dashboards. Also, Vicky, maybe you could share in chat? Like, what is it that is bothering you with cleaning and modeling data? What's coming in the way? What's in the way?
Because at least for me, I know when I'm working, once I've got clean datasets, building dashboards on the fly is a very easy thing to do. I think we spend most of our time wrangling and fighting the unclean data.
I do believe there's a second poll at this point, Darinka, that you wanted to share?
Yes, exactly. So based on that data cleaning problem, we wanted to ask, where does your data live? Thanks for the reminder, Vicky. Is it in the cloud? Do you usually work with on-premise data? Are these flat files on your computer like Excel and so on? Or is it all of the above?
I was going to say, I have a question, Darinka. In your experience, would you say that it typically is a mixture of all of those things that you tend to see, or would you say that most people have moved over to cloud warehousing now for all data sources?
Yeah, that's an excellent question. Thank you, Vicky. I see a mix of everything, to be honest, but I see a certain tendency that we're moving towards cloud just because it's easier to work with in the long run. Yes, it's something that you invest in, but you can leverage so much afterwards.
Also for scalability if it's a very large organization. You cannot depend on on-premise or flat files, definitely not. So as your organization grows, as your team grows, you definitely have to think about moving to the cloud, at least part of your workload.
An interesting answer that's coming in. I see that cloud warehouse is almost halfway there, but on-premise databases are still in the lead. All of the above, yes, definitely.
And I'm also very pleased and happy to see that local files didn't get any hits. So something that we as data analysts really dislike is when we get local or random CSV or Excel files to work with. So not an example of good practice.
And another question that we at InterWorks get often is, you know, what if my organization doesn't have a good data model? Is ThoughtSpot going to work if nothing is set up, you know, as neatly as it should be?
And the usual answer that I've given, I think it's the proper answer, is that as with any BI tool, is anything going to work that well if you haven't properly implemented all of these good practices and your data model? And if you haven't set this bare minimum in the first place, all you're going to get is struggle, definitely. And I really like in this way that ThoughtSpot is pushing us towards those best practices and forces us, you know, to hit that checkmark on all of the things that were prerequisites for a successful business intelligence platform.
And actually, I think there are no other answers. Nothing will work well if your data model isn't set up properly. So just again, inherently forcing best practices and reminding us where the pitfalls may be. That's something I really appreciate with ThoughtSpot.
Awesome. Thank you so much for engagement. This is really interesting stuff, especially from the poll.
And Vicky, I believe there was a small question in the chat or a comment. I'm not sure I can see that.
Yeah, sure. So just somebody commenting in regards to, I believe, the dataset for the dashboard or ad hoc requests require multiple joins, requires specific filters. I suspect that's a typical ask for any analytics consultant anywhere.
Yes, yes, definitely. And again, moving a chunk of that, you know, data engineering work to cloud will make your life much, much easier.
Good. And I think at this point, it might be a good idea to kind of delineate or give a side-by-side comparison between search-driven analytics and dashboarding. Most of us, even me, are coming from a traditional dashboarding background. And now comparing them in terms of strategy and tactics, there are some important differences.
First of all, in terms of content ownership and usability. So in the case of search-driven analytics, you know, the end users are actually the creators, the authors, the ones creating all the queries by themselves, hunting for important insights, and they can leverage that user-friendly interface, if possible, by using a type-in search bar, natural language prompts, and using keywords, different synonyms.
And like I mentioned, recently ThoughtSpot has also launched a GPT integration called ThoughtSpot Sage, which we're very excited about, and it allows us to leverage the power of GPT-like search experience and get those needed business insights very quickly.
A very important difference in search analytics is, you know, here, users are creators, but it's not the same with dashboards. In the case of dashboards, BI developers are the ones who are scoping the request from the end user, then exchanging with a pool of stakeholders and owning that entire cyclical process. And then you will get into tactics of dashboard design, like the process, how that works, and so on.
And search analytics on the other hand is there to maximize flexibility, you know, on points like filtering, drill-downs, just giving users the flexibility in exploration of that data. Whereas dashboards are designed to answer a limited set of business questions. These could be very important questions, previously scoped out, outlined, but in search-driven analytics, users can drill down into those attributes and customize their analysis however they like, and ideally, again, leveraging anything that's readily available in that data model that we just talked about.
The content in search analytics is highly personalized, very shareable. And in dashboards, a lot of the user paths are predefined. And these may answer top-level questions, but, you know, from time to time, you will need to dive in deep to understand your numbers better. You may need to request a new view, a new filter, and similar.
