Orchestrate all the Things - GoodData and Visa: A common data-driven future? Featuring GoodData CEO and Founder Roman Stanek

Episode Date: June 1, 2020

From user, to partner and investor. That's not a very common scenario for software vendors, especially if the user-cum-partner-investor is someone like Visa. GoodData is evolving more than its relationship with select users. In this backstage chat, GoodData CEO and Founder Roman Stanek and George Anadiotis discuss the ins and outs of the deal, the data landscape, the way data are used to shape directions for organizations big and small, and what's next for GoodData. Article published on ZDNet, May 2020

Welcome to the Orchestrate All the Things podcast. I'm George Amatiotis and we'll be connecting the dots together. This is episode 3 of the podcast featuring GoodData founder and CEO Roman Stanek. GoodData, one of the key players in business intelligence and analytics, announced a partnership with Visa in May 2020. This is an interesting development for both sides for a number of reasons. Roman and myself connected to discuss the ins and outs of the deal, the data landscape, the way data are used to shape directions for organizations big and small, the quest for a single version of the

truth, and what's next for good data. I hope you will enjoy the podcast. If you like my work, you can follow Link Data Registration on Twitter, LinkedIn and Facebook. So yes, thanks again Roman for taking the time to connect and as I was saying we may as well start with where we left off the previous time. It wasn't actually you with whom I had the conversation. It was with GD, huh? Yes, exactly. Yeah, very cool. And it was a very productive one, actually. And it's not been that far back in the past since we had it,

but I thought it's a good place to start. And actually, for an additional reason, because it was about a partnership, and this is about a partnership as well. A partnership, the last one, with Amazon. And this one may even be bigger, but. The partnership, the last one with Amazon, and this one may even be bigger, but let's start with the last one. So how is it coming along, basically?

Are you happy with that? Yes, I am. Yes, I am. You know, AWS is, you know, the biggest player in the cloud, and, you know, we are, you know, historically we were very close to AWS. And we are very much looking at partnership with AWS as the next development for the company for availability of good data around the world. It's all connected with everything. It's all connected. If you look at our bigger vision,

our vision is to help companies to get what I would call return on data. So help them to really kind of use data to cut costs or get profits or build new products. So we are not focused on some silly dashboards and so on. We focus more on what you use. The notion is that we need to protect the privacy and security of the data.

And so we have to be kind of present around the world with different data, privacy, data regulations, and switch because they have so much presence. So we can actually leverage their presence in their territories. And it's also aligned with, you know, plan to launch our new Kubernetes product on AWS this summer. And so it's all kind of coming together. to launch our new Kubernetes product on AWS this summer. And so it's all kind of coming together. It's the vision of being kind of a global player. AWS can support us. The vision for a new stack that's fully Docker and Kubernetes based. And now this endorsement from Visa.

So it's all kind of, again, it's all connected. Okay, okay. So, yeah, the Kubernetes aspect is interesting in many ways. And to be honest with you, what you just said, I think kind of hints that you have broader plans for your Amazon partnership, because initially it was all about Redshift, if I'm not mistaken.

But it was maybe just the beginning. That's just the beginning, exactly. If you look at AWS, to look at AWS only as a Redshift player, that would be too limiting. We really see AWS as a major partner in other expansion around the world. And if you look at it, I'm going to use the term that not many people know how to use it. I call it Balkanized. The web is becoming Balkanized. You no longer

can manage data from one place. You have to have a data center in Europe, a data center in post-Brexit London, and US, and Canada, and so on. So it is more and more challenging for data providers like the data to have all of the presence around the world. And that's where, again, it goes beyond what we did initially with AWS in terms of Redshift partnership. Okay. Since you already touched upon the Kubernetes aspect, what we did initially with AWS in terms of Redshift partnership.

Okay. Since you already touched upon the Kubernetes aspect, you said your plan is to make it generally available, I think, in a month or so. So you mentioned July, or did I get it right? It's not general availability. It's going to be more like a beta or public preview. It is a big shift. It is a big shift for us. Moving to a completely Dockerized version

is essentially a re-architecture of the data. There are so many new concepts, and I'm super excited about it because there are so many concepts, there are so many tools, so many interesting kind of development that we can leverage in the Kubernetes world so that we are making good data

like a very, how would I say, a very good partner in that kind of technology ecosystem. If you look at Kubernetes from the whole cool new ecosystem of tooling and management tools and provisioning tools and deployment tools and so on. So for us to kind of fully leverage that ecosystem, it's a very large engineering effort. But at the same time, when we are done, it will be like magic. Yeah, I agree. I mean, you already have, and it's one of your defining characteristics, I would say, a big integration ecosystem in terms of ingesting data, data command. Yep. And I would say that going the way of Kubernetes would actually broaden your role in the data ecosystem

in the sense of allowing good data to not just ingest data, but actually also export data through Kubernetes to things like other dashboards and application monitoring tools and this entire ecosystem. Yeah, absolutely. And Kubernetes is just such a modern way how to host. If you look at it, good data is a very large application. We have 17,000 CPUs.

