Building Strong
Relationships With Business Stakeholders

Join Egor and Nik in this insightful webinar as they delve into the importance of fostering strong relationships with business stakeholders in the realm of data leadership.

Replay Available On Demand
February 21, 2024
4:00PM EST
Join Egor and Nik in this insightful webinar as they delve into the importance of fostering strong relationships with business stakeholders in the realm of data leadership.

This engaging discussion highlights key strategies for success, including the use of shared language, understanding business needs, and translating technical capabilities into tangible business outcomes.

Strategies Covered:

- Importance of Shared Language
- Understanding Business Needs
- Translating Tech Capabilities into Business Outcomes
- Delivering Quick Value
- Guiding Data Strategy with Business Needs
- Shared OKRs and Time to Value
- Empathy Towards Business Stakeholders
- Role of Metrics in Business Success

Throughout the conversation, Egor and Nik draw from their own experiences and share real-world business scenarios to illustrate these concepts. Whether you're a seasoned data leader or just starting in the field, this webinar offers valuable insights to help you navigate the complex landscape of data leadership and build stronger relationships with your business counterparts.


Read the Transcript: 

Egor: [00:00:00] Good to see you. And we are live. Hello, everyone who is joining us. Thank you for tuning in to another exciting live stream that we're going to be doing today. I have Nick Acheson, who is the field CDO over at Dremio. And we'll be talking about building strong relationships with business stakeholders especially as that pertains to data leaders and how they can.

Egor: Build that sort of empathy and understanding within the organization. To kick us off, Nick, do you want to introduce yourself a little bit and especially tell us about the role field CDO sounds you're playing relationship builder on both sides of the of both internally and externally and helping your customers build those relationships.

Egor: So can you tell us a little bit more about your

Nik: role? my name is Nick. I've mostly lived on the other side of the table as a customer or a prospect

Nik: I've run data analytics platforms globally for major companies like Nike, Zendesk, American Eagle. Philips, Concur, and I started my career working at the National [00:01:00] Security Agency, so things like security are always at my heart. I switched out about almost two years ago now, somewhere around there.

Nik: And I became the Chief Data Officer over at Okera. Who had recently sold to Databricks. And prior to going back on the other side, I decided to give it one more shot. Did not really want to go to Databricks. And I reached out to Dremio and we basically created this role over here for me. So the field CBO really at the end of the day I, I travel almost every single week to major customers and major prospects of Dremio.

Nik: And really, I spent about 30 percent of that time talking about Dremio. And what I'm really doing is helping them think about like their data market, their data analytics, architecture and strategy, but really tying that to business value. I always like to say nobody funds architecture. So transformation really happens at the people level an organizational impact and, and.

Nik: data culture. So really helping companies make those ties where then technology kind of invisible underneath the covers allows folks to scale. taking all those different levers and helping our customers build out their own [00:02:00] strategy.

Egor: that makes a lot of sense. And when I think about the market and how long the term data driven organizations.

Egor: And now I think recently we call it data informed organizations have been a concept. It makes sense that that's an required role. How do I fit? Dremio? How do I fit any sort of tool into my overall strategy and how's this going to help my business?

Egor: that's one of my favorite parts of my job as well. I get to talk to our customers, tell them sort of what's your data strategy, what's your journey, what's the next year look like, and how does big I fit into that? How can we help you get from A to B? And a lot of those conversations rely around bringing the business closer into the data, making the data useful for the business, making sure that we're the right people are getting the right information at the right time.

Egor: when I talk about business stakeholders, I often think about the C suite, the VPs, who are consuming the data to make some sort of informed decision. And. They are the ones who need to be convinced that this is [00:03:00] a valid investment. You need a data platform. a data strategy to make you successful as a business decision maker.

Egor: Is that similar to what you've seen?

Nik: Yeah, it's even funny. I reached out to Dremio before I was about to take another position with a competitor and I told him, I was like, honestly, I think your tech's better, but I think you're missing it. I was like, you're, you're missing the real opportunity out there because the reason why I reached out is you know, I was one of the early adopters before kind of even data mesh was a thing and building out a lot of those capabilities.

Nik: And 1 thing I just always anchored I call like the unified access tier, you know, single location to access their data, regardless of where it's at and just again, making it much simpler to shop and consume data as a customer and driving that transformation and pulling the hooks out of individual systems.

Nik: And when I reached out to Dremio, Dremio's main message heavily was on the acceleration side, right? making SQL run faster. now in our messaging, we're a unified lake house platform for self service analytics. taking those lessons learned in the field, bringing them back in and even [00:04:00] help us evolve our strategy of, yeah, it's, you know, you can, you can serve 20, 25 power users in a company where you can serve 2, 500 and make those 2, 500 power users.

