Data as a Product – Tom Barton, Contino

This is texturing TV. Well today I have the great pleasure being joined by Tom Barton Tom is head of data at Cantina welcome. Tom could be talking with you.

Thank you very much. Great to be here. I'm an old data guy from a long way ago.

So this is a great topic. I love talking about data. So we're definitely want to get to that but please introduce yourself tell folks about you and a little bit about you know, sure one Tom Barton head of the data practice for continuo and what that means is that I'm responsible for data engineering across all three clouds artificial intelligence machine learning data strategy and everything in between.

That's because Contino is, you know, a digital consultancy that really helps you transform the way you do business part of that partial that is data and really I could not do what I do in terms of data modernization and monetization without the compatriots that I have that do Cloud modernization uplift site reliability engineering, you know, Really really modern practices to help you engineer the best experience for your business and Company and of course data goes along with that. So if I don't have a secure space Cloud Landing Zone a really well architected, you know Cloud place for me to put the data then I fail as part of my practice. Of course, my practice has some of those skills as well.

Really we focus on as I mentioned the kind of engineering and outcome based, you know approaches your data. I was having a hit conversation with our our new GM of our research business. We were talking about the how things are so connected between microservices and Cloud native and data and devops and Rapid delivery of software and Ai and how that all fits into it.

You know, no one thing is what it is is about it's about all but how you leverage it together, right? It really it's an interconnected world. I mean we're you know, it's an add some more complex world.

If you think about years ago, we used to talk about the full stack like a full stack developer. I think we're so far beyond that now and there's really no one person that that can know every every Nuance of the stack really really well and that's what we focus on a couple areas and even within data as I focus on it. I find myself, you know talking to customers about, you know, the appropriate place to put their information and and the choices that they have in terms of storage with regard to speed accessibility again monetization.

What's the business outcome you want with? Data, and that directs the solution and we give them and there's just so many, you know, if you just think about the three major Cloud providers alone. They have a plethora of things to pick from so if you're a company, you know faced with making that decision around where your data is and what you do with it for the business outcome.

How do you how do you arrive at a choice? And that's really where I find myself, you know in most of the customer Journey these days helping them navigate that space in addition to talking about just the general strategy of how to use their date. You know, we're in such a Telemetry data censor data, you know financial transaction customer experience data.

Yeah data about what's happening in our applications and observability. We're not lacking for data. That's for sure right here.

Look kind of overrun with data Maybe that's part of the challenges. Where do you start when you're kind of hey, how do you setting on all this information? How do you decide what to tackle first?

It's tough. I mean you you mentioned that the amount of data there's one customer that we have as a billion iot transactions a day. Can you mention that at every day and there are other customers just have data coming in and they don't know what to do with it.

So where where do you start as a often a question we get asked. And what we focus on first is you know business priorities. I mean, that's that's the thing these days is that we're you know talking about a technical solution.

Yeah, it's kind of obvious. But the thing that we hear people pounding on the desk today about is, you know, I need to use this data for something and that is almost always the place to start. So what what's the outcome you're after and where do we start for you?

You know, that is a that's the micro view of the macro problem. You just talked about with like all the data that you have but you you almost can't approach it like all data at the same time. Right you you most modern business today unless you're you know a small to mid-size Market you have and even some of the mid Market companies have multiple divisions multiple offerings multiple product capabilities and services multiple company customer segments.

So now you have all this data segmented probably siled and saying where do you start with all that is is gonna be daunting and you know, What we try to do is bring it back to the logical approach of what's the what's the business outcome you want? Let's prioritize those outcomes first. Then there's underlying data services and capabilities that we can go after to help you with that outcome.

That's the best place to start and then productize it and it and it sounds a little funny when we talk about it, but productizing the data making some responsible for that solution will eventually help you monetize it and give you a pattern for approaching the rest of the data. And so that's that's kind of the place that we're in now talking to customers about, you know, the approach initial approach to data. Sounds like an evolution of we used to talk about the data owner.

