[Radar Recap] Value Creation with the Modern Data Stack

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com.

Hello everyone and welcome to Data Framed. I'm Adele, data evangelist and educator at DataCamp and data framed is a weekly podcast in which we explore how individuals and organizations can succeed with data. So we just wrapped up our Radar 2023 conference and boy oh boy, it was so much fun. So what we wanted to do this week instead of releasing a standard Data Framed episode is share some of our favorite chats and panels from Radar. So throughout this week, we'll be releasing four different Data Framed episodes covering some of the best sessions we found in Radar. Now, just as a warning, the format will be slightly different for these episodes as oftentimes we take audience questions as well, and moreover, the audio quality is that of a zoom call. But we hope this brings additional perspectives from Radar to a wider audience. So starting off this week is a session we had with two previous DataFrame guests, Bar Moses, CEO of Monte Carlo, and Yali Sassoon, Chief Strategy Officer at Snowplow. In this session, we deep dived into the state of the modern data stack, how data teams can succeed in building high ROI, data functions, and much, much more. Throughout the week, we'll be sharing insights from Jordan Murrow, Cindy Housan, Valer Logan, and a lot more. Now let's get started. All right, all right, all right. Hello. Hello everyone. Welcome to one of our first sessions of the day and Data Camp, Radar 2023 on value creation with the modern data stack. My name is Adele and I will be your moderator today and I'm thrilled to be joined by a fantastic set of panelists. Would have deep expertise on the state of the modern data stack today and how to drive value from it. Arguably, one of the biggest challenges data teams are facing today in this 2023 is the ROI challenge. We're seemingly at the cusp or in the middle of a recession. Boardrooms are looking to weather the storm. They want to see higher efficiencies, they want to see higher ROI from Data teams and over the past few years we've seen really the maturation and the evolution of the data tooling space with the rise of the modern data stack. And the modern data stack in a lot of ways has promised to drive that ROI and value for data teams that they are looking for today. However, the main question that we have for today is are tools alone just enough to drive value from data? So here to discuss this more deeply are Bar Moses and Yali Sassoon. Starting off with Bar Bar Moses is the CEO and co founder of Monte Carlo Data, a data reliability company and creator of the data Observability category. The that is backed by Axel, ggv, redpoint, Iconic Growth, Salesforce Ventures, IVP and other top Silicon Valley investors. Previously she was a VP of Customer Operations at customer success company called Gainsight, where she helped scale the company 10x in revenue and among other functions, built the data analytics team. Bar, it's great to have you on.

Thanks for having me.

And now on to Yali. Yali Sassoon is the co founder and Chief Strategy Officer at Snowplow Analytics, a behavioral data platform that empowers data teams to solve complex data challenges that is backed by nea, Atlantic Bridge and mmc. At Snowplow, Yali gets to combine his love of building things with his fascination of the ways in which people use data to reason. Yali, it's great to have you on.

It's great to be on. Thanks for having me.

That's awesome. So I want to make sure that the audience as well will have some brief time for Q and A. And let's get started. So, you know, as someone who's been in the data space for more than five years now, I think in a lot of ways one of the more illusory or confusing terms that I've seen seen is the modern data stack, because in a lot of ways it refers to a category of tools and architecture and even sometimes an approach to building data platforms. So we definitely have a lot to talk about today when it comes to the modern data stack, but I'd love to first start off by really dispelling any confusion around the word and really get your definition for what the modern data stack is. Bar, I'll start off with you.