Nik: Right? So. how do you really bring the business in instead of constantly looking at a space where nobody has enough engineers, nobody has enough analysts.

Egor: and that's something that going from that tech focused sale to that business focused sale, is, Particularly valuable, but it is targeting a different kind of market.

Egor: Dremio, BigEye, we're technology companies, but we have to build that abstraction layer on top of it to say, this is how we're providing the value and you don't worry what's underneath. What's your take on exposing the what's underneath? Because That's, I feel like that's often a easy to explain to the engineering team, much harder to explain to the business as to why one solution versus another, because the names are all the same in the market.

Nik: Yeah, I got a fun, I got a couple of fun stories. I guess I'll spread them out throughout the call. Instead of burning them all now. But maybe I'll start with [00:05:00] one is, you know, when I, when I joined, I'll just say one company, I normally do my tour around and go, Hey, do you know what we provide on data analytics capabilities,

Nik: what are you missing? And. I remember very clearly one of the leaders at the company within the APAC region and they were doing massive, you know, they ran the sales organization said, the number one skill I have to hire for is not folks that understand the customers, not folks that understand our technology, nothing.

Nik: Number one skill I have to hire for is SQL. I wish it was just easier to work with the data and to allow the insights to make better decisions. Find me a way not even to work with SQL, let alone Python and Java and everything else in the world.

Nik: What the business really wants is to be able to get the insights they need to make better decisions. So if you can serve that in a way that's non technical, I think more often than not, you'll see massive adoption.

Egor: And I feel like the common model, is you have lines of business and verticals

Egor: A set of analysts attached to them. And the analysts are that human [00:06:00] bridge between the technology and the business. And they're the ones who are talking to the stakeholders to understand what they want. And then they're talking to the data engineers to understand, well, how is this even represented in underneath?

Egor: When I was at Uber, We were the central data warehouse team, but a lot of my work was with these lines of business teams. I talked to Uber Eats and Uber Eats had embedded analysts and they had their own set of data engineers and they said, well, here's our data model and this is how we think about the world.

Egor: And our analysts are talking their PMs to understand how the the world works. But then have to translate back to the data engineering team and say, well, this is the tables. We want to represent them. These are the dashboards we're gonna build.

Egor: panacea of let's cut all the human labor out of it and connect business directly to the answers that they want. How far away do you think we are from that? Realistically?

Nik: Yeah, I think a lot of the time we're much closer than we give ourselves credit for. I think we focus on some, maybe the wrong questions at times.

Nik: And so [00:07:00] like when I built a couple of data product teams in the pastone of my favorite examples is American Eagle we had over a hundred dashboards in one particular business unit. And when you got really down into the greens and so forth, you ride along with, with your business unit owners and the digital teams, at the end of the day, like, there's only like five to 10 questions they keep asking, right?

Nik: And it's just like, how do we sell more jeans today? Did we sell enough jeans today? Did people buy more jeans when they bought other things? And when you strip it down there, we got down to about 10 key dashboards for that business area. And one of them was only six data elements in one dashboard.

Nik: it's not as complex when you go backwards and say, what problem am I trying to solve? And I've got some part where it's like, what data model do you need to serve this? No, how can I help you answer that question faster?

Egor: I think that's where that, I feel like that's where that communication breakdown happens a lot though, is because As a data engineer, from most of my career, I think in terms of databases tables pipelines, and getting the data to you, [00:08:00] and this is obviously useful, but then the part that has always been missing is that All right.

Egor: The data is there, but nobody knows how to use it. Nobody knows how to access it. And that building that relationship, building that communication of here are the words that the business is going to use to describe what they want. The business is going to say, I just want to know how many jeans I sold.

Egor: Right. But then the data engineer is going to hear. where's the gene data coming from? Are we have to pull it from our sales? They

Nik: They just want to know that the data they're reading is trusted. Whether that hits four or five systems, whether it's updated near real time, they're not making decisions real time in most cases either, right?

Nik: They're just like, how do I improve? Selling more jeans. How do I make sure that we don't put crop tops on the website again? Nobody was even searching for it, right? being able to bring the data to them and the insights to help them answer questions They weren't even considering that's a true partnership

Egor: Yeah.

Egor: AndEduardo asked do data teams ever express to stakeholders suggestions in terms of prioritizing [00:09:00] business objectives and how's that received?

Egor: You have to already be at a place where the business trusts you to make a suggestion. What I've noticed a lot, is engineering teams will focus on the value that the technology can provide without fully understanding why that value is important. And then that becomes a difficult conversation with the business because the business says, well, what am I getting out of this?

Egor: My needs are actually not, you're not even talking the same language as I am. once you're talking about not pipelines and POS systems and tables once you get into the language of all right, you have stores and stores sell jeans and there are customers that buy those jeans, then the business is going to be much more receptive to nudges and recommendations on meeting needs and business objectives.