So we have a very large application with servers and so on. And so Kubernetes makes it much more manageable. It makes it much more kind of, again, much more modern and more flexible and so on. So we are super excited. And again, that's kind of part of our strategy is to offer customers the most advanced analytics platform. And AWS is a big part of it. But if you build it well on Docker and Kubernetes,

it will work on any hybrid cloud. It will work on IBM cloud. It will work on IBM cloud. It will work on Google cloud. You know, it's going to be really exciting how, you know, your comment about integration is very interesting because you add another angle, and that is integration with public clouds and integration with public cloud infrastructure. Yeah, I think basically it may even enable a new market in a way for you. I was just discussing the other day with the CEO of Grafana Labs, the company, and we were talking about their growth as a company combined with the trends in the market in terms of cloud versus on-premise

and multi-cloud, hybrid cloud, and so on. He shared a piece of information which I found very interesting. So he said, well, okay, our growth is basically 50-50 comparing on-premises versus cloud. We have many more clients that sign up for our cloud offering, but we have much less

clients that choose to go on-premise, but these are the big lines, actually. Yeah, yeah. You know, and it's actually, it's really interesting the on-prem, you know, it's for us, you know, historically we were always hosted The on-prem, you know, it's for us, you know, historically we were always hosted in the cloud

and, you know, kind of managed by good data. But this new Kubernetes development will allow us to, you know, allow our customers to manage their own good data infrastructure. So that's new for us. But also it will allow us to manage good data in small public clouds. It will allow us, again, to launch good data in AWS in India or Middle East and some other places. So it is kind of, I see managed

by good data and managed by customers as two important parts of our strategy. But the beauty of Kubernetes is that we will have the same technology. Yeah, indeed. Like you said, it's a big investment, but it opens up your horizons in a way. Yeah, absolutely. Absolutely. You mentioned that initially you're going to go with a beta release and I was wondering

if you have any idea, if you have any specific strategy, which of your clients you're going to be making it available first? I'm sorry, what do we... I was wondering if you have an idea of which subset of your clients you're going to be making available, the beta version of good data based on Kubernetes first. Yeah, it's going to be the clients that are highly technical. You know, that's the difference between the cloud version and the Kubernetes version is that the cloud version of the data is very, you know, we manage it all end-to-end. And the Kubernetes version is for, at least initially, it's going to be for geeks.

It's going to be for people who want to kind of go and explore Kubernetes and Docker and the whole ecosystem, you know. And so initially, we are looking for a highly technical audience for that launch. But over time, we will kind of get it to the place where we will actually be able to offer it to anyone. But at least initially we are looking for someone who will actually be highly technical, who will work with us on making sure it all works.

Okay. Okay, so I guess with that we may shift our focus to the main topic of the day, which is your upcoming partnership with Visa. I think if I'm not mistaken, Visa was already a client, so in a way it's a kind of upgrade, so going from working together with them to a new status. Would you like to describe a little bit how exactly that came along? Yeah, and I actually think it's, you know how it is. It's at the end of the day, what really defines partnership is,

good partnership is shared vision. And I believe that we have the shared vision that data can make a world a better place. And, you know, so we spend less time talking about the dashboards and the reports and so on. And we spend more time talking about the real business impact. What kind of insights can we give people? You know, what can make, what can help a merchant get access to better information, better marketing information, better credit information, better cash position, or anything like that? It's the shared vision that data is a very important business tool.

And, you know, and that's kind of, that's what drives the relationship with Visa. It's, it's, it's, the technology is important, but having that right kind of business perspective and, and helping companies to actually get value out of data. That's, that's what I believe we have the same shared vision. Okay. Okay. Well, the reason I'm asking is because it's not the most usual thing you see in partnerships,

actually. I mean, it can happen. But this is actually, you now, I would say, have a kind of triple relationship with them. I mean, presumably, they will obviously keep using your services, your software. They have invested in the company, which is, I would say, a bit unusual. And then, of course, you now have a close partnership. And I don't know, this is a totally speculative projection on my side, but I was wondering if, on their part, this may have to do with the fact that there's a new regulation in place recently for the last year or so from the EU, which is called PSD2. And what it says in a nutshell is that basically financial institution APIs are now open to anyone. And third parties can use them to offer banking services to clients.

