The Data Stack Show - 87: Why Is Now the Golden Age of Data Analytics? With Cindi Howson of ThoughtSpot

Episode Date: May 18, 2022

Highlights from this week’s conversation include:Cindi’s career journey (2:36)Major shifts in analytics (6:34)Where we are in formation of the modern analytics cloud (9:07)The process of moving into the cloud (11:01)How to accelerate the digital transformation (17:29)Common patterns amongst company cultures (19:42)Data regulations affecting change (22:34)ThoughtSpot customer base (24:06)The need to know SQL (27:42)Power users leveraging the AI Insights (31:24)Who should audit technology (32:33)The ways that education is happening best (36:28)Stuck in descriptive analytics (40:43)Changes in company structure (43:54)Defining an analytics engineer (46:57)The impact on IT as a function (50:33)Enjoying data analytics (53:06)The Data Stack Show is a weekly podcast powered by RudderStack, the CDP for developers. Each week we’ll talk to data engineers, analysts, and data scientists about their experience around building and maintaining data infrastructure, delivering data and data products, and driving better outcomes across their businesses with data.RudderStack helps businesses make the most out of their customer data while ensuring data privacy and security. To learn more about RudderStack visit rudderstack.com.

Welcome to the Data Stack Show. Each week, we explore the world of data by talking to the people shaping its future. You'll learn about new data technology and trends and how data teams and processes are run at top companies. The Data Stack Show is brought to you by Rudderstack, the CDP for developers. You can learn more at rudderstack.com. Welcome to the Data Stack Show. Today, we are going to chat with Cindy from ThoughtSpot, and she is the Chief Data Strategy Officer. She helps their customers with their data strategies, so she gets to see all sorts of interesting ways

that companies are using data. And she is actually, as we'll find out, a user of the ThoughtSpot product herself. So she knows all the ins and outs. Costas, what I want to know, and this is just a personal suspicion I have, partially based on experience and a lot based on conjecture, but I think a lot of companies still live in what I would call sort of relative to the options available, like a primitive world of analytics, where you get the you sort of, you get the basic reports that allow you to run the business and have a sense of what's going on. But it's pretty hard to move past that. But that's based on my own limited view. And Cindy, you know, has worked

in data analytics for over two decades. And so I think she'll be able to help us see like, is that actually the case? So that's more of my personal burning question. How about you? Yeah, I think Cindy is like the expert when it comes to digital transformation and transformation in the organization that has to do with the use of data. And she has seen this happening from starting from like the mainframe

and like doing analytics on the mainframe to the era of cloud that we live today. So I think there's a lot to learn from here on what it means to adapt and use your data and what it takes to do that as part of an organization. So yeah, I'm really looking forward to chat with you. I think today it won't be that technical of a conversation, but I think it's going to be a very, very unique conversation.

I agree. Well, let's dig in. Yeah, let's do it. Cindy, welcome to the show. We are so excited to talk about all things analytics and even more with you today. Thank you, Eric. My favorite topic. Right. Well, could you just tell us, you've had such an interesting journey in analytics because you really were doing the actual work of analytics when it sort of became a real thing, you know, in terms of technology. So can you start

there and just sort of tell us about your journey and what you're doing today? Sure. So I have to take you back more than 25 years ago now in Switzerland, and I was working for Dow Chemical. I was known as the Lotus 1-2-3 macro queen at the time and then writing focus reports on a mainframe computer and you know lots of detours or let's say different paths in between worked for Deloitte and Touche for a little while in Texas. They were just then starting their BI and analytics practice. So you might have to think back to why 2K was the big emphasis, not data, not data warehousing. So we were starting that practice. And then I started my own company, the BI Scorecard, which there were two emphases, there were two angles to this business. One was comparing BI products, hands-on, kind of the consumer reports of BI. And the other was helping organizations have a bigger impact with data. I licensed the rights to that research to Gartner and joined Gartner in 2015 and then joined ThoughtSpot coming up on three years now in 2019. Wow. Okay. I just have to ask, is there a particular report that sticks out in your mind from your time at Dow Chemical, you know, that was like particularly challenging or surprising?