Egor: Because you're talking their language.

Nik: Yeah, I can't attribute this next comment for, for a good reason, buta CMO once told meI was like, well, you know, how are you thinking about the problem? And about [00:10:00] 45 minutes into the conversation, he just paused and was like, look, the, how we're going to do this is going to matter.

Nik: I can't fire everybody on Monday and they're going to keep asking you to pull wires and set up landlines. I need you to show up with a cell phone and sometimes you're going to have to teach him how to use it. And in other times we're going to have to keep phones in certain buildings and cell phones in other people's hands.

Nik: And he's like, it's going to be hard for some, it's going to be easier for others. So how you prioritize where you do those relationships matter. And at times being humble enough to show up the right way because in another company as part of our financial forecast modelwe walked in and said this doesn't look right and we built a whole another model with additional data sets in about two weeks and the model clearly showed we're never going to hit the numbers that we project to in the next 3 to 5 years

Nik: And it was right. But that was delivered the wrong way. Like that's gotta be done in partnership. So even in times where the trust isn't there sometimes you have to still build that trust, even though the insight's the right thing to [00:11:00] do. that's where real transformation happens versus innovation.

Egor: And I, I think that's a hundred percent of trust problem. You can't, if you're never going to have a good conversation, if the first thing you do is you go to some business leader and say, you are wrong about everything you've been doing. Let me show you

Nik: why.

Egor: Oh. I feel like this is why engineering gets a bad rap because it's all fact based rather than value based and trust based

Nik: Oh, yeah. I had a, CEO once tell me. He's like, I don't care what the data says it looks right, but I don't like the way it makes me feel. So I'm not going to listen to it. this is why storytelling and partnerships matter so much.

Egor: And this is something that I've always found interesting is the data analysts have historically been the ones who are doing that proactive trust building and communication with the business. As you said, Nick, I think we're edging closer to where the business can go directly to the technology and start asking questions.

Egor: I've seen some cool demos out [00:12:00] there of LLMs within BI tools that leverage semantic layer technology and other annotations and tagging, but allow you to go in and say, build me a report of My gene sales by store for the last year, month by month, and it will break that down and build the right SQL, generate the right visualization.

Egor: Now I'm sure it can't do anything. That's a little bit more detailed and nuanced than that, because that requires the relationship. And so my question then become who is in charge of that nuanced relationship, Upleveling the data analytics teams and turning them truly into business insight teams where they say the technology is now no longer the limiting factor.

Egor: You don't have to hire for SQL. You have to hire for business acumen and translating that down to what is important. Yeah,

Nik: that's, it's actually a very unintentionally amazing question because even in your example, you, you noted store. When I was at American Eagle, we had [00:13:00] three different definitions of store.

Nik: When I was at Zendesk, it was hard to actually say what was sales yesterday because everything had an asterisk on it, right? So I think what we're going to probably see in the next year or two, as, as we really start scaling out LLMs and Gen AI and so forth, is that massive need for governance and quality data.

Nik: And not even in terms of the hallucination problem that we're seeing in some of the data, but just even to your point is how would it know what sources to go look for when I said it? You know, what was sales yesterday in North America, what does our contract look like with ESPN? ESPN as its individual, ESPN with Disney and ABC, like all that foundational data and maturity needs to be there, which is where I think some companies are getting it right,

Nik: And we need to make sure we're mature in that so we can only point at gold sources, for example. But still governing that ecosystem is important, but all that the business still doesn't care about, The business wants to ask in natural language the question. So it's imperative to the data [00:14:00] team to agnosticate the underlying source.

Nik: But we still need to mature it, It doesn't mean that you can go crazy.

Egor: I think that's a good indicator of where data strategies are going in the next couple of years. I've been closely following semantic layer and how you describe your data in a way that makes it easy for you to navigate it, search it, ask questions of LLMs because they are trained on your semantic layer.

Egor: And so Is this really where the turning point where data strategy moves from infrastructure strategy? What database am I using? Am I on Snowflake? Am I on Databricks? DBT? Matillion? are we with the advent of AI and semantically are in a lot of this tooling on top of the data infrastructure.

Egor: Are we getting closer to a world where data strategy becomes more aligned to business strategy and just say, how do we. Eat better, serve the business

Nik: faster. And I think again, kind of working backwards isthere was a good article that came out recently Harvard [00:15:00] business review article, I think Randy Bruin and team sponsored this one.

Nik: And one of the things in there I just absolutely anchored on was a note that like CDOs with a business connection spend on average 5 million less to reach the same revenue goals. As those that don't, right? We're not focusing on trying out new tech.

Nik: You're actually applying it directly to a business priority, getting those shared OKRsfor example, every single thing that we build and we work on, we only count the impact we make to the business. Right. And that needs to feed back into that OKR process.