I thought it was just a buzzword. Is it not? Just kidding. Yeah, it is. You know, I would say it is used sort of widely as a buzzword. Right. But I do think there's sort of something behind it. I think, you know, at a very high level, it's a way to help organizations realize value from data. Right? Realize value faster, more effectively and for more use cases. I think one of the folks that I've heard speak you know the best about this is actually Bob Mulia. He's the former CEO of Snowflake. You can kind of think of him, I think of him as like the father of the modern data stack, if you will. And he had some specific criteria for the modern data stack that I really like. How we thought about it, first of all, sort of level setting, it's cloud based, I think that's kind of like level setting. Second, it's modular and customizable which allows for best in breed solutions. So if in the past folks were looking sort of like an all in one solution, right now folks are choosing the best tool for a specific job. The fourth thing that he mentioned was metadata driven. I think that speaks to the rise of the importance of metadata in the modern data stack. And then finally Bob's take was that it all runs in SQL, at least for now. So I think that's an interesting take for sure. I think at the end of the day, again it has to do with expediting value for data teams and stakeholders. In the past it was very on prem siloed stacks and today it's a focus on easy to get up, easy to see value from and kind of most importantly scalable and reliable.

That's really great. And Yali, from your perspective, what is the modern data stack and your definition?

So I think I mean exactly the same thing by modern data stack as bar. I think I just define it from a slightly different angle. I think it's the ecosystem of tools that have grown up around the cloud data warehouse. So the cloud data warehouses, the snowflakes, the databricks, the bigqueries, they're the core technology at the center of the modern data stack and they've been transformational because they've enabled organizations to accumulate data from lots of different places and compute on it in a way that was really hard previously. Before those cloud data warehouses came along. Was it really hard and or really expensive? And what's been interesting is this ecosystem of tools has grown up around it that has all the properties that Bar talked around. They're modular, people select them because they're best to breed dedicated tools just for getting data into the warehouse. Dedicated tools for processing and transforming the data in the warehouse, taking it out, activating it, and then managing things like observability and metadata across it. So yeah, agree with everything, but I think really worth emphasizing the cloud Data warehouse or lakehouse in the middle of it.

That's really great. So we have kind of a really good working definition of the modern data stack now, both from the characteristics that Bar shared, but also around the ecosystem of tools that exists within this space. Now what I want to talk about is how data teams can create value with the modern data stack. The crux of today's conversation is all around how data teams and data leaders can derive value from these different tools. Value creation with data is top of mind for a lot of data leaders today, especially in 2023, with a tighter focus on ROI. You've worked both of you with a variety of data teams, whether at Snowplow or at Monte Carlo, using your different tools that are incorporating the modern data stack. Walk us through maybe common patterns and traits you found value creating data teams have when it comes to their usage of tooling. I'll also start off with you here, Bar.