Yeah, because it was the first, you remember the beginning and the end, I guess, but it was the first one where my manager at that time was asking for plant capacities. And we felt like it was not possible because our product codes were in one database and capacities were in another. And they didn't think it was possible. They thought I would have to download everything into spreadsheets, merge it together, and it would be a one-off analysis. I'm like, no, I think, you know, we can merge these different tables in the different databases. And he was staggered

that we did it. It also stands out in my mind because as we iterated on this, really the question he was asking was not just what are the capacities, but it was the why. And it was because there were some cracks in a new facility we were bringing online and he wanted to take it offline for a period for some maintenance. And he couldn't do it in time. The cracker exploded, nobody was hurt. But I go back to that because I think Bob Lee was the manager. And I was like, if Bob could have asked his own questions, like you can in ThoughtSpot, we would have been able to just operate faster. And so I just look at how the technology has changed, but the problem of letting the

business people ask their own questions, that problem still persists. Fascinating. I want to dig into that, but let's just touch base on the technology side of it. So one thing that I would love your perspective on is you have been working in and researching analytics tools, you know, since the beginning, right? And I think in many ways, like the, you were, you had a front row seat to sort of the explosion of analytics tools, right? I mean, there are just so many out there. There's tons of different ways to do things. Could you just walk us through, what are some of the major shifts you've seen over the years in terms of major sort of milestones in analytics and in the technology specifically?

Yeah, so I think of it really as four chapters. Now, the first chapter really was about report-based analytics without generating SQL or coding in SQL. And this was the era of business objects and Cognos. Now, when there were acquisitions, so IBM acquired Cognos, SAP acquired business objects, Oracle acquired Hyperion and Siebel Analytics, all in the 2007-2008 timeframe. BI was elevated then because of the scale of some of these deployments, but it was very report-centric. The second generation was really about visual-based data discovery. And this was pioneered by Qlik and Spotfire.

Tableau came along and made it easier and outpaced both those vendors. So that was the second generation where it was about the visualization and empowering the analyst. The third generation is really augmented analytics or at ThoughtSpot, we call it search and AI. So it's about using search, natural language processing, but then also AI to generate insights that you didn't know to look for. And the fourth generation is really the modern analytics cloud, still with augmented analytics, but never having to move your data powered by a cloud ecosystem, whether it's the data platforms like Snowflake, Google BigQuery,

or it's about the data science as well, but all in a cloud ecosystem. So finally, customers can have best of breed capabilities without moving the data. So you mentioned that we have like four stages and we are, from what I understand, already in the fourth one, right? Like the modern analytics cloud is like other formation right now. So at what stage of this stage in terms of maturing we are, like how you have seen companies embracing like the clouds and what has been built and exists out there and what remains to be built and figured out? Yeah. So really, you can't do analytics without your data and you can't get to granular analytics without the scale that the cloud offers. So if I think about where was cloud data platforms, if we go back to say, you know, a few years ago,

less than 10% of the data was in the cloud. The pandemic really has accelerated every organization's migration to the cloud because it accelerated digital. As soon as you're digitizing all your customer interactions, then you get into the volume of cloud. It also is the growth of these cloud data platforms. I mean, Snowflake last year, the biggest IPO period in the software industry. So I think it gives that credibility and comfort that cloud is the way to go. And there's predictions out there that by like 2023, more than 80% of the data will now be in the cloud. We'll still have some on-premises data stores that people just won't go back and replatform, but that's a huge change from just 10% a few years ago. Oh, yeah, that's incredible, actually, from 10% to 80%.

That's crazy. And I know that ThoughtSpot started as a company that was very enterprise-focused, right, with a lot of on-prem installations and did the migration to the cloud. And this happened in, I don't know, like two or three years. So it's not like a lot of time, right? How you have experienced like all these companies that you were working with them

like three years ago and they had like these on-prem installations going to the cloud, like how the process of moving into the clouds has happened and has happened like also so fast, right? Like it's... Yes, it has happened fast. So sometimes I feel like I'm watching a movie. I feel like I'm watching an action movie. And I do think what ThoughtSpot has accomplished in two years eventually will be written up

as a Harvard Business Review case study because I look at some organizations, let's say Adobe, for example, and their transformation or even MongoDB and how long it took and how quickly we were able to shift really. And there were two big shifts. So as I mentioned, you can't do analytics without the data. And our customers were asking us to not, we started also leveraging our own highly scalable distributed in-memory engine Falcon. That was the first generation of the platform. But then our customers were asking us, well, as we implement Redshift or Snowflake or Google BigQuery, we don't want to move the data again. So help us connect to that. And so we released what we call Embrace Mode in early 2020. And that was the first part of the journey to the cloud. Now,

many of our customers were running already in a private cloud, our software. So that was also part of it. So here we are, January 31st, 2020, we released these four connectors in embrace mode and the pandemic was just starting. So think back to that time. And we knew and saw that organizations that had already started with ThoughtSpot were handling this chaos of the pandemic better. They were handling their supply chains, their working capital better. And we said, well, how do we enable this for organizations of all sizes? And for that, we needed to go to a full SaaS offering, software as a service.