And obviously, if an organization wants to do that and aggregate data from many different data sources, then having a partner like yourself would come in handy. So, like I said, it's a totally speculative projection on my side, but I wonder if you have an opinion on that. Yeah. No, again, I don't want to speculate either, and I cannot speak on behalf of Visa. But as you can see actually in the press release, they absolutely believe that, and we believe that

the data is valuable. And again, you will have to talk to Visa why and so on, but you're absolutely right. We have a multifaceted partnership. We are good with Visa as business customer, we are an investor, we are building these new data products. So it's a multifaceted partnership and you have to ask them for the justifications and reasons and so on. But from our perspective,

it is kind of obvious that every company will have to become a data company. Every company will have to look at their data and say, how can we actually leverage the data to improve the processes internally? How can we actually get the data insights to our partners, to our business networks and so on. And so that's kind of that's good data vision for a while now is that every company needs to become a data company.

And that's missing in a lot of discussion in analytics is only focused on the technical aspect. It's only focused on different types of visualizations and different types of approaches like natural language processing and so on. So people spend most of the time talking about the technology, but not enough time talking about the business of data, real business impact. And in this environment where everyone's looking for new ways how to compete and save money and so on, the data is even more valuable. Okay. So I understand it's actually the investment branch of Visa that will act in that capacity and invest in good data. Typically, when we see deals like this, it's quite common that the investor gets to place a board member

in the company in which the investment is made. Is that the case here as well? You know, I'm not... Yeah, we definitely have, you know, Visa have certain rights. I don't want to talk about like what they

have and so on. It's, it is, it is kind of, we are a private company,

but if you're not at liberty to discuss the terms, it's fine. Okay, okay,

we can move on to the next topic actually, and we have, we have quite a few. I spent minutes going through your latest news, and there's quite a few things to talk about there. So one of the things that we discussed with Zenex the previous time was the modeling aspect and the semantic layer,

which, again, is, in my opinion, one of the strong points of what you do. And I saw that recently there's been a new development around that. So it seems that now people can work in a collaborative way around that. And to me, it seems like a very reasonable next step. And I wanted to ask you to say a few words about, again, how that came along and what is the reception you're seeing? Yeah, yeah.

So it's, you know, the semantic layer is really kind of the critical feature, especially what we are trying to do here. You know, if you look at, you at, there are two types of analytics. There is a Tableau type of analytics that's very exploratory, that's very much like the analyst is analyzing the data. Good data is used for that as well, but we also use much more for communication of data. Companies are communicating data to their clients,

to their merchants and so on. And the data model, the semantic data model, actually helps communicate the data, and it actually helps to get to a single version of the truth. So if I'm a merchant and I'm looking at data coming from Visa, I don't need to understand how they store data and all of the semantics of data because in good data, it's all kind of hidden and it's all made available

via the semantic model. So it actually works really well for our clients who are communicating data to people who are essentially not data scientists. They are like someone in financial department or someone in travel department and so on. So that's the main value, and that's why we are doubling down on making the modeling much more intuitive,

much more easier for new people who are new to data modeling and new to data. But it is a very critical part of our story. And that is that we always, when we do data, you just don't see the data. You also see the data with a certain perspective or point of view, and that's the logical data model. Do you have any insights you can share in terms of roadmap for the evolution of the semantic layer and the

collaborative aspect? Because to me, this was a major milestone, uh, being able to, to, to work on models in a collaborative way. So I wonder if you have any, uh, idea what's next. Yeah. Yeah. Again, as I said, it's, it's, it's, you know, we have to make it augmented. We have to have, have companies to discover new data.

We have to make it more collaborative. Um, uh, and, and it's, it's kind of a, you know, you know, I know it's a cliche, but it's a very important stepping stone on a path to like a single version of the truth. If we can actually have all the people to see the data the same way, internal and externally, they will be coming to the same conclusions. And so we are investing in it.

We are making it broader. We are making it easier to use and so on. But as I said, it is kind of critical. You're absolutely right. It's very critical for good data success and success of our customers, how to communicate data to people who are not technical, who are not data scientists. Well, what you answered kind of triggers me to do a follow-up question because you mentioned a single version of the truth.

And as it happens, my background is in data integration as well, and large-scale data integration. If you ask people who have this kind of background, they will most likely tell you that achieving a single version of the truth, well, the web is the ultimate counter-example to that. It's practically impossible to do. What you basically end up having is this kind of islands, let's say, of locally shared models around which there is some kind of agreement. In that respect, I would say, intuitively hearing your answer,

I would say that maybe the next step would be enabling these islands of data modeling agreement to talk to each other, to do data model mapping. Exactly, exactly. That's kind of the collaborative part, and do data model mapping. Exactly, exactly. That's kind of the collaborative part, and that's kind of the augmenting part. You know, if someone already did some modeling in other part of the company, can I reuse it? Can I actually connect to it?