And lastly, like we mentioned, ThoughtSpot does have an on-premise version, but primarily it's made to work with very live data in your data warehouse. So you can act on that real-time data. You will have a point-in-time answer and get actionable insights really quickly. On the other hand, dashboards work off of a number of data formats. Some of them are also live data sources, but we're also working with extracts, and these need refresh schedules. Maybe these are also connected to flat files that need updating and so on.
So these would be some of the main differences between the two types of analytics. Again, one is not necessarily better than the other, but some key differences remain.
And at this point, I'd like to run through the usual process of dashboard development and what it looks like. How we create and how we use them and share them with the end users.
And again, speaking for myself as a dashboard BI developer, I can confirm it does look like this. So number one, it's the ideate phase, where we define what we need to work on. We gather data requirements, maybe we get that data model in place, we build out the first wireframes, checking with designers to make sure, you know, this is intuitive for the end user to use. And then we do some additional data discovery. So gather data requirements, get those first wireframes, and then play with data.
Then we move on to the second phase, which is developing—the most fun one, I'd say. We're creating higher-fidelity wireframes here or version one, the MVP of the product. And there is this very laborious process of validating business logic that someone communicated to us. Maybe some of that is not new to us, but some additional requirements may come in. Then we're going to go through the design UX process again, making sure everything's pixel perfect. And, of course, data source development, which can take a large chunk of the development time. Data sources are always an issue.
In the refine step, we are gathering very important user feedback, doing quality checks, exchanging questions on data and completed work so far—very important stage.
And all the way to the launch day where we're training new users how to work with the dashboards that we made, sharing documentation and existing features, you know, socializing, as we say—dashboards—sharing them more widely so maybe we get some additional feedback.
And into the maintenance phase. So definitely even after all this work, after all these steps, additional feedback will come in, and dashboards are almost designed as, you know, it will never fit everyone. So we'll always have users requesting ad hoc reports, you know, custom views of that data, and you'll be collecting requests maybe for bug fixes, ad hoc reporting, and so on.
And this is how we typically look at the dashboard design development process. We spend a lot of time in this loop between the refine and develop phases, you know, validating logic, checking if this was the initial requirement, and coming up with what are the priorities of this dashboard.
And it's hard to complete this phase, especially if your stakeholder base is very large, you have very nuanced business questions. And eventually, there's also a lag maybe between the maintain and develop phase. Again, some end users coming back with additional feedback that you will need to incorporate.
And definitely with time, your inbox can start to look a lot like this. So if it's not managed well, this will increase, you know, that time lag, but also maybe training requirements for the end user. How many ad hoc requests can be answered? You might have to cut some corners there and try to decrease the overheads in your teams.
Now where can ThoughtSpot help? It can definitely help here. And again, one of the main value propositions is related to users. Users can create these ad hoc queries and visuals and leverage that, again, natural language search and create all these charts themselves.
And here's a screenshot of how that might look like. You just start typing in the search bar, something that you definitely remember from websites.
Then we can drill down into our data. It's also a big one for me. And as you can see, you can essentially drill into any part of your data model in one click. So it doesn't have to be like year-month-week hierarchy. It could be also something else.
Then we've got personalized KPIs and, you know, content tailored for each user. When you first log into ThoughtSpot, you're immediately shown this open space. You can experiment. There are no wrong answers. You can delete steps and come back. It's almost also similar to social media. I can see what my team members are most interested in. I will see trending answers and live boards here on the right.
And definitely sharing, it's one of the core features of any good BI platform and definitely ThoughtSpot. So once we've collected all the important insights, how and where can I share them? And again, in ThoughtSpot, there is a wide array of sharing options, even from Slack to other business tools that you might have, like even Salesforce.
And a very cool feature of ThoughtSpot is definitely liveboards. So I can group all of those charts that I created, queries into liveboards. So if I want a more executive-level overview, something to come back to, I could have all of that combined in a liveboard.
Another quick poll, maybe if you'd like to share what's one of your greatest challenges in self-service analytics if you have it in place. We've got data quality and trust, user adoption, data security and compliance, complex data integration.
I can definitely see user adoption as an important one. Data quality and trust.
Yeah, yes. I can definitely agree with the most clicked on right now, which is data quality and trust. Thank you for sharing.
I definitely agree on data quality assurance, and especially if you have one well-working model, yeah, you can reduce that worry.
This kind of launches us into building our first use case for search analytics. So we might have some business objects like dashboards in place, maybe they're not meeting all the needs that we have.
So imagine we're a BI team. We have a dashboard, and it meets all the requirements we fleshed out initially—clear KPIs, there are good visual practices, answers to important questions. So all of these dimensions are in place. So we've also spent a lot of time thinking about the requirements.