So ThoughtSpot Cloud, we released early in June of 2020, had some beta customers there, and then fully launched it September of that same year. And now cloud accounts for, I want to say, 85% of our new customers and now 60% of our annual recurring revenues, roughly. So that is a huge shift. It's fast. And I think a lot of it is that we had the foundations already there for scale. But then I also think it's the mindset of this team that our customers needed us to do this. Yeah. Okay. I have to say that I cannot wait for our business review to come up with the case study.

So I have to ask you, let's say, built on a very focused way, like with very specific go-to-market motions that they have to do with like on-prem installations. Everything is like different when you go to the cloud. And I'm not talking about the technology here, right? Like this is probably to be honest, like from my experience, it's hard, but it's the most easy to manage as a transformation process at the end, because the engineers know what they have to do and how long it will take to move from one to the other. But changing the culture of the organization, like talking to your sales teams and like tell them that like now we do something different, changing the language, the position for marketing. How can you do that and how you can do that scale and show past? Like, what's the secret there? Yeah, so there's a couple parts to this, Costas.

I think, first off, the founders and the executive team make the hard decisions. Our mission has stayed the same, to create a fact-driven world. It's really just the distribution of the software that has changed. So if we can accelerate that distribution, then people are all in. But it has taken hard choices. Now, I don't think the problem is internally in ThoughtSpot because we are a tech provider. What is harder is our customers and the culture there. They may not want to move as fast as we move. Although now I see when I connect our customers, they want to learn from the digital natives because the digital natives are very quickly usurping and disrupting what risks becoming legacy competitors. And there's classic examples. You have Hulu disrupting Blockbuster. You have large banks that have now digital-only banks threatening them. So culture is a big barrier

to transformation. There's research from Randy Bean, the author of Fail Fast, Learn Faster, that 92% of the organizations he surveys say that it's the combination of culture, people, and process that are the bigger barriers to being data-driven than technology. That makes a lot of sense. And how a vendor can affect that culture that the customer has, right? What ThoughtSpot can do to help, I don't know, the Bank of Americas out there or big institutions, huge and complex processes embrace this change and like accelerate the digital transformation. Yeah. So there are some best practices here. One is start small and align to a particular pain or opportunity and take that as your proof point for the new operating model of where you want to get to. Emphasize the why, not the how.

The technology and cloud, that's the how. But the why, maybe for a banking customer, it might be delighting your customers and getting an increased share of wallet. So aligned to the why from a business viewpoint and the technology is the how. Now with that, we also do focus on changing mindsets and behaviors. We do that through things like the Data Chief podcast. We have a Data Chief Slack community. We do executive roundtables and lunch and learns so that we are helping with some of that cross-pollination between the leaders and the laggards. That's very interesting because you make me feel like you have access to the way that these organizations are thinking and working on a very high level, which is not easy to get access to. And I'd like to ask you based on that, from interacting with all these organizations and the people behind them, right? What's like, have you managed like to identify some patterns in these cultures that you see

out there? Because obviously like each company has their own culture, right? Like that's for sure. But there has to be like some common patterns that we see out there that we can probably identify them either as like, let's say, a common obstacle that we see out there to embrace the digital transformation, or in some other cases, also like something positive or something that in some cases, like it accelerates the process. So what's your experience there?

What have you seen? Yeah, so if I look at organizations, so I often say culture and technology is two sides of the same proverbial coin. And somebody said to me, it's a very expensive coin. But if I look at a company that has a legacy data stack, so they're not in the cloud yet, they're still either just based on that second wave or even worse, the first wave report centric, then I will see a culture of complacency, settling for the status quo, siloed thinking, and resistance to change, lack of leadership. If I look at organizations that have embraced cloud and the modern analytics stack, I see a high degree of experimentation and they don't see

risk-taking or failures as failures. They see it as learning and rapid prototyping. And there is a high degree of trust between the data team, IT, and other functions lines of business. They're very much purpose-driven with a can-do attitude. Yeah. Can you give us an example of like a company that you think that they are doing like a really good job in like embracing change and experimentation? Something that we could get, you know, get inspired by what? Yeah. So I think of some of the Hallmark customers that I have the opportunity to work with, and they are public references. I look at the likes of Verizon or Medtronic in medical devices.