Right who's the owner now? It's like sort of like apis productizing apis productizing data, right? Yeah, you have to and you bring up a really good point in terms of data ownership.

Someone is always responsible for that data in terms of dispensation. How you share it Etc. But now we took the reason we talk about it as a product is because more than one constituent whether it's a customer or an internal constituent or you know, your colleagues Etc or third party needs access to that data for some purpose more than ever before.

So in the interconnected world, you you mentioned, you know, a product set or an digital experience is governed by multiple data points. Not just one more and so you want to mix them and match them and build something to its best efficacy to do that. How do you get out the data?

How do you share it? So there's your product owner data owner and you have to offer it like a product explain it like a Product and and I would say the other thing we're seeing is the more consumer driven approach to data consumption and if I could leave you with one thing out of a lot of this conversation, it would be how people are consuming the data and utilizing it both your customers and your employees and and that's an important thing to think about. How are your employees using the data to further your business?

That's gonna be one of your biggest challenges because you know, if you looked at the education marketplace, right and who's coming in as an employee and it's it's difficult to hire right? You will have people with quantitative Finance degrees. That's a real degree, by the way, and you know and you'll have a woman on your staff who knows python in addition to all our financial data skills, and now she's saying hey, I really like access to the data because I'm not in Excel anymore.

I'm writing up, you know python script to do what I need to do with the data or to prototype or to visual. Some way how do you impression or something to visualize it? Yeah.

How did she find it? How did she asks that one of the hardest problems? It's interesting too because I also put Partners in that as kind of a third part of that and it seems like one of the drivers when you put a strategy together around data.

It isn't just like this throw it in data Lake and put some you know Tableau or some tool on it and let people go farming right and I kind of go after it you that that's you know, inviting chaos. I think. Yeah, but there's oftentimes there's a sense in the business of we're sitting on a lot of x a lot of data about customers a lot of data about our history of manufacturing or whatever.

It is the business that you're in it seems like there's always this pent up demand of I wish we could leverage that in some way and of course you could flip that of add to the end of it to benefit our business and this way to improve the customer experience or get new customers. Whatever the drivers are is that is that a pretty common experience for you as well. It's a much more common experience these days and and I think you know using it for the benefit of the business we could talk about that a lot.

Ways I've heard it talked about in terms of value. What's the value of the data to the business in terms of product creation. I've also heard it talked about in terms of monetizing data not as much because that the connotation with monetize somewhere.

It's a bit pejorative people think one of you need sell the data. Yeah, like so my personal information. No, no we're talking about right?

Right, but we've had customers say before this is really a real customer use case say we think we're only monetizing 25% of our data meaning, you know, we know there's inherent value in there, but we of what we're able to access and create in terms of new products and customer experiences. We think we're just you know, a quarter of it is actually providing real value to us and that monetization comment is is within the last year so much much more prevalent much more so than the other m word which is modernization. And you know, we've we've heard that too.

So helping customers monetize your data get access and bring the value out. In one shape way shape or form where that is creating new product new service customer experience or we're selling it right? There may be some legitimate uses for the sale of data, but mostly it's you know, the former You're the a totally get because you could interpret monetization the wrong way.

The the mental model of that works for me is it's like capital and working capital, right? You have money in the bank and you have working capital that's helping you you're using you're getting value from it. It's putting it to work in ways that that's helping you customers your business all of that.

It's kind of what you're talking about with monetization. Yeah, you know, it's the difference between if you think about it. Doing something to the data and doing something with the data like you can store the data.

You can move it around that's doing something with the data. You can even modernize it and you know shift it somewhere but if you're not doing anything with it, right actually taking action on it, that's the key Point driver in monetization and and it's again modernizations on a bad word what I like I like data modernization but few if you think of it this way if you're modernizing your application stack only so you can get better access to the data and that tells me a story that you need to Modern, you know modernize your data stack or think about monetization that your data stack or shifted in a different way. So you can access that better.