Nik: So I think it's imperative of it's got to be connected.

Egor: a good example of a business connection for a data team? I believe that you can achieve your revenue goals. Faster. If you have a business connection, have you seen that work in practice? And what are some tips that might help other teams?

Nik: Yeah. the last three ish places I've been dating back to, even to Nike as we kind of moving to like these mesh type of architectures and frameworks is building out [00:16:00] those data product teams, right? So for me, I've generally looked at product managers three different ways from a data product perspective.

Nik: One is just a core platform, right? You got to have that trusted underlying data and having a product owner, But then you have that product owner that's in that business side, right? It's the thing about what dashboards are rebuilding, what models are we building and treating like some of those top 10 dashboards as an actual product, right?

Nik: And maturing them and thinking about them intentionally connecting them to that gold level data. Most of those models classically kind of get delivered and get ignored to the next model. Like really thinking about maturing them because things like LTV and, you know, next best message and so forth, all those require constant dynamic taxonomy is addressing them.

Nik: Oh, and then there's that third space, which I'd say is probably ignored a little bit. Which is that innovation space, having those one or two things of like, what can we actually deliver as an innovation arm of the company?

Nik: And we're also a revenue generator. I think that's important.

Egor: [00:17:00] Yeah. Ava is asking, how do you count the impact from the data team? measuring data team impact is notoriously difficult. And I think this is what Nick you're alluding to is having those shared OKRs, making sure that the data team is aligning with the business and the business goal?

Egor: If the goal is sell more genes, then the impact of the data team will be, how do we help you sell more genes? What information do you need to sell more genes? Maybe how do we automate restocking of genes? If you keep selling out of them, that's a project for the data team to automate that process.

Egor: And what I've noticed a lot is that innovation bucket that you talked about, Nick is oftentimes prioritized over other things, because as engineers, as technologists, we like. To think about stepwise step change. And we like to think about what's interesting and innovative. And oftentimes that comes at the expense of, well, there is a two week solution to this, like if I go and write a little [00:18:00] script and I run it on a schedule, this is going to save the business a million dollars or make a million dollars.

Egor: A lot of times that gets overlooked because that's not considered. And I think that's something that we as a data culture, can start changing our perspective. It's not about the technology. It's about having tools in your toolkit and using them to make business impact.

Nik: No, a hundred percent. You've got to realize, even though being an innovation arm in a data organization, it still has to go into the business. So maintaining a very clear relationship, whether that's a product team or an in function business team, success for this isn't living in your organization.

Nik: Success is being able to hand that off, and that has a continuous life cycle, living on past you, so you can continue to work on the next thing. And I think a lot of times there's a pride factor in there too,

Nik: That's driving so much more revenue for our business. Or I built this really cool model that's delivering this next best experience to our customers. I want to [00:19:00] continue to build on that. There's somebody that owns that relationship every single day, right? So partnering with them early how do you impact their day?

Nik: And that's part of the value stream mapping, right? It's like if I can take, you know, an in function, you know, merchandising person that could never do something to now being able to do something in five minutes, that would have been infinity. That's great. But taking something that is measured and says it takes them three weeks to do this thing that I just made it happen in one minute, that materially changes your business.

Nik: So I'd say being able to measure that internally is just as important as what can we do now tomorrow that we couldn't do today.

Egor: I think that's a lot of time we talk about time to value and especially as vendors, Nick, your job is sell Dremio. My job is sell big guy.

Egor: our job is measured on time to value. How quickly can I get up and running? How quickly can I show value with This new tool, it's exactly the same thing when we were at Uber, best time to value for us was [00:20:00] actually, you take a query that is being run by a thousand analysts.

Egor: What five times a week each, and you materialize that into a table and update that table every two hours. everyone's getting changes at the same time. We're cleaning up the data we're monitoring. No one has to rewrite their queries. Every time the underlying data model changes and from a business perspective, saves the business thousands of hours, but the time, the time it took us to get there was.

Egor: A few days, like it took us a few days to just get all the data, build the pipeline, done.

Nik: And most companies still take quarters to do that, So it was cool that you did that in a few days.

Egor: Uber circa 2015, the data team was phenomenal. The pace with which we moved, was.

Egor: Amazing, that is still something that is much shorter time to value rather than us saying, well, we can make these queries run faster. Let's bring in new data warehousing technology. Let's teach the analyst how to build their own pipelines. That's not going to bring that same amount of value.

Egor: It will bring eventually the same amount of value over a much longer [00:21:00] period of time. I think that's another way of measuring your impact is not just how much value you're bringing in, but. How quickly can

Nik: You get that? Well, I think this is the part of like bringing that product model in, right?