Yeah, it's a great question. I think there's, you know, if I think about the world today versus like three to five years ago or even five to ten years ago, I think we were all maybe, if I dare say pretending to be data driven 10 years ago. And I think when folks say today we're data driven, that's probably, I think there's more meat behind that and you know, some of like the data or the sort of arguments behind that one. There's just more data products actually. So in the past, if like the core product of a company oftentimes was decoupled from data and data was really used for internal purposes, like counting how many customers we have, which is, by the way, that's still like the biggest challenge any company has. Tell me one company that is able to count how many customers they have. I don't think that very hard to do. Turns out counting is still the hardest thing to do with the modern data set. But I think you see way more data products, right? Like not uncommon for organizations to have hundreds of data products. I think Shopify mentioned that they grew from like 50 to 250 data products within a year or so. I think the second big chain is like there's obviously way more people working with data. So in the past, if you had like a data analyst today, like pick your title, right? Like data engineer, analytics engineer, ML engineer, engineer, whatever you'd like. Everyone is working with data. Think to Yali's point, there's sort of like emerged kind of these four core pieces of the modern data stack which, you know, folks are kind of working around the first, again, just to recap what Yali said, the data warehouse or the data lake or data lake House. The second is orchestration. The third is a BI or sort of, you know, any advanced analytics platform like Snowplow actually. And the fourth is what I think like data observability or metadata. And so the first sort of primary, you know, I think typically what data teams do, the first thing that they do is like set up the modern data stack. And then the question is like, well, you know, the business asks what have you done for me lately? Right. And so, you know, I think that is something. By the way, it's interesting. I actually wrote a blog post in 2019 about how to think about the ROI of data teams. And it's suddenly like, so it's like back in fashion, it's like back in popularity, right? So it's actually like, I'm pleased to see it. I think it's a healthy dialogue and really great for us to think about that. So, you know, what we mapped out back then is basically kind of a framework helping folks understand how to think about roi. And it's probably, I'm a former consultant also, so I was thinking two by two. So it's basically a two by two. And like one of axis is basically like what is the role that data teams have in delivering value to the business? And I think there's like four or five core areas. The first is increasing the ROI of operational initiatives, which basically means making teams, internal teams, more productive or effective. So for example, working with the marketing team on a campaign analysis to help them understand which campaign is most effective so they can manage their budget. The second is actually driving a totally new initiative, for example, doing analysis of user segments to identify a particular user segment and a hypothesis about it. The third core driver of ROI is uncovering totally new opportunities. So it might be like a totally new product feature or totally new marketing idea, something that the team hasn't experimented with. The fourth is, or fourth or fifth, think of it as the same as one basically figuring out ways to support scale or new capabilities to support that. So maybe it's a recommendation system or an experimentation platform like those things in support of the business. And so the idea is you can take those drivers and map them against any team really that the data organization works with. And the data team oftentimes doesn't work with every single organization internally. It might start with the marketing team and then move on to product, extend into sales and customer success and kind of work across internally. So that's like one way to think about it. And then the second, which we're obviously seeing more and more of, is actually data teams working closely with data in production and with like products that customers are using. So, you know, building data products. So, you know, I think it's very early days of being able to communicate value in this way, but starting to think through like, okay, what are the specific ways in which data teams help or drive value for the business? And then can we create like an impact scorecard in some way or kind of allocate dollars? Many teams are obviously very proud of kind of the value that they bring to the business in terms of a value in terms of a dollar perspective. Can you actually allocate that? I think that probably the most sophisticated teams I've seen do that. Curious if Yali is seeing other things.

Yeah, I think the, if I look across the different data teams that I'm familiar with, I'd say the big difference between the ones that are successful at creating values and the ones that are less successful are the ones that are successful doing a couple of things differently. So they're partnering with the business to identify those use cases bar that you talked about, and they're really working with the business to execute on them. And then they're putting in place the technical, they're using the technology to support the use case. So the use case comes first. So with a clear use case, it's generally reasonably straightforward figuring out the roi. I think you gave the example of a marketing campaign, and that's a very measurable thing. You can show uplift, you can show, you know, through more effective campaign, we've acquired more customers with the same amount of money or we've acquired more customers same amount of money or the same amount of customers with a smaller amount of money. Figuring out the outcome isn't the problem when you're at the use case level. It's actually figuring out what's the most effective way of executing that and making sure that the business gets all the way to value. So the kind of anti pattern there is data teams that stop at democratizing the data or giving the business the ingredients to go and realize the value, but then not make sure that they've gone and taken those additional steps and actually realized the value. And then the real anti pattern is the data teams that put the technology before the use cases. And this sounds crazy every time I, I say it, but I do know organizations that have followed this approach. They decided to go in on the modern data stack and they start assembling it like a Jigsaw. Jigsaw puzzle. And they're like, when I finish the jigsaw puzzle, then I'm going to go to the business and start using it to drive value. And that's all back to front. And it means you've spent a lot, a lot of money before you've shown any value. So the, the big thing is start from the use case, I always say, and work back. I don't know if that resonates, Bar.

Yeah, it does. 100% for sure. I mean, just as an example, maybe to make this a little bit more concrete, like, one of my favorite examples is JetBlue. So folks know, like, JetBlue is sort of a leading airline and their data team is super interesting, super strong user of Snowflake and dvd. I think maybe one of the world's user biggest users of dbt. And I just love that example because it's so tangible, like tracking data about, like, flight information and where's your luggage. Like, we're in a virtual conference, but imagine we were all traveling to this conference. You'd want to make sure that you know what your connecting flight is and you know where your luggage is. Right. And as a data team, when you're responsible for that, like, yeah, you're touching real customer experiences in a very meaningful way and obviously adding value to the business in such a clear way, it's also an awesome data team. So that's a different story.