Maybe leave your application alone, maybe focus on the data itself you're perspective about it's with the temporal nature the time nature of data because No news to you or any of us, right? We're living a world where things change faster and faster and I just technology but code that we deploy product and we're going to experiment with this new service or add something to an offering and do it in this region of whatever part of the globe that we're doing business in in our cloud. And and you know in the monitoring world we have observability now and we talk about, you know, a temporal nature of data because things things go away.

They just appear things change the codes change from the time. We saw whatever event occur that kind of happens with data too. Is that one of the challenges you're starting to see more customer struggle with you know, he it's it is definitely something they're struggle and in a multi-variate set of ways, right?

So they are yes, the data itself changes and and more over time as you point. As you pointed out just like anything else the underlying technology the storage medium of it changes and more over time. So you're dealing with several levels of change on data and you know, you could lose access to data because of that change, right?

We all know at least some of us who've been around for a little while know data stored on tape, you know may not be accessible anymore in the same way. So we need to move it or modernize our access it the data format. I mean no longer be supported.

Comment I would say right there's still a lot of support out there. But when I be common and more maybe difficult to convert or get at and then it may not be described in a really great way. So that makes it even more difficult to to get at it and then I talked about the technology difference a little while ago about what's your choice of, you know, data solution that you want today, you know pick it and every cloud provider has it in multiple versions of it and it's dizzying and then I think there's a the other part of the data changing and and over time which is It is it is ephemeral in some cases and it should be ephemerable and then there's the governance part of data that we don't often talk about but we do in regulated Industries a lot.

Right which is you have to keep data for a certain amount of time like a minimum 70 years if it's certain personal financial data within a company or or things like know your customer you have to keep it around, you know for regulatory purposes where you can't transport it if it's in the EU, it can't be moved to a us-based data storage medium. Those are things we you know about and then how do you expire the data to some cases? You want to get rid of the data?

Right? So you do want to change and more for or just be deleted of a time. So the question we're being asked now in terms of really you're really good question about what the change in nature data is how do we apply policy change to data so we know data changes and we have policies for great.

What is a way that we can apply those written policies in an automated way to data so that we're not caught flat footed in terms of That's formatting. It's dispensation. It's you know or its status in terms of deletion.

Interesting. I'm curious to You talked about treating data as a product. Imagine.

There's sort of a life cycle of people go to really thinking about it more strategically about data as they go through monetization, you know kind of processes. Are there a typical set of evolutionary steps or that people take and maybe you'd be have any without naming customers. But if you have any kind of story real world stories or might be relatable for folks there, you know there is and we've dealt with it quite recently.

I actually a couple customers who have gone through the process of looking for new capabilities of services to use their data with and it was a real pressing need to to the point where one of the customers was in a modernization kind of project and they stopped it because the outcome from modernization would have been too long for their potentially use case for the data which is hey we think we can use the data for customers to exchange information or product with each other and Represents, you know multi-million dollar kind of opportunity for us if we can get it done now. So how do we do that? But how do we approach that?

And what are the steps we take to approach it? And the good news is that there was half the battle was already done was knowing what they want to do with the data. The rule number one right always have an outcome and a desired outcome for it not just modernization or modernization for its own sake without some sort of outcome focus is always almost always do to failure like half the time I think so that was great.

You've identified a potential outcome. We work with you on the strategy to help you get it there. Now you have to find the data, right so finding it knowing where it's at is one thing accessing.

It is another. And then defining it in terms of its utilization and usage. And then moving on to actual action where we're making it, you know work for you over time at a product level.

So those are some of the steps that we recommend and we've taken with clients actually action the data and and today you know, one of those products is is in capability at the moment where clients are moving data with each other. They are trading and kind of a Marketplace and that value conversation and that data Journey, you know was was groundbreaking for the client. And now if you repeated enough times, right if you think about that one microcosm of how he's once set of data repeated with others, right?