Nik: Because if you told me, hey, there's this new technology that can help me go from weeks, months or even quarters, two days and folks can self service. Like that sounds really cool. Right now, Hey, if we need to change our strategy on how we're offering or even making changes to our UI and that UI impacts.

Nik: is going to change how we serve the next best product recommendation to a customer, right? If I say we can make changes, if you learn something, Right. If I go to a meeting and you say, what would happen if we wanted to test this? we can do it tomorrow that gets funded. it's like teaching yourself to speak in business terms.

Nik: We'll get you funded all day long, but it forces you to measure the [00:22:00] impact of this in business terms? do you make those decisions every day or once a week. So how do you make that real for the business?

Nik: Because at the end of the day, they have your funding.

Egor: I think there's, we've talked a lot about how data teams can work with the business stakeholders. the business thinks that they want specific things from the data team without actually.

Egor: Understanding the implications or the complexities of it, or maybe even not even, not even sometimes understanding why it matters. I have a great example here was streaming when streaming was just taking off and real time data was the next best thing. We, there are a lot of efforts to build real time analytics and dashboards that are updating every five minutes and every 15 minutes.

Egor: And the biggest challenge I've had with that, is that real time analytics is very rarely necessary. there are very few real business use cases aside from vanity metrics where real time analytics is useful because the decision making cycle of a business is much [00:23:00] longer than 15 minutes.

Egor: It's longer than the refresh cycle on your data. Why does this have to be at a certain rate? Why does it have to be a specific data set? It doesn't have to update on a certain time schedule.

Egor: And how have you seen data teams successfully do that, Nick? How do you push back on the business and say, Look, we understand that you want this, but actually you want this other thing. The outcome is going to look slightly different than you envisioned but you'll get the same business outcome.

Nik: As a data team, we often think in functional and unfunctional requirements and break into a detailed architectural spec. my favorite example isWhen I was over at Nike, right. And we were trying to build out like a connected product strategy as part of like a supply chain of the future

Nik: Right. So where we can track inventory all the way from. Initial idea to the warehouse building it all the way through delivery and even doing cool stuff Like I want to scan my shoe and possibly win and meet meeting LeBron James, you know after even a purchase, right? Yeah, you think about that whole flow just in that middle part and connected product, I got a [00:24:00] 50 plus page requirement deck from the architecture team with the in function engineering team of the business

Nik: 97% uptime. I need real time inventory eventing as part of every hop. And we need a deployment in EMEA as much as we need in us and we need to run hot, hot. I'm like, nothing you said has anything to do with the business.

Nik: if I missed a hop, for example, from it, getting off the plane, going into the truck, and then a truck going into a factory, as long as I hit the factory, does that make a difference? Well, no, we can't miss two hops. Okay. So you need to maybe once delivery.

Nik: Great. And that also kills the real time insight requirement. I just pulled you in nine months. Is there a performance requirement that you actually going to make a decision based on, when the data hits? Oh no, but we need to make sure the data is there. I can get rid of the EMEA requirement right now.

Nik: I just pulled in another three months and now we're probably saving up somewhere around six to 9 million. We can probably deliver this thing. [00:25:00] In six to nine months in an MVP cycle and actually test this thing out in market we can start going after fraud as one key material, business level of impact, which is now, obviously, as you know, many millions of hours to Nike in the black market.

Nik: As we continue to firm this up, it's always that fun adage of how fast, how much is it going to cost and how quick to prove value anchoring on that, especially the value piece.

Nik: The business doesn't care about any of that. So making that shift, it's hard for data people, right? Because you always think in functional, not like deliver the need.

Egor: And I think this is a great example as. A product team.

Egor: when you go to a customer telling you requirements, technical and functional the easiest way to solve that problem is to say, what are you trying to accomplish? And then once they start describing that, then you can reverse. Reverse engineer the, okay, well, this is what you want rather than what you think you

Nik: Yeah, no, the metadata catalog we built over there. When I joined Nike, it was so far [00:26:00] behind schedule and the requirements were crazy They're like, well, we have glue and it's pretty shaky and it goes down like twice a month and we don't really know why. And I'm like, well, if I can deliver you the same performance of that, I was like, is that okay?

Nik: You're like, sure. I was like, you just threw out two thirds of my requirements, right? So again, it's just like being able to get on that level makes, makes your life so much easier, but also makes the business life easier because it gets them into a flow of not asking for technical things.

Nik: Come and tell me that you need to see, inventory. a near real time view to make these types of decisions. I got five platforms I can deliver that for.

Egor: And tying into Adriana's question What are some of the biggest mistakes we see data leaders make when trying to collaborate with stakeholders? coming in with the technical rather than with the business as a communication gap.