That's really great. So, you know, in a lot of the answers that you've both laid out here, what's really clear as a pattern for data teams being successful is that they are very closely aligned to the business. Right. Both in operations and how they manage their roadmap and how they prioritize their tooling. Right. I think one, one pitfall we've seen data leaders fall into the past is they're not close enough to the business and they tend to be siloed off as a support function. Right. Yaliu actually came on Data Frame, talked about that pretty extensively, and we actually had Shane Murray, Monte Carlo's field CTO as well, talk about that on the Data Frame podcast. So maybe both of you, I'd love to understand why you think that we still have data teams today that are still siloed off as a support function. And how can data leaders take steps to fix this? I'll start off with you, Yali.

So I think there are a few barriers that data leaders need to overcome to partner effectively with the business. There's a lot of data leaders have come up in the data space and there's a lot of knowledge they've had to acquire around data science and data analytics and data engineering and data data to data to data to data, all these different sort of aspects. And so it's easy if you're working in data to be a bit insular and focus on our profession and our tooling and our approaches and our innovation and so on. When, by the time, I mean, I think that's a failure mode for anybody at any level in the organization, but by the time you become a leader, that's, you know, it's critical. You've got to be outward facing and you've got to be spending more time with the business, working with the business to identify where there's opportunities to apply data than you do with the data team actually realizing those opportunities. But then with the rest of the business, data is a, can often be alien, it can be scary, it can be different. Data people have their own language, their own jargon, their own. So there's often. A reluctance or reticence from the rest of the business. There might be a lack of data literacy in the rest of the business. And then sometimes the data leaders not necessarily set up for success. So I think, I, like, I'm sure both of you and many people in the industry genuinely believe in the transformative power of data. But if the data leader isn't, if she isn't senior in an organization, she's gonna have to fight to have the influence that she really ought to. So I think their challenges, some of them are within our wheelhouse to solve because they're on the data side, some of them are outside the data side and some of them are at the broader, the way the organization is set up and structured.

Yeah, no, I'd agree. I was just going to agree. Yali, we should disagree about something. But. No, I totally agree. I mean, I think, you know, I agree that there's, there's a couple of factors, right? Like one is sort of, and Yali alluded to those, like the DNA of the organization and how important data is in that organization and then the structure, meaning what is what, in other words, what is the role and remit of the data leader? And oftentimes those are not things that are necessarily in our control as data leaders or data teams. I think there are some things that, you know, or thinking about the things that are in our control. I think there's something about like having a sort of, you know, I like this phrase, like what you do matters mentality, what we do matters as data teams, and that's true for any organization. And I think as a result of that, there's a couple of things that we can do. First of all, know what data matters to who and why. So so often, I mean, data teams of like, you know, if data breaks somewhere, they're not totally sure about the impact and like, who cares about that data and actually being like, hey, this data is really important because it ties into our marketing model or financial model or whatever it is. Right? Like, maybe this is tied to ad revenue or product metrics, but like being able to speak to that and know why the data that you're working with matters as an extension of like, what I do matters. I think the second thing that's successful data teams do is tie themselves to important company initiatives so they can like, you know, plant the flag with like, we help with this new product launch and we help with, you know, reducing churn and increasing expansion. And we actually created this amazing experience for our customers to help personalize a recommendation for, you know, the, you know, in fintech companies recommending like the right financial solutions or, you know, if you're in Netflix, recommending like, you know, the best show, that's very important. I watch the recommendation closely. I don't know about you all. I need new recommendations. Recommendations, right. But like, if the data team can have a part in creating a great experience, like that matters to the company. And then finally I would say is like, find ways to measure your team and use that as opportunities to highlight what you do matters. So, for example, you know, we work with lots of folks who, I mean, obviously we're data observability companies, so we work with folks who are thinking about how to measure the reliability of their data. And so maybe you implemented a new sla, for example, on making sure that your Data is fresh 99% of the time, you got, you know, some, you know, you got some measurable, measurable way to sell to see that, celebrate that as a win for the team. Right. If you're able to tie between, here's the metric, here's the win that we have, here's why this is important in tying to a company initiative, and here's who should care about this in their organization. That's a huge leap in terms of what we do matters. I think stepping into that and owning that is something that we have full control on and we can do easier said than done. I recognize that. But I think thinking in that mindset has certainly helped kind of many of the best teams that I've seen that's awesome.