So make this an area of approach you'll then build a value chain of data that will eventually have you modernize your entire data stack and kind of strangle away the old pattern and then give you a new capability. Interesting. I bet a lot of folks in relate to that.

Yeah, that's why I definitely can I'm curious, you know, a big part of our audiences is devops people doing yeah doing building software using those types of techniques and processes. How do you view data now and how that fits into a devops world an agile devops kind of environment. Yeah.

We're behinds data is behind we have to admit that and we've sort of seeing it come in the last five years you're in the conversation about data Ops data Ops data operations. Ml Ops and AI Ops if I can say that this is the probably the big buzzwords in the last couple years with it, right? It's still a nascent kind of capability, but it has to occur because we hamper devops processes, you know buying not keeping up with their Dynamic ability to you know, move product and releases and this friends that they do.

Reliably produce a set of results or capabilities. Dynamically, that's not being done with data and point. In fact, this is another customer example where they were looking for the ability to spin up data in a devops sort of way a data Ops away.

Hey, can you help me get data so I can build applications. I need test data good test data management principles and data Ops principles to build data capabilities and subsets that could iterate and build these new products and services that I'm trying to get to Market that already agile and devops enabled right? They're reading moving through a devops pipeline really well and so we were we were holding them up at this client this particular client by saying, can you hang on a second?

I do data refreshes every quarter. So I'm gonna put it I'll put it, you know, someone will spin it up for you and put it in another data storage layer. What do you mean a quarter?

I'm on a two-week Sprint. I cannot wait for you. You're holding my release cycle up.

That's a nightmare. That's Nightmare for devops professionals everywhere and I can guarantee you if we had a devops one of my devops partners on with us he or she would say the same thing right? It's it is a pernicious strain on devops developers if they are including data as part of their cycle.

Yeah, that's one of those to the included just go off and do their own do they get it sometimes and get it somewhere else and they do the most times they do wrong which you know, then my nightmare then it leads to data sprawl, you know data replication, you know on non-reliable copies data the whole nightmare of the data governance, you know professionals career is in sconced in what you know, independent devops developers can do when they go off on their own and spit up copies of data and you what date is the source of truth right now what you know, because my numbers your numbers don't work right here. Yes, worst case scenario were you you know, then they show the those numbers the business and they go where to get those numbers like is that true? And now the data original data owner has to answer a whole new set of questions.

Not a fun place to be. Yeah lots of true conversation. Yeah.

Well great. It's been fantastic talking with you and you parting thought is We need a phoenix project book for data the data Phoenix bread. I think we do.

It's been I mean I could talk about this all day long. I think this is a phenomenal conversation. But we do we kind of needed a project Phoenix book for data.

I think we would collect that right take us you I think we've got the first chapter written here exactly. We'll do a transcribe. We'll put it out there see me be down.

It's a little bit then we'll write the rest of it together. All right, it's been fantastic talking with you a working folks find out more about congino and the things that you do maybe some resources that you have absolutely they can find us that casino that I owe and that has a bunch of resources including the latest white papers capabilities and opinions we have about great Data Solutions. And how do they most often engage you as a consultative of conversation or is it's typically a consultative type of conversation and look these kind of conversations.

We're I am happy to have With clients or prospective clients, you know we too often it's very transactive in this world and you know, just this conversations like this alone can help guide you and instruct you and then typically if it goes deeper than this we do, you know advisory work where we help them really figure out a strategy all the way to execution. And and that's the main point is that you talking about the data and building it is strategies one thing but then executing a prioritizing it like you had mentioned like what do you focus on that's the thing. We're really good at which is helping you focus on this outcomes one at a time.

You know, what's what does this saying? Go eat the elephant? How do you eat the elephant Fork full at a time?

Right? Yeah. And with with the stick elephant, right?

All right great talking with you Tom Barton who is ahead of data the data practice that continue. Thanks again for joining us. Hope you come back.

We have some more to talk about we got another chapter. You're right absolutely would look good. Thank you.

Take care. All right.