And after the launch day, there are also some additional questions from the end user group. Maybe there's one franchise in this restaurant example, one executive who says, I only want to see twenty restaurants with higher costs.
And inevitably, you're going to have more and more of these questions and asks for ad hoc reports. Users are very curious. As people, we're very curious, and we want to drill down into data to understand more. It can be the case that, you know, this is a middle person serving data answers to someone else, creating reports for other teams, acting as the messenger between the two.
And then you have this chain process of how these questions come in, and, you know, the lags and how quickly a data analyst can act on those insights, but also the knowledge gap widens until the end person can get that answer quickly. And with ThoughtSpot, these pain points get a lot easier, actually.
And how does that happen?
So first step in building your use case for search-driven analytics is, you know, first understand before you start building out this use case. What is the user base like? What is the overall cost of investing in the solution, but also what is the cost of not investing? Like, what is it that you're struggling with right now? Maybe users have access to the data but not in the correct format.
Prioritize. Organize your data by data attributes, store them by business department they belong to, adjust this to a user base—is it fifty, one hundred, five hundred people you're serving these answers to? What are the actual use cases coming in?
Step three is to build. And again, assuming you have your business use case defined, data and analyst teams need to make sure that you have a well-working data model in place, security built in, and prep that data model, you know, with built-in keywords, synonyms to those keywords, to make all that data as searchable as possible, and business users can then leverage all of that by typing in, again, natural language form in the search bar.
Step four, definitely test and validate. As you build your worksheets and searchable charts, it's always a good idea to put together a task force of power users and have them test around, make sure the data they need is searchable, connected in that data model you built, and just make sure that you optimize that user experience for them. And what I wanted to mention also is what's good about ThoughtSpot is that testing and user adoption, which I saw was one of your pain points based on the poll, is it's very easy to do with ThoughtSpot. So these iteration and testing loops are quite short. All of the feedback can be brought in almost immediately and implemented quite quickly, which brings you faster to the launch day, right?
So make sure, first of all, your users are trained on how to use search. They're familiar with company synonyms or keywords. Make sure you teach them how can they drill down into different visualizations, save their answers, combine charts into liveboards. How can they share all of that with their colleagues?
And a good tip is to create tutorials, why not, maybe some support technical documentation, just as you would with any BI tool or new application launch.
And like I mentioned, ThoughtSpot lets you track that user adoption. You have a saved list of what are the most searched-for answers or most interesting KPIs. So it also gives you a bit of an overview in your business as well.
Step number six is definitely align. So make sure that that user interface aligns with the rest of the content that you have in your organization. Maybe you're working with existing dashboards, but, you know, why not use the option to give the user that search experience as well? And this can be done with embedded analytics, and what you're seeing right now in this small gif is Curator, which is an embedded analytics ecosystem, and here you can combine multiple BI platforms and basically put a website or a mask on your company's dashboards, important BI objects, and let your users move through all of that content, almost feeling like this is within my company. I don't have to leave this intranet or website to search for other reporting tools or open other reporting tools.
And I think this might be a good time to pause quickly. Vicky, were there any questions in the chat maybe?
We do have a question that's come in. I'll read it. It's quite a long one.
So how does ThoughtSpot identify how to integrate multiple datasets and give the final results as such? For example, there is an order item table which gives revenue at an order item level, but there is another table with product details. If we want to identify revenue for specific products, how does it identify and pull the multiple datasets and create the final data pool?
Yes. Thank you for that question. I actually like this question very much just because I can see you've been thinking about this already. Yes, maybe we are connected to a data warehouse, but there are multiple tables that you can work with, and I'll show you this in a minute.
Within ThoughtSpot itself, you will be firstly creating worksheets that connect to different data sources and connect them on an existing or whatever logic you have in your business. And then allowing your users to search on these connected tables.
So a lot of people ask me as a data analyst, what can I do in ThoughtSpot? What if it's already doing everything for the end user? Well, it's not doing everything. You will be working in the background, connecting all of these tables, figuring out the business logic, even different keywords, adjusting synonyms.
So it's basically pulling all of these threads into one meaningful canvas of data that users can then, you know—this is what I can work on. I can type in my questions.
But again, you will not get an answer if your data tables are not connected. So short answer to the long question is, yes, you will need to connect these data tables first in order to pull in a meaningful answer.
But thank you for that question. Really nice that you're interested in how this works.
And I think this is the most fun, and everyone has been waiting for a quick demo and how this looks like in the wild.
So let me quickly show you.
So one of our visitors today asked, how can I ask this question? How can I connect these data sources? So once I open my ThoughtSpot, you can see the famous search bar.