But then I also look at some of the digital natives like Cloud Academy. They're all about upskilling. Even bagel brands. I'm thinking food now. Bagel brands, Einstein bagels, they're doing a lot of innovative things. And you reference large banks i mean they're they're not all slow thinkers like bank of new york melon is doing some innovative things and has standardized on thought spot as well just because you mentioned you mentioned banks and

banks and like there are like some specific industries out there that regulation is also like very important around data right yes so how do you see that as a factor that affects change i mean obviously like 10 years ago the landscape out there like the legal landscape around like how we work with data was like completely different than it is today and probably it's also going to change again right like we are not we're not done yet. So how do you see regulations playing a role in the adoption of the cloud's native mindset, let's say? Well, regulation is not saying don't have your data in the cloud. It's that regulation, you don't get to do away with your responsibilities for making sure it is protected, encrypted, whether at rest or in transit, and then respecting the privacy of individuals. So the fear of the cloud there, I think the cloud is more secure unless you actually have your own data

center that is physically guarded and you're always on the latest levels of encryption. So regulation, I think, is often used as an excuse. There are some constraints related to this, particularly with respect to privacy, but even highly regulated industries, pharmaceuticals, for example, life sciences, I'm still seeing them embrace cloud and modern analytics. Okay, that's super interesting. Let's talk a little bit about the people that are using ThoughtSpot, right? Because we've talked so far about the collective side of things and culture, but culture comes from the people who work there, right? And who is the main user of ThoughtSpot in the organization today? Yeah.

So our North Star is the non-analyst, the actual business person who has the questions. And the analysts historically have been a bottleneck between the decision maker and the data. So we want to elevate the analyst so that they can work on more high value, high impact analytics, not doing silly things like spending hours formatting a dashboard or adding a sort button or worse, formatting something in PowerPoint. As one leader said to me, PowerPoint is where data goes to die. So we don't want that analyst bogged down by the drudgery of the backlog of requests. So we want them to help the business user ask their questions, optimize their questions. So it's empowering the analyst to do more with less, but actually have everyone ask their own questions. And how is AI in Maryland's natural language processing important?

Yeah, so AI is infused throughout the platform. The most obvious thing will be AI generated insights. So telling you maybe what you didn't know to ask. So just because Eric and I were talking about coffee, maybe an AI generated insight would be, is there a certain demographic that is more likely to drink their coffee black? And it would tell you what the outliers were rather than the demographic that likes all the caramel and whipped cream on it. So the AI will generate insights, but AI will also nudge you in terms of saying, most other people will filter their search by this

particular attribute. Have you looked at that? It will also choose the best fit visualization. It will give you things like trending live boards, trending content. So those are more the subtle forms of AI that are baked into the product. That's super interesting. And I mean, I think that our industry, like since its inception, I mean, is trying to do exactly that, right? Like to empower non-technical people to ask questions to the data and get like the answers that they are looking for. I think like a very good example of that's like SQL itself as a language, right? Like it was never like created as a language that was supposed to be used by only engineers or like technical people. Actually, I can think of many engineers that they don't know.

I feel like that's the one language I learned that's been timeless. It is, it is, it is. And today we are talking more and more about it. But yeah, it's not like you don't have like to know SQL to become like a front-end engineer, for example, right? So, and even then, it's like a very different, let's say, thing of like using SQL to ask analytic related questions and using SQL to build like an application, for example, right? And like create records and things like this. So how close are we because to ask their questions without the need of like even probably knowing like SQL. But have we achieved this?

Let's say vision that like exists in the industry for the past, I don't know, like 30 years probably, maybe more? Or we still have work to do there and there are like still problems that we have to solve. Like, how do you feel about this? Well, as an industry, of course, there is work still to be done. And in ThoughtSpot, we say we are only ever 2% done because we really want to make it easier and more pervasive.