Egor: if you're not speaking the same language, if you're misunderstanding [00:27:00] the terms or the ideas or the concepts, then you're just not going to have a productive conversation. And I think a lot of time the onus is on the technical person to understand what the business business person, the non technical stakeholder is trying to express, and then mapping that back to the technical rather than trying to get the tech, the business stakeholder to understand the technical and map it to the business.

Egor: And so that's, I, I always consider that a communication gap using the same words and concepts and making sure you both agree on what those words and concepts mean.

Nik: You always want to when you say that.

Egor: What is a word?

Nik: Without that word. I think understand. It's that space between understanding the mission and empathy, right? And I've had really good conversations, even with tech teams are like, well, why is that customer still on Oracle? Right. How do, or I go in there, how do we get the supply chain team out of SAP? do you think they want to be in [00:28:00] there?

Nik: What they care about is delivering a customer experience. What they care about is having confidence and knowing what's in between the four walls, What they care about is going for a SaaS company that I'm not going to create unintended friction that creates a churn signal. Like those are the things they care about.

Nik: So when you come in and say, I fell on my face early on my career, and I learned this lesson on the mission side. If you come in without having understanding of the mission and why they're doing what they're doing, you're always starting out on the wrong foot and you're going to have a short relationship and lifetime in there.

Egor: Yeah, I think that makes a lot of sense. let's say you agree on what you're building, why you're building it, why it matters. I think the communication still continues throughout the whole project lifecycle.

Egor: going back to your example of the glue catalog. can I just make glue more performant? you still need to keep that communication up as like, are we still experiencing the same challenges a month later, two months later, three months later, when is good enough?

Egor: Did we [00:29:00] misunderstand each other in a previous conversation? Now we're implementing things the wrong way. Once I get my requirements, I'm just going to go off and build. But. It's important to not just go off and build and but check back in periodically and say, are we still building the right thing?

Egor: Is this still a priority for you? Because business priorities, shift a lot faster than technical priorities do and keeping that aligned and making sure everyone's on the same page over time is

Nik: Well, I've always prioritized that platform team of product and kind of separate it from the analytic and data science side of it,

Nik: If you're building the right foundation, the business priorities can change, if I want to think about running a distributed architecture and how do I do that without being stuck in proprietary systems it's like, well, what does that mean to the business? What I'm actually building is a faster way to move into customer 360 implementations,

Nik: And if we do it this way, I can shift intoour messaging system, right? So I can do a next best message. I can also shift into supply chain and do [00:30:00] next best product recommendations, And I can also shift into product itself and go, what's the next best thing we should even be building based on trends

Nik: Like, whatever priority we shift into the business, I've built modularly enough that I can pivot in and continue to deliver fast. And I think that's the part where if you pull them out of the tech a little bit, if you're building the right way. It's easier to be a good partner to them to say, sure, if we want to switch to next best message to next best product, great.

Nik: it's a different dataset. I can pull that in the same way I'm executing this currently. And when I come back, it's foundational. we might be able to experiment with both.

Egor: And I agree 100 percent with you. I will warn our listeners of a mistake that I have made multiple times which is over optimizing for the future, thinking that I know what patterns are going to come up and try to predict the business use cases and predict the product functionality and [00:31:00] It makes the original implementation overscoped.

Egor: This is why I like to talk so much about time to value is at the end of the day, do it a few times in a maybe not super scalable and sustainable way, but you will find that pattern faster than you would if you tried to preempt it ahead of time. So that's just, that is a warning that I have, that's a lesson I've learned the hard way.

Nik: It's why I say, build incrementally to that state, Like everybody moves towards a north star, you're going to take on debt. As much as you might not want to, but that's where it goes back to, how much is it going to cost? How long is it going to take? How fast to deliver?

Nik: Deliver fast, And build incrementally. So you have to take those pivots. the pivots are more on the use cases, not necessarily architecture, but you've built modularly and open enough that you can take those changes.

Nik: The cool thing about working with Dremio is if you want to rip it out in seven years, it shouldn't take you 10 years to rip it up, we're not building proprietary. And, you know, even though some platforms may say open, good luck ever taking your code base out of them.

Nik: I think it's possible to build something that is[00:32:00] easier to migrate.

Nik: Two years, not 10. you're just moving the logic at that point. You're not changing everything.

Nik: And, that's hard.

Egor: I feel like, as a slight tangent, we as a SQL, as much as we say, Oh, SQL's a standard, and it's all ANSI, and let's be real, there's no such thing as standard SQL. And every database migration project takes years because 80 percent of this translates directly. And we can copy paste these, this code and the rest of it is custom functions and proprietary, and we're rebuilding everything in the new environment.

Egor: And so I think this is where open standard meets the harsh realities of proprietary software still rules.

Nik: That's just like a customer telling you they have a standardized connector. I haven't seen two companies do the same Kubernetes deployments,


Egor: think that's something that is very true. And I want to go back to a little bit on attention using The same terms and concepts You mentioned a great example. What do stores [00:33:00] mean? What do genes mean? Well, I guess hopefully genes mean the same thing everywhere, but who knows?