And I want to kind of be able to take a framework from you both and see how we're able to kind of templatize that thinking and share that with the rest of the data community. That's especially watching here. Both of you mentioned on the importance of one measuring impact of the data team, right. Whether that's through measurable ways that are KPIs that are really important that the data leaders should track up. The other one is making sure that the data team is working on really important company initiatives. And this is really closely tied to what you said Yali, on the choosing the correct use cases and starting from the use case. So maybe bar to follow up on one of your answers here, what are great ways that you've seen data teams can measure the ROI of their activities? Right. How do you approach, you know, that particular measurement in terms of being able especially to attach a dollar amount to the initiative of a data team?

Yeah, it's a great. So one I would say this is hard to do. Not, not. I haven't seen very many teams do this. Well, that's kind of actually why we put together this framework, this kind of two by two with you know, the, the idea is like map out the projects that you done or that you'd like to do if you can. You want to think about it as a project base. You can actually think about like the data team's roadmap or product roadmap if you will and then map that against like here are the, here's the value that I've created against like those four or five sort of different areas that I've outlined. Whether it's like making a team more effective or uncovering a new area, mapping that against different functions and then allocating that in some sort of scorecard. So I think that's like a framework that's very helpful. I think there's some things maybe taking a step back from that for a second, one of the things you can also think about this is like leading indicators and lagging indicators and actually Lior Solomon, who leads data at a company called Drata, which is one of the fastest growing security companies and he was previously at Vimeo, which for folks who don't know it's video streaming platform, they have like 200 million users and the data team there was instrumental in helping actually like drive that business through Covid and helping grow users through the team through the, through that time. So the ability for the data team to say like hey, we actually helped Meaningfully during this tough time is a huge win. But he actually speaks about really interesting kind of different metrics. And one of the ways that he's thinking about this sort of like threefold, one is, you know, leading or lagging indicators, sort of an attribution to a particular revenue impacting or impacting in terms of customer experience. The second is actually like leading indicators, which oftentimes look at usage. So he's talking about things like, you know, number of reports that are being actually used, number of teams that use data teams, and then looking at operational metrics, like, for example, time to detection. So, like, how often do you detect data issues, time to resolution of data issues, and then overall, like, issues, you know, are you plagued with an issue every day or every year? And then finally, the third category is team morale and team performance overall. So many folks do, like an NPS, for example, you know, 0 to 10, how likely are you to recommend a data team or a company? And so there's different ways. I think that's like a very sophisticated way to look at it. And it's quite controversial, by the way, to look at usage. I think we don't like to measure ourselves by, like, how many reports we created. We actually wrote. We wrote a funny LinkedIn blog post saying like, or just a LinkedIn post saying, you know, how do you measure the ROI? Only wrong answers. And there were hilarious responses there. I encourage folks to check it out, including things like, how many lines of code did you write? How many reports did you actually create, et cetera. And all these, how many snowflake credits did you use? And all these unreasonable answers. So I think there's more wrong answers than right in these cases.

Yeah. And Yali, from your perspective, what are ways that you've seen data teams also measure the roi?