So if I look at some customers and that have been on this journey longer with us, I'll take Schneider Electric, for example, you know, there, and this is where we were talking about best practices aligned to a particular mission or vision, the why, and then ThoughtSpot is the how, or the technology is the how. Well, Schneider Electric really has a mission to ensure all people have access to energy. It's a basic human right. So part of their goal in rolling out ThoughtSpot was to free up people's energy and to enable, they started mainly with a people analytics use case to identify top talent to make sure that they retained the top talent. And their adoption rate is 78%.

Those are not analysts. These are people managers, HR professionals. And if you think the industry average is about 25%, and that's mainly power users. So individual organizations, I would say, are there. But as an industry, we're nowhere close. That's fascinating to me. And I actually want to... So those numbers are pretty wild. So 25% adoption mainly by power users. I actually want to circle back and ask a question about the AI sort of enabled analytics or insights, of course, in ThoughtSpot, but just in general. And I'm saying this in part because it's really interesting to hear about, you know, are the power users leveraging the AI insights is one question that comes to mind. But I think

even beyond that, and I'm speaking a little bit from just personal experience here, as AI assistance and analytics has sort of become more common, even in a tool, say like Google Analytics, right? Where now you're getting served, oh, your number of sessions from this particular channel is lower than last week, right? A lot of times I will just dismiss those, right? Because it's like, okay, I kind of know what I want, and I don't really know what's going on under the hood there. They're producing some sort of insight and they don't really understand my business context or the questions that I'm asking. And so this is complete conjecture, but it wouldn't surprise me if there's a low rate of adoption for AI-assisted analytics because

especially the power users, on some level, they're like, I don't trust this as much as I trust myself or that I trust my SQL. I know I can get to the right answer. Do I trust this machine to like, what if they give me the wrong number? That could really screw things up. So I just love your perspective on that. Yeah. So Eric, it's interesting because designing for openness and trust are some of our key design tenants and you're not going to trust black box AI. So what we give you is full transparency into seeing what was the SQL generated, what were the algorithms used, and even what were the inputs to that model. And you can control, maybe there's certain attributes, or if you're a data scientist, features that you don't want to be input into that model. So really, without that transparency, it can create noise and we don't want noise.

Yeah, super interesting. Yeah, that's fascinating. Eric, I have a question here. So that's like very interesting what you just said right now. So you give like all the auditing trails that are needed there for someone to go and audit what's going on. But who's responsible for that? Who's going to audit the AI?

Because the power user, I mean, and that's like the interesting part here, right? Like we create abstractions so we can let the people that do not have like the technical knowledge to go and use the technology. But there are times where we have to audit the technology. So who inside the organization should do that? This is where we see the role of the analyst shifting from just a drudgery dead-end dashboard developer to what we call the analyst of the future.

And we're seeing an emerging role as well for the analytics engineer. But this would be the analyst who is optimizing the AI or looking at the AI-generated insights. But they don't have to code it. If they want to take something and pass it back into a full data science platform, they can do that. But we don't want it to be a blank canvas. We want to give them that starting point. That's super interesting.

I love this, how empowering, let's say, the end user or the non-technical user also transforms the existing role. So it's not like the roles get obsolete. They just change and have a different focus. Yeah, not obsolete. No way. And look at our labor market. The need for talent in this space is so tight. I was talking with a customer last week, a new customer, and he, I think I can share, he was a pricing manager and he said, I lost my analyst. He's like, I'm not a data person. I just need to figure out this pricing and the prices that we're setting. And he could teach himself ThoughtSpot. And he said, thank goodness I can, because we haven't been able to fill this wreck for the analysts that we need.

So we have to let the analysts stop the lower value work, teach other people to fish, and that frees up the time for the analyst to get to the new data sources, ingest them faster, but then also work on these AI optimizations. Cindy, one question in there, because I was actually thinking about this and you just said it, the education piece of it. And I'll use a specific example, maybe that will resonate, you know, with some of our listeners who have worked on the analytics side. So I love the idea of this self-serve, you know, let's let the decision maker, you know, let's empower them, right? And in many ways, I totally agree with you. Like, let's remove the drudgery from the whole data engineering analyst role, right?