Egor: I think that's something that a lot of teams struggle with today is. Standardizing on definitions and metrics. And when it comes to data, a lot of what we talk about are those business metrics. What is, back at Uber, what is a trip? Well, Uber also had, I think, a handful of definitions of trips. There's a driver trip, rider shared trip, eats trip.


Nik: I'm sure they were the same on every system that was consuming trips.

Egor: But that is something that I feel becomes a headwind for a lot of these efforts is when you are talking to the business, it's matters which line of business you're talking to, and it matters who you're talking to.

Egor: For example, if I was talking to the Uber Eats team. And we were talking about trips. I knew exactly what they meant by trip. And that was not the same thing that is meant by ridesharing. And that is something that [00:34:00] I realized pretty quickly, where the word might mean the same thing, but the definition and the implication of it is very different.

Egor: And that's something that I found very important, and that's why I think Metric Catalog is Has actually died off as a concept, but it's still massively important for large organizations.

Nik: Yeah, I know I, I, I speak probably too much length sometimes on data catalogs in general. Cause I get a little fussy when people are like, Oh, what's your data catalog?

Nik: I'm like, for what? Right. Cause you have your metadata catalog, you have your business terms and glossary, and you also have generally that space where people shop for data. As a data catalog, and they can be three different things and even how you're doing pub subbing and a bidirectional relationship between your catalogs, right?

Nik: My favorite example is like if somebody is doing something in Dremio, right? And interacting with data that's in S3, but that data is also being consumed by Snowflake as part of a marketplace, right? But then someone's also [00:35:00] interacting directly in S3 at the bucket level. Well, when changes or enrichments happen, what does that mean?

Nik: how do you manage the flow of that? what's the source of truth? So like, this is where I get in fussy fights sometimes with our customers. I usually make it easy when I go, cool. If you want to standardize on Unity catalog and Databricks, is every single data element and every interaction going to PubSub back into Unity.

Nik: no, absolutely not. then neither one of us should be your enterprise catalog, right? we should talk about facts and how we define facts and the path to being a fact in the business. And who owns that fact?

Egor: Yeah. Ownership, I think, is massively understatedbecause It's not just about aligning on terms.

Egor: It's about on who is allowed to even define that term. Who is the ultimate person who says this is what that term is and every other definition is wrong. And catalogs, I feel like give you a false sense of security because you see the words on a page, but then who actually put the words on the page?

Egor: It becomes a different conversation.

Nik: And [00:36:00] I would say there should always be a business owner and a technical owner. start in finance. There's someone owns what revenue is.

Nik: So how do I capture that? How do I make sure when changes happen in Salesforce all the way upstream? there's somebody that actually owns the downstream impact of that and starting there is much easier than starting at things like Who owns trips when six parts of the organization do.

Egor: Adrian asked another great question. What advice can we give for working with stakeholders who are resistant to change or are difficult to collaborate with? it boils down back to empathy. why is that stakeholder resistant to change or difficult to collaborate with?

Egor: Maybe they just haven't had a good experience with the data team. Maybe they've had requests that have just gone ignored for a long time. they're saying, well, data team never gives me anything. So why should I work with them? starting from zero, building that trust up, building that empathy, understanding where are they coming from?

Egor: What do they need? think small incremental progress. What make small differences in their life and they will learn to love the data team, then be more open [00:37:00] about their needs. That's what it boils down to.

Nik: Yeah. What's really fun Adrian is I have one slide that I've used for the past eight years across about five different companies.

Nik: It is the same slide. I just changed the company name in it. It talks about the technology transformation, People want to move to data driven. They want to move to event based, decisions and, algorithmic, intelligence but then there's the people.

Nik: Right. Being a primary landing house for analytics everybody still wants that same stuff. The technology part is so much easier, right? Who wants to run cubes anymore in a contemporary architecture environment? People know how much of a headache it is updating cubes. People know how brittle cubes are, right?

Nik: But that still runs so much part of the world in so many parts of the businesses. And it's like, how do I get them to change that? my example I'm giving is actually a major company in the U S that I actually had this exact conversation with. They're worried about driving more adoption because the teams are resistant to change because they're, comfortable with cubes it's the [00:38:00] transformation part of this and the culture part that matters so much.

Nik: Not the technology stuff. You can sit down with them, run through the new architecture. But guess what? I'm running cubes tomorrow to answer all my business questions, right? So bringing them along is massively important every organization I've been in and then I advised to, it's a slightly different answer.

Nik: You know, the data culture in that organization matters, the maturity of analytics in those organizations matter, how mature they are thinking about the business matters. But I will tell you the one thing everybody can always anchor on his mission. So if you can connect it into how this would impact their day to day, that matters.