So I think that the teams that do this best focus on those lagging indicators because those are the ones that really move the rest of the business. So I'm not disagreeing with the leading ones, Bob. I think a lot of data teams fail to articulate the value of what is the difference they're making on revenue? What is the difference they're making in increasing revenue or reducing cost? There's often a lot data teams are doing to help productivity, but I think businesses don't care as much about productivity as they should. I wish they cared more about productivity. And team productivity, I think, is one of the most impactful things you can do in data. But I think in most organizations, you've really got to spot those use cases that have an impact on revenue or an impact on cost, and they appear all over the place. I was really, there's a very strong data team at Globe and Mail, which is a Canadian newspaper group, and they desperately wanted to measure the dollar impact of every item of content that was produced. And if you think about it, that's something that every media company should want to do, but very, very few of them actually get to do it because it's complicated. Item of content might drive somebody to subscribe, might drive a subscriber to retain, and it's going to drive a certain amount of ad revenue against this. There's nuance in how you calculate it, but they use the opportunity that the modern data stack presented to pull together all the data to make that calculation. And they could then show the business, they could show the editorial, the journalism team, hey, look, this is the dollar impact that each of you has had to the business. And the business could start asking and answering a whole new set of questions around. And is our investment in this sort of category of content producing the same return as this sort? So really, I'd urge data teams to punch through to revenue and cost and start helping the business understand the drivers of that revenue and costs. And as soon as that happens, and you can make an argument that through data we were able to drive this much more revenue, this much more cost, the status of the data team and the business is going to totally change and the business will start pulling the data team into the places where you guys will need to be to have the maximum impact. It's more of a tactical answer. Start with that and then all the good stuff will follow.

So I completely agree, and I actually definitely agree that it's very hard to actually measure the ROI of content initiatives as someone who works on content quite a lot. So that's definitely a great use case. Which ties into my next question. We talked about how to measure ROI of different data initiatives, but we also talked about aligning the data team with great use cases. Right. Yali, I'll start off with you. Since you mentioned this, walk me through ways data teams can really scope out high quality, high leverage use cases within their organization.

It's an art, not a science. It starts with an understanding of the business and what are the key initiatives in the business and then an understanding of how data can be used to make those initiatives more effective. So to go back to the Globe and Mail example, if you want to make a media company more profitable, media companies in the business of producing content, and that content somehow creates money and Being able to map one to the other is at the core of taking a data driven approach to making that media company more successful. So it's kind of interesting to follow that example through a bit. The Globe and Mails start being able to report on the dollar contribution of each content item. They start to empower their editorial teams to make more decisions. But then they can start using that same data to optimize the placement of content through the new cycle to maximize engagement. Engagement is one of the leading indicators that then drives lagging indicator of revenue. And that's an example of moving from an analytical use case where you're creating reports that help the business make decisions to actually what Barb was talking about earlier, which is a data product that's optimizing your site to improve the user experience, make it more engaging and drive more value then. And from that they started looking at the subscription flow and optimizing that, personalizing the subscription journey to maximize the chance that a user who's not subscribed goes on to subscribe. So this is a data team that are focusing on what are the core challenges that my business has. Acquiring subscribers, retaining subscribers, then using data to help with that, and then they're using each use case to build additional use cases on, on top of that. So it does, it does start with the, with an understanding of the business and then working back, working back to the data. But it is, it is a bit of an art. It's going to look very, very different in different sorts of organizations. I hope that's not a cop out answer.

It definitely is not. Bar, do you want to add anything here?

It's a tough question. The only thing I have to add is I do like the sort of the idea of treating data as a product and I think that approach lends itself to making a higher impact. And that includes actually having a data product manager and being really clear about who your customers are. So having that mindset sort of forces you to be really deliberate about what the impact is that you will have, who are your customers, what is the feedback from them and what do you need to do to deliver value. So, you know, I think that's like a push in the right direction. Generally. I think that's, you know, I'm kind of a fan of like not reinventing the wheel. And since there's a lot of things from software engineering that have worked so well, I think it, you know, would be, would be great for data organizations to kind of adopt as well. I think it'll just make our lives easier and having data as a product being one of them is one of like the key things that I think are really, really beneficial for, for data teams.