Because it is just, I mean, report building can be brutal. Cleaning it up, redoing the visualizations 20 times, all that sort of stuff. It's like, great. If we can get rid of that, everyone will be happier. But there is an educational component, right? So if I think about, I'll just use a specific example. We were working just on our team on some cohort reporting, both rearward facing and then sort of forward looking forecasts. And those can be incredibly helpful. But if you're not familiar with it, or you haven't looked at them a lot it can be a little bit disorienting right it's kind of like okay you know you say okay i need to think about like this isn't you know point in time reporting you know on a on a timeline you know for everything

it's like related to just a very specific subset of users you know and sort of you know there's time decay elements and all that sort of stuff what are the ways that you know, and sort of, you know, there's a time decay elements and all that sort of stuff. What are the ways that you've seen that sort of education happen best? So you just mentioned the analysts sort of moving from maybe just a report building type role to actually actively educating. Do you see that in the products, you know, ThoughtSpot, other products as well? Is it the responsibility of the analyst? Is it a culture-wide thing? It's all of the above, Eric. more finer grained analytics, whether it's clustering or cohorts and what have you. So what we have to do is make the technology so easy that we can shift the focus to the data. And what is it really telling us?

So I, you know, I'm a daughter of a DJ here, so I use music analogies. And I think about how you create a Spotify playlist today. Did you watch a YouTube video to learn how to do that? Did you read a manual? Did you go sit in a training room class? You learned it largely yourself. Now, I'm going to suspect my college-age kids probably gave me a few tips to get going. But I think of my father, the way he would create his playlists and the sound system, you really were a professional. The same thing has to happen with data and analytics. So the analyst may teach the business user the starting point. So give them a live board as a starting point and they might become a coach or somebody. I love this term. Somebody used the word a data Sherpa.

So kind of a guide, just a quick half hour. Here's how you get started. You might follow up with some lunch and learns and then more sophisticated things are baked into the product. Let me show you your data. Let me show you how to explore this cohort, for example. But teaching the language of the business, then this is really data fluency. And that has to be ongoing at all levels. Sure. Yeah, I was talking with someone who worked on the analytics team at Swift and

they're doing some really interesting things, but it was so interesting. They said, we do monthly lunch and learns with our marketing and sales teams. And so we actually push reports and then we'll do a lunch and learn and we'll walk through it and they can ask questions. But what struck me about that was, I was like, wow, that's so simple, but so great. And then I thought, it's so rare, right? I mean, it just seems like so few companies actually do that. Would you say that's true? I don't think that's true anymore.

Or every customer that I work with, they are doing lunch and learns or data and donuts, things like this, ThoughtSpot Thursdays. So data fluency even now is a boardroom conversation for most organizations and, or let's say for most forward-thinking organizations. Sure. When it's not, I would be concerned about those organizations. Yeah. Yeah. No, that's really encouraging. Similar question just around analytics. So if we think about the analyst who's just doing report building, I think in that environment, a lot of times you are kind of limited to sort of your basic analytics, right? I mean, you have these

like unbelievably powerful tools and you're doing the most basic, you know, sort of four or five reports, you know, that you kind of use to manage the, you know, the couple of KPIs that drive the business, right? But there's this whole world available. Do you see a distribution of that? I mean, I know ThoughtSpot customers, it sounds like they're a little bit more forward-thinking. Industry-wide though, do you see a lot of companies that are sort of using powerful analytics technology, but are still sort of in the kiddie pool, as it were, in terms of the analytics that they're running? So just stuck in descriptive analytics. So I think of these tiers, descriptive, what's happening, and that's report

centric, diagnostic, exploring the why, and then predictive, you know, what's likely to happen. I actually don't like the term predictive or it's too confined because I think about segmentation, that is an advanced analytics. And then really prescriptive, what is the action I should take based on this insight? And so we do know that many are stuck in that very basic descriptive reporting. But I think this is where, again, the best companies, they are well on their way to diagnostic and predictive. Got it. Super interesting. Yeah. It's really encouraging to hear, actually, we all have our own sort of little slices of the world that we see. And it's really encouraging to

hear sort of in your wide purview that there is a big change happening in the industry and that a lot of these problems sort of seem like, okay, we're stuck in descriptive analytics for years and now organizations are breaking through that. Yeah. And maybe, so because you say the lens or the bias that we view things through. So I think it's important to dwell on that for a moment. One of the reasons why I joined ThoughtSpot is when I was at Gartner and I would do customer reference calls, it felt like I was always talking to the more innovative companies that were ThoughtSpot customers. And so I do think, yeah, my lens now is very much the leaders and they're also creating new markets. So another customer from Europe, now also part of the U.S. brand, Just Eat Takeaway.