And so on. And, you know, I'm absolutely with you. You know, there's no, like, a single version of the truth, like, across the company. You know, it's impossible impossible and we know it. At the same time, in certain domains, if I'm preparing data for my merchants, I want to

have a single version of the truth, at least in that domain, so that when I talk to 20 people, they don't come with 20 answers. Yeah, that makes sense. It's all about scale, basically. Yes, exactly. On a global you know it's it's all about scale basically why example that on a global scale is practically impossible and some people would even argue not even desirable doing it on a local scale is actually necessary oh absolutely yeah you're absolutely right you don't want to stop the

company and argue for a year what is the the global version of the truth across the whole company it's impossible but every every department every you know argue for a year what is the global version of the truth across the whole company. It's impossible. But every department, every domain needs to be able to agree on their own kind of nomenclature and share it. That sounds much more pragmatic, I would say. Another thing that I would classify as pragmatic, which I saw in your latest news, was the Accelerator Toolkit,

which seems like a starter, getting started toolkit for people to enhance data analytics in applications. Would you like to say a few words about that and again, how it has been received so far? Yeah, absolutely. So if you look at the history of data, most of the data today is presented in a form of kind of static dashboards. And these static dashboards have very, you know, very low adoption. They are difficult to understand.

They have very limited flexibility. There's like essentially every company puts four charts on a screen and they call it a dashboard and they call it a success. And that's not what the consumers of data want. They want to have much more dynamic and much more intuitive interface. And imagine that you go to Uber and instead of going to a map that shows you cars, you would have to go through some dashboard that you would actually see where the cars are and you would have to find the car and go to the dashboard and see the price and so on.

So we are not used to deal with data through the dashboard metaphor anymore. We want to have a much more dynamic metaphor. But that requires some JavaScript. That requires some development. That requires some libraries. There is some sort of a learning curve and with our accelerators we are actually kind of making that learning curve easier so that people can build much more dynamic data applications and make data accessible much easier okay so it's not um would i be right if i said that

it sounds like it's a combination of technical components. So, I don't know, probably things like JavaScript wrappers or API abstractions and maybe also a little bit of non-technical approach to that. So, best practices or training material. Yeah, and also some layouts and some stuff that actually makes the building, the data application better. So yeah, we're actually working on, and then maybe the next time we talk, you know,

about Good Data University and how do we actually help people to onboard and so on. But the accelerator really is all about, like, how do we actually help people, customers, you know, build more complex applications easier? Anyone can, again, anyone can put four charts on a dashboard. That's not difficult. But the user interactions are then minimal

and building more complex applications is just too complicated. Okay, I see. It makes sense. And since we're almost out of time, no conversation in this point in time is complete unless there is some COVID reference at least. So you also published an interesting initiative, which basically, if I got it right, collects data from various e-commerce initiatives

or perhaps your clients, you can clarify that, and tries to publish and to get and publish insights based on that data? Yeah, and we work with partner MRCs. It's a marketing application. They have a lot of e-commerce data in their application. And we got together about a month ago and we said that it makes no sense

to send another COVID email that deals with infection rates and so on. People are probably sick of it by now. And we want to create, again, we want to create real value. So we started this initiative and Amaris has led it. And it's hugely popular because it really gives people insights into the data, what's happening in the world. People can make better decisions about their planning

for e-commerce and so on. And that's kind of going back to our Visa relationship. That's very much the vision for the relationship is that Visa that's sitting in some data warehouse is not really helpful until someone has a good access to it with some good perspective. And this MRCS is essentially our perspective on what is the impact of COVID on e-commerce and in different verticals and territories and countries and so on, so that people can make more informed decisions.

Yeah, I think it's an interesting perspective and I totally agree with what you said. I mean, in the beginning, there was this flurry, this proliferation of dashboards about infection rates and so on. But honestly, I think it got quite banal quite fast. So to see something more targeted, more domain-oriented and all bills, I think is a good idea. Yeah. But the key here is the source of data. And I know we are kind of at the end of the hour here, but the main thing is the source

of data. You cannot do it if you don't have the good source of data for that. And you also have to protect their privacy, you have to protect their security, you have to protect everything. So that's the comment you made about the new laws. So, yeah, it's kind of not easy, but at the same time, it's very obvious with a, you know, the power of data so that with MRCs, you know, it is very valuable information for people who make decisions about planning for e-commerce and marketing and so on.

But it is a combination of the data plus the model, plus the UI, plus the vision and so on. So, yeah, it all starts with data, but the result is, you know, an impactful application. Okay. Well, thank you very much for your time. I think it was a very interesting conversation and we covered lots of ground

and again congratulations on upgrading your relationship with Visa and yeah good luck

with all your plans for the future yeah and we will be in touch we will have more announcements about that Kubernetes stuff and so on

so I'm looking forward to talking to you again soon. I hope you enjoyed the podcast. If you like my work, you can follow Link Data Orchestration on Twitter, LinkedIn, and Facebook.