Okay, that is awesome. So we only have 10 minutes left and I want to make sure that we also keep some time for Q and A. Right. So I'm going to cap off our conversation by asking where you see the modern data stack future. Right. So you know, we've been definitely seeing quite a lot of movement in the research space, in the data space, generative AI space over the past two years or so, kind of look through. I want to be able to take your insights here and kind of gaze into the future a bit. I'd love to understand from you. Given you both operate in the modern data stack, both in data observability in the case of Monte Carlo and data creation in the case of Snowplow. Where do you see the spaces you operate in evolving in the next few years and what would be your predictions for the modern data stack, let's say by 2025? I know this is a bit of an annoying question, but Bar, I'll start off with you here.

I wanted to punt it to Yali. Are you sure I have to go first? Look, I'm not in the future predicting business, right. If I was, I probably would be doing something else. So I can't tell you that that's a forte for mine. But I think there's a few things that I'm very bullish on. And by that I mean like I can't imagine the world. I can't imagine a world where like 5 to 10 years now data is less important. It just doesn't make sense, you know what I mean? Like if there's something that I'm willing to bet on is like data is going to be more important than ever. Data teams are going to be more important than ever. Even in this like macro climate. Like you know, data is like the one thing that's continuing to grow and organizations continue to invest in. And so I'm extremely bullish sort of long term on data no matter what. Like the, you know, the, the horizon is, you know, I do think the modern data stack will become the de facto data stack. I think it is, I would say like new and emerging in the sense of like the large majority of people are actually on prem and need to move to the cloud but, or are moving to the cloud. But I do think the modern data stack will become the de facto data stack in 10 years. I don't know, maybe that's a controversial statement. And I do think, you know, maybe finally I'll just say like, I think data in a sense, you know, we kind of like we were off the hook in many ways as organizations. We, you know, could, you know, maybe had less tools than like engineering organizations and we could get away with like, maybe some inaccurate data here and there and maybe the products were like, you know, not the most scalable and we did things in like more manual, ad hoc kind of ways. I think five or ten years from now that's not going to cut it. Like we're not, I don't think we're going to be in a place where we can afford to do that. I think data will sort of get the diligence that it deserves, if you will. And then finally I do think that, or you know, maybe this is wishful thinking, but I do think five to 10 years. I'm not sure that like engineering organizations are getting asked about their ROI. I'm very curious if 5 to 10 years from now people are going to be like, oh, what's the ROI of a data team? I'm pretty sure it's going to be so obvious we're not even going to have to like ask that question.

That's a great prediction. Yali, from your perspective, I'm thinking about.

Bar's last prediction about ROI and engineering teams. I think I see engineer product engineering teams being asked to reflect their roi. So yeah, maybe, maybe we have found a point of disagreement. Just before the call wraps up in terms of how the modern data stack is going to evolve, I'd make a few predictions. So I think the. You mentioned generative AI, Adele, I think you can't. Everybody's talking about it. I think there's gonna, I think we're gonna see AI tooling evolving right now in the modern data stack. It sits, it's in bi. Sit alongside one another. But I think we're gonna see quite a lot of innovation in those tools. And potentially the distinction between them dissolving as tooling to put AI in the hands of more people gets better. So I think that's one prediction I'd make. The other is I think we've got to acknowledge that one of the challenges with the modern data stack is there are lots of pieces of the jigsaw puzzle and that creates integration headache for data teams. So I think all of us vendors snowplow Monte Carlo and everybody else, I think we've got to work hard to make integrating with one another easier so that it is more like putting pieces of a jigsaw puzzle together and less like integrating disparate applications, that is. And I think potentially we're going to see a bunch of consolidation that forces some of that to happen. And then the last prediction I'd make is I think we've all again as vendors we've all focused on being best of breed on doing something very particular, very, very well, which has left our users being the people that have to do that integration piece. So we've, we're giving them something which is best to breed tooling but we're forcing data people to do all this integration work. And if as well as integrating better we can kind of think about common use cases and making them easier to execute across the stack, I think we'd be doing everybody, data teams everywhere, a huge service. And now I've got to stop making predictions. I think somebody's going to walk into my hotel room. Hang on.