Food delivery was a niche segment three years ago. Mm-hmm. Now look at the volume of data you're creating from your app and then powering restaurants to tell them what people are ordering. That has disrupted whole markets. And thank goodness we've had that in the last few years. So I do think my lens skews more towards the leaders, but that's the natural evolution of things as well. I think the ones that stay stuck with just reporting, they're not going to be able to keep up. Cindy, I'd like to ask you to share a little bit of your knowledge around the changes that

are happening in the structure of the companies because of this transformation that is happening right now. You, for example, you have the title Chief Data Officer in ThoughtSpot, right? So what is the role of the Chief Data Officer? And also, what other roles roles you see emerging or changing or getting transformed? We mentioned already the data analyst, how from being, let's say,

the person who is going to build and maintain dashboards becoming like something different to support this new environment, new reality. But what other roles do you see changing? And if you feel like something else is going to be introduced also? Yeah. And to clarify, so my official title is Chief Data Strategy Officer. Oh, okay.

I work with our customers on their strategy. So minor detail, but this is the CDO role within many organizations now, I think is, I want to say it's well-established, but it continues to evolve because the first generation CDOs were really the data guardians and gatekeepers. And as they've continued to evolve to really how do we apply the analytics for business value, their roles have been elevated. I've seen some evolution of the title as well, Chief Data and Analytics Officer. So I think at senior levels, it's evolving. And where these roles may have started in IT, I increasingly see them reporting to the chief digital officer or to the CEO directly.

That's at the senior level. Within, let's say, the contributor roles, one of the predictions I wrote this year is that the analytics engineer will replace the data scientist as the world's sexiest job. Tom Davenport and DJ Patel came up with the sexiest framing of the data scientist. But, you know, data science was slowly losing its luster because they're not working on high value things. Many of the models they build are never operationalized and they're done off to the side. A lot of the work is about data quality, data cleansing. And many of the data scientists lacked the domain expertise to apply the models or even build the models. So I think this analytics engineer has the domain expertise. They are

well-versed in cloud technology, so they're able to work in a more agile way. And it's the delight of working with newer technologies as well, that you can cleanse, transform the data, operationalize some of these things all in an open ecosystem. So how you would define an analytics engineer? What's the background of this person? Where they come from? Yeah, I mean, they might have evolved from the role of a data scientist, or they might have evolved from the role of the analyst who is looking to have a bigger impact. And how do you see the more traditional engineering roles, like, relate to that, like data engineers, for example.

Well, even data engineer, I would say, is still a relatively new role. If we were talking in an on-premises world, then we might be talking about ETL developers, whereas data engineers, now you've kind of shifted the order of things, extract, load, and transform. And so the data engineers are responsible for these pipelines, but then the analytics engineer, they might also have some of their own pipelines, but they also know the domain and their end goal is not moving the data into a data lake house, but their goal is really about creating the analytics. So another role where some of these things come together is the data product owner. And this is where people get very afraid or passionate because we talk about concepts like the data mesh and people are like, wait a minute, what about master data? This is going to be chaos. And yet we have to go down

this route. The more we have conversations about it, we'll tease out what are the problems and pitfalls to avoid. But for sure, the idea of building one centralized data warehouse and one centralized IT team is going to do this and it's going to take six to nine months. That's not the pace of business these days. Yeah, yeah. I have a feeling that like we used to say during like, let's say the past decade that I call that decade the SaaS decade because it was all about like building SaaS and moving like all these into the cloud, all the applications. And we used to say that like every company is going to become like a software company,

right? And I have a feeling that the next step is going to be every company is going to be a data company. Yes, I agree. We agree. Yes. Or let's say, well, every company, data powers every company. And data is now part of everyone's job. So this is where the workforce is going and where industry is going. Somebody was asking me, what is the digital economy? And I said,

the digital economy is the economy and you don't have a digital business operating well, unless you are leveraging the data that is generated by these digital interactions. If you think financial services, the days of knowing your personal bank teller down the street, you don't know them. You only know your customers through the interactions on an app for the most part. 100%. I totally agree.

That's like a very, very interesting point. One last question from me, and then I'll give the stage to Eric. Just as a continuation of the roles that we talked so far and the changes that are happening there, what's the impact on IT as a function in the company of all these changes? Because we used to have IT as, let's say, the function that takes care and controls everything that has to do with the technology inside the company.