Nik: Like some people are motivated by that. Other folks are motivated by buying new stuff. Other people motivated by making more money and having more opportunities

Nik: So it's really connecting with and understanding how the, how people tick and then helping assure how you can drive that transformation and not just that person, but then that team and then that organization,

Egor: I got you. I agree with that. people get used to their workflows and they don't want to think about new workflows and think about how to change just because it's hard to break from a habit.

Egor: If you've been doing the same thing for years, you have to be really convinced that the new way is going to be better for you. that is going back to us teaching that person, teaching that team and building that understanding of this is why this will be better for you.

Nik: Yeah. And back to the empathy, it may not actually be resistance.

Nik: they know this runs the business today. And I think there's a level of maturity on our side, advising them. why would they make that change knowing that any of those changes could impact their workflows, their [00:40:00] customer so there's a risk profile that has to be addressed there.

Nik: So I think how we show up absolutely matters.

Nik: Maybe I'll give you one more fun story while we're, while we're sitting back. So I speak a lot about data mesh. It's obviously even core to Germany's value proposition in the market. Back to meeting the business where they're at. One of my fun stories here is I was actually talking to an E suite E staff leader in the company I was working at, and we were pitching them running a centralized versus a decentralized model, talking about data mesh and [00:41:00] so forth.

Nik: The person which I love and I'll take it with me forever. He says, I don't care because I'm like a mushroom I'm kept in the dark all day. I'm fed excrement. And every time I grow, I get cut down again. And then I get bigger numbers to go meet again.

Nik: Can you help me solve that problem? And when you think about how we deliver to the business, they really don't care at the end of the day, what they're caring about is how do they keep growing? So when you talk about things like data mesh and fabric and centralize and virtualization and all that jazz again, connecting it into what does that mean for delivering value for them faster and helping them grow their business?

Nik: That's why these trends are picking up so much speed in the market All right, Egor, I shared another story while you're gone.

Egor: I love it. you are a wealth of stories.

Nik: There's a lot of pain behind those stories though.

Egor: It's either that or my 100 open tabs that I never end up closing. One or the other. I know we got about seven minutes [00:42:00] left. Happy to answer any questions. We've talked a lot about the human side of this.

Egor: About the empathy building, mutual vocabulary, about how to interact with these stakeholders. So now we're doing all this successfully. We're engaging our stakeholders.

Egor: We're aligning our OKRs. working on the same things. How do we communicate that back out? What does success mean? How do we talk about joint success rather than this is just success of the business, doing business things.

Nik: As a platform or data team, I've always focused on the marketing side of it, I never talked about how many dashboards we built. What we talk about is, the impact we've made in the business back to those OKRs,

Nik: Here's what's possible. I'd have product managers product teams and even the engineering teams. I used to go into any engineering team and say, if I can talk to one of the engineers and they can't tell me the impact their work's going to make, I will shut it down because I'm accountable for that extreme ownership.

Nik: Every single thing we're doing is, foundational to the next things we should do, People always talk about Legos, but in my mind, [00:43:00] like Legos are really about continuing to deliver, right? And building and building and building and building upon. So if you tear it down, it's easy to build other places without having to start all over.

Nik: I think that's the cool part. When I think about like the architectural history, when you think, you know, through where data architecture is, is continuing to mature down and mature out. It's. How is that continues to be a foundational pattern for what could be next?

Egor: A little bit and then we get you to business outcomes faster.

Egor: That's awesome. How do we follow up on this? Any resources, communities that we can recommend for conversations like this? My, my personal favorite is the Joe Reese podcast. He, I think he has two of them. He has his own podcast and then he has one with Matt. But he oftentimes we, we, I have exactly the same conversation with Joe and he has them with other data leaders. So I would definitely tune into his stuff for more conversations like this.

Nik: Yeah, I'm in the same boat as you. I go to different places, [00:44:00] depending on what I'm looking for. Like if I want to learn about openI look at Alex Merced, who's an influencer at Dremio, but talks about the whole space.

Nik: If I'm looking at, general data quality there's a couple of people I follow their data contracts, So I more so go like there's particular people that are just amazing at the spaces that they're in. And their content is always so good, because you get a lot of these kind of stories when I go in and talk to a customer, it's not like, look what we could do, it's, hey, look what Coca Cola is doing.

Egor: We've come full circle. Well, Nick, I appreciate you taking the time today.

Egor: Thanks a lot for the conversation. And thank you everyone for tuning in. we will send out a link with the recording afterwards and a link to a blog post. That outlines a lot of these topics as well. if you're interested in checking out big guy, feel free to message me on LinkedIn or visit us at Bigeye.com

hosted by
Egor Gryaznov
share this episode