No worries. So I think with that what we can do is take quite one question from the audience and this is a data engineering question. So this is quite in line with both of what you do. Bar and Yali. Bar, I'll start off with you. I'm going to show the question here on stage. It's asked by Suresh, lead data engineer at olx. So what according to you would be the top five metrics that would help in measuring the ROI of a data engineering team? Both infra data teams and data engineering teams that generates data products. Yali, since this is about as well data creation, I'll start off with you.

Maybe so it's very tricky to start with the ROI and a data engineering team. I'd look at what are the teams that that data engineering team is serving, what are they doing with the data, what's the value of what they've, what they've created with the data? So you know, what's the additional revenue they've been able to drive, what are the costs they've been able to save? And then if I were a member of the data engineering team, I'd say, look, none of that would have been possible without the foundational work that we did. So we want to say that revenue and that cost, we made that possible, but also by putting in place this infrastructure, we made it possible for other teams in the organization to come and have a similar impact going forward because all data engineering are in a sense a kind of platform team that enable this stuff to happen on an ongoing basis. And I'd be really bullish about putting that business case to the rest of the company.

Anbar from your perspective, top five, let's see. I think in an ideal world I would choose, you know, one, the financial impact of the value like revenue generation. If you can tie to that, that would be my top. You know, I think time to insight would be the second in the sense of like helping, helping customers internally and externally use data or unlock the value of data. The third I would say is around the reliability of data. So making sure that you have minimal data downtime and that you as a part of that time to detection time to resolution of data issues and then number of ad hoc queries.

Duh.

Just kidding. That was a joke. I'm going to cop out and say four. Those four that I mentioned.

Okay, that's definitely all. Okay, we still have one. Time for one more question. Given that we're three minutes away from the end of the session, this one is a bit more career oriented. Right. But essentially this is something that I wanted to ask about as well. You know, with the innovation in the data tooling space that we're going to see in the future, what would be your advice for those just starting out their careers right now, looking to add value with the tools in the modern data stack.

Right.

Especially for data professionals that are more junior or mid level. Start off with you, Bar.

Yeah, I would say, you know the, the data, the data landscape is changing a lot. You know it's, I'd be hard pressed to find, you know, many industries that are changing at this rate. And so I think being curious and you know, staying, you know, continuing to learn, continuing to see sort of what's out there on the one hand and then on the other hand I think there's a big role for focus. I think there's, you know, you can easily do a lot of things in various sort of areas and actually not have an impact in the business. And so actually focusing, you know, we talked about like the core sort of four areas as data warehouse, data lake, orchestration, bi observability, like thinking about what's the value that you can unlock with those. I would say. And then, you know, I think the other thing is. Yeah, I just go back to what I say earlier. Like remember that what you do matters significantly and like be proud of that. And so whatever it is that you want to do, just go do it. And yeah, don't let your fields hold you back.

Any parting words from you?

Spend time with the rest of the business, understand their problems, understand their pains, understand what they're trying to do and you'll just find a whole world of opportunity where you can use data to help them and, and that's where the impact starts.

That is really great, great way to cap it off. So on the numbers, we officially have 1,300 339 people joining us for this session. I really appreciate your time Bar Yali for joining us today. Bar, I know that you have a hard stop and you have another meeting after this, so I'll let you go. And thank you all so, so much for joining the session. We're going to take a 15 minute break right now and then we'll start off with the next set of sessions. But do make sure to message bar yali on LinkedIn in case you have any additional questions or and do if you want to amplify if you like what you see, want to share this, do use the hashtag DCRadar2023. But with that in mind, thank you so much Bar and Yelly for joining us today for this session on Radar.

This was so fun.

Likewise.

Thanks Adele. Thanks Bar.