But it seems that like things are like radically changing. So where is IT going as a function in the organization? Yeah, so Kostas, I'm a little nervous to answer this because like some people will get angry and upset. But there was a very provocative article in the Wall Street Journal. Is it time to get rid of the IT department? And at first I looked at it, I was like, wow, somebody's just, you know, this is clickbait. But when you read it, the arguments were very well presented. And it's more the concept of a centralized model was designed for an on-premises world where you had to buy a big mainframe or a big server or what have you and operate that. So I think we need technologists,

but those technologists may not be only centralized. I think it's really about when do you have a federated model and you federate to get domain expertise. So if you think about even going back to a transactional application, maybe you're buying your CRM in the cloud and it's that domain person that will decide what's the best CRM, but then you're going to look for economies of scale in terms of when you want to centralize something. So I think it's still about, you want centralization for some governance and career development,

but every domain will have a technology aspect to it. Yeah. And I guess at the end, like as every other function, the compound goes through transformation. So IT has to do that, right? And like in both, right? So make still the same.

Thank you so much, Eric, all yours. Yes, well, I think we're close to time here. I just have one additional question for you, Cindy. And maybe I was going to say this is a personal question, but maybe it's actually not. Do you still enjoy digging into the numbers and doing analysis? I know that you run a lot of strategy, but I also know even just on a personal level, it's not always a great use of my time to sort of dig into reports, but I still love doing that for some reason. And just with someone with your sort of history and experience, I'm just interested to know if you still enjoy doing that.

Oh, absolutely. I do. Even this morning, actually, there was a new feature that we released on the platform. I'm like, oh, let me try that. And then it was like, oh, which data set do I want to do this with that I actually am allowed to share? So as a business, I look at our customer satisfaction numbers. I look at our interactions. I look at our sales. I also look at the diversity that we have in our organization as well as in our industry. So my husband told me he had to ban the word data on our anniversary. He's like, for one day, can we just not hear that word data?

And then I think we had the news on and data was on the headline. So I think I still had the last word there. That's so great. No, that's really fun. And it's just fun to hear that, you know, even though you've sort of done so many interesting things with research and strategy, that you still have a love for digging in. Well, the other exciting thing is, Eric, so having done this for 25 years, I feel like it's now, finally, this will be the defining decade for data. It's not some afterthought. It's really a forethought and it powers business. And then it makes the world

a better place. But maybe we have to save that for another conversation. Yes, I know. We definitely should. Cindy, this has been such a wonderful conversation. We've learned a ton about really the history of analytics and then also the way that the most forward-thinking companies are doing this both from a technology standpoint and I think more importantly, a cultural standpoint. So it's been really helpful. Thank you so much for your time again. A pleasure talking to you, Eric and Costas. Okay, here's my takeaway, Costas.

I thought about this a lot, actually, throughout the episode, maybe to the point of getting distracted, but you asked some good questions, so that didn't matter. So thank you for that. But it was really clear that some of her earliest experiences as an analyst working for Dow Chemical were formative in large part because they had an explosion in a plant, which could have harmed people. I mean, no one got hurt. But the way that she described remembering that experience and then that being a really driving, a big driver behind her passion for even the product ThoughtSpot and saying, if we had this, we could have avoided it.

I just think that's a really special dynamic. When you believe in a product because you know that it could have solved a catastrophe that you were a part of in some form or fashion earlier. I just thought that was really, really neat to tie that formative experience to what she's doing today. And I just always appreciate people who have that level of authentic belief in what they're doing and the product they're working for.

Yeah, yeah. A hundred percent. I agree with you. From my side, I would say, like I found extremely interesting the whole discussion that we had around the transformation of the roles inside the organization and how this is also associated with cultural transformations

that need to happen in order for companies and organizations out there to embrace and utilize the data that they're generating, right? I mean, we should spend more time with him discussing about that stuff. I think he has a lot of insights and insights that are very interesting for everyone who is a professional in this space. So, yeah, I loved the whole conversation that we had, but this part of the conversation, I found it extremely insightful. Totally agree. All right. Well, thank you for joining us again on the show. Subscribe to get some great episodes and your favorite podcast listening app, and we will catch you on the next one. We hope you enjoyed this episode of the Data Stack Show. Be sure to subscribe on your favorite

podcast app to get notified about new episodes every week. We'd also love your feedback. You can email me, ericdodds, at eric at datastackshow.com. That's E-R-I-C at datastackshow.com. The show is brought to you by Rutterstack, the CDP for developers. Learn how to build a CDP on your data warehouse at rutterstack.com.