And here on the left, if I see I'm connected to a data table called InterBurger Run, if I click on there, you will see that I can actually choose which data source I'm connecting to. So it doesn't have to be always the same one.
And again, I'm working with InterBurger Restaurant today. And let me click on that search bar, and you will notice as soon as I click on there, I can actually see some suggestions or recommendations and also ranked on previous usage.
So I can see that net sales is very popular in my team. So if I hover over that or click on there, it will continue giving me recommendations based on that first suggestion. So I've selected a metric, net sales, and then maybe I can start typing in as well. I'm going to say, show me net sales for today.
And again, nice auto-fill option. Once I'm happy with my query, very simple again for the first demo, net sales today, I'm going to hit that Go button, and we'll see it's querying that in the background.
And again, with a very quick answer. Our net sales for today or as of today are 6.7 million, and it was within seconds. And I'm not sure anyone has checked, but if I were an executive, CEO asking this question, maybe I don't need an entire business dashboard for this question. I just want to know what are our net sales as of today, and I can just start typing in that question.
Now if I want to do another search, I could just remove all of this, I could again click undo. Let's go for something more similar to our natural language search. I'm going to say give me top three—and it's again suggesting these as you go. I'm going to say top three franchises for each state. And again, this is the ThoughtSpot version without Sage, the GPT one that I mentioned. This is, let's say, the standard version. But again, it's very similar to that natural language search.
So we're looking for top three franchises for each state, and yeah, why not? Let's give net sales. Let me go again. I'm going to say Go just as if you're inside that Google search environment or even Amazon.
And there's my very nice visualization done within seconds. I didn't have to build complicated calculations or anything like that.
And a nice thing about ThoughtSpot, as I mentioned, is the drill-down option. So if I—which state? I kind of like Texas. If I right-click on Texas and say drill down.
Maybe I'm interested. Why do we have high sales in Texas? So if I drill down on there, as I mentioned, this will not give you a limited set of options. It will give you many, many dimensions that you can use to cut through your data.
And let's see what I can use. If I start typing in, maybe by city, I want to see top cities in the state of Texas where we're making most money, and it seems to be Austin—total sales above nine million.
And again, as I mentioned, for those of you who are used to crosstabs or are interested to see individual numbers, it doesn't have to be presented as a chart. I can click on here, save as table.
And this gives me a very clear understanding, yes, Austin. I didn't know there was an Athens in Texas. Interesting. You know, how much money are we making by city if I'm an executive? I just want a quick glance, okay, which are the top cities? Maybe I really want to see those numbers.
So, yeah, I hope that you found that interesting. You can definitely save all of these charts or searches that you have in the form of a liveboard.
So here I've got a flight data liveboard that one of my colleagues created. You could save a couple of KPIs that you constantly return back to, maybe even a nice Sankey chart, which are difficult to build, I cannot recommend.
And yeah, have that as a starting point before you start again, you know, drilling down into any of these charts if you have further business questions.
We have one final question, Darinka. We're just on the cusp of running over.
So say, for example, there are multiple analysts creating worksheets for identifying the same KPI. If we have multiple worksheets created for searching the same KPI, will this confuse the end user which one to use and which one is the most reliable?
You would ideally have one worksheet for that. I'm not sure what's the actual use case that you're talking about. But again, maybe coming back to what are the top issues that we come across is that, you know, data management, why would you have multiple worksheets or multiple versions of one thing?
So maybe first think about that. Ideally it should be one centralized thing that you use and then that your users can search upon.
And maybe finally now to wrap the session up is just to again repeat what are some of the best practices for building that first search analytics use case. Definitely understand your users, what are their business needs, what ad hoc queries are coming in, and organize. So, you know, you will need to do a pre-brainstorming session with your users. What is the most common business need? Definitely test and validate—the cyclical nature of the entire process. And finally, align your analytics. Maybe you don't have to give up your existing BI tools. Maybe ThoughtSpot is something that you can use next to it.
Good. I think that brings us to an end.
Here are some helpful resources and links. If you're interested, definitely check out our blog again. We've started a ThoughtSpot 101 blog series. There's a lot of interesting content there. And yeah, if you have any questions, feel free to reach out. Thank you.
Thank you so much, Darinka. Thank you everybody for joining us today. We will be getting a copy of the webinar through to you. And as Darinka said, if you do have any further questions, please feel free to reach out to us. We're happy to discuss them further.
Thank you, Darinka, for presenting today. Much appreciated.
Yes. Thank you, Vicky, for your support. Have a lovely day ahead. Thank you, everyone.