How PrivOps Can Revolutionize Data Integration and Governance

Intro: [00:00:07] Broadcasting live from the Business RadioX studios in Sandy Springs, Georgia. It’s time for Sandy Springs Business Radio. Now, here’s your host.

Lee Kantor: [00:00:24] This episode of Sandy Springs Business Radio is brought to you by Mirability, providing unique ID solutions, leveraging cloud, AI and more to solve business problems. Here’s your host, Erik Boemanns.

Erik Boemanns: [00:00:38] Good morning. Yeah, this is Erik Boemanns with Mirability. And today we’ve got a great guest. His name is Tyler Johnson and he’s with the company PrivOps. So, let’s get started Tyler, maybe introduce yourself.

Tyler Johnson: [00:00:47] Hi, Eric. Thanks for having me on. I’m. You know, I know we’ve known each other for a couple of years, and I’m really excited about what you’re doing with Mirability. All right. So as far as as me, I’m an entrepreneur, an older entrepreneur that spent two decades in industry and then decided to take the plunge. And I’ve got a little bit of a history of innovation. For example, 20 years ago, I invented all the test automation that Hewlett Packard uses to test their Unix supercomputers. So really bringing lean agile techniques into software and test and it. So that’s kind of my specialty. I’ve also worked with a lot of the big players in the world, Rackspace and Microsoft and Google and so forth and so on.

Erik Boemanns: [00:01:34] Right, right. No, that’s a great background. And so, you decided to take the plunge. PriVops is the plunge. I assume that you took. And so tell us a little bit about what PrivOps does.

Tyler Johnson: [00:01:44] The the simple answer, the three word answer is DevOps for data. And you may not know what DevOps is. Think of it as as an assembly line, right? 125 years ago, we were using assembly lines to geometrically increase the rate of building physical goods like cars. Ten years ago they came up with DevOps. So Jean Kim and John Willis and Patrick Du Bois and a bunch of leaders in that space, and what they did was they applied assembly line type techniques to building software, and that’s called DevOps. So what we’ve done is we’ve extended that concept into data, which is why we call it DevOps for data. So think about what we do is enabling companies to build assembly lines for building integrations. So this isn’t speeds and feeds for data. This is how quickly can we add connections, change connections, remove connections so that we can get the best use of our data. In this world where we’re hampered by cybersecurity constraints, where we’re hampered by data privacy regulations. And just from the the scale problem, like how much data from how many sources do we need to be able to feed into artificial intelligence and other tooling to make those tools as useful as possible?

Erik Boemanns: [00:03:06] Got it. Yeah. And so you talk about data and you talk about all the sources. I’m curious. So maybe if there’s some examples of what those data sources could be, what kind of data are we talking about.

Tyler Johnson: [00:03:15] Well, it’s it’s everything. Data is a lifeblood of an organization. So you’ve got you know, obviously you’ve got customer data. So the interactions that you’ve had with your customers, who they are, um. You’ve got employee data. So your your your employees, your contractors, you’ve got contract data. Uh, things like expiration of contracts and terms and, uh, the actual storage of the contracts themselves. Uh, you’ve got product data. There’s there’s just a whole and one, one type of data people don’t think about is identity data. So, so, uh, so one of the things that we do that’s unique in this space is, is we use identity along with, uh, the provenance of the data. So where it was created, where it came from, uh, when it was created, those kinds of things. And we, we combine all that together, uh, to enable access to data based on both data provenance and identity.

Erik Boemanns: [00:04:19] Got it. And so hit go to take a step back to you talked about cybersecurity. You talked about some of the privacy. Obviously those are two things that are only increasing every single week at this point. Right? Between breaches between new cybersecurity laws getting passed. What is that connection that you see? We’ve got all this data. We’ve got to protect it. You talked about DevOps for data. Where’s the convergence point? How does this all come together in a way?

Tyler Johnson: [00:04:45] Well, it’s. I’ll give you an analogy, like building your house. Right. You have a contractor and you really want your general contractor that’s building your house to think about all of the plumbing and all the electrical as they build the house, knowing that the inspector is going to come and make sure everything was built right. So you don’t want plumbing or electrical to be an afterthought when you’re building your house. And it’s the same thing with data integration you don’t want. Security or data privacy or data governance to be an afterthought, that you’ve got to slap a band aid solution on right now. And that’s that’s really how most people think about things like data governance. We think about it as by design. We build it in from the ground floor. So. So as you extend your capability to move data from point A to point B, you’re able to have that security built in by design, have that data privacy built in by design, and that takes friction out of the process and makes you move faster.

Erik Boemanns: [00:05:54] Yeah, absolutely. And I think that’s a key thing. If you’re working with data, you’re working with some of the regulations that by design is almost required just for the sake of those listening. What would you what’s your definition of data governance? What does that term mean to you? And DevOps?

Tyler Johnson: [00:06:08] Well, data governance traditionally refers to the controls you put in place, making sure that data gets, uh, tracked and audited so that you’re not passing customer data where it shouldn’t go, for example. So that’s really more of an audit, uh, like a after the fact, you know, so that and then we build, uh, controls in place. And those are and what’s different about what we’re doing is since we’re automating everything end to end, uh, we already have all of that tracking, and we can demonstrate that that data was not, uh, wrongfully accessed for something that wasn’t a legitimate use of that consumer’s data. So from my perspective, data governance is about. Control through automation, whereas the traditional term for definition of data governance is having controls in the place to minimize the chances of things going wrong.

Erik Boemanns: [00:07:07] Right. That makes sense. And I think the automation is a key point. I’m curious if you have an example or a story of where that automation helped a business process or helped help somebody manage their data better.

Tyler Johnson: [00:07:17] Yeah, absolutely. So we we have a customer that’s a mortgage bank. And they wanted to their regional and they wanted to expand into California. And California passed a data privacy law called the CcpA, the California Consumer Privacy Act. And then there was a follow on Cpra. Whatever its legal stuff, Eric, that’s that’s your bailiwick, not mine. But, uh, but yeah, so we wanted to be able to make sure that we could still allow our salespeople to have access to their customers data and to be able to use that data to create a better experience for their customers. But then we now have this new regulation to contend with. And so what we did was we replaced all of the company’s data integrations for reporting and analytics, uh, using our automated data fabric solution. And what that gave them was, uh, first of all, it it dramatically improved their ability to get data to where they could analyze it. But what it also did was because we had the data privacy built in by design, uh, we could open a branch in California without having to worry about going through a bunch of current state assessment of, uh, you know, our data governance maturity and rationalize that that to the regulations in California. We don’t have to do any of that. That all went away because we knew that if California comes back to us, to the bank and says, you know, hey, prove to me that if a California residents asked to or opts out of having their data processed or used for sales and marketing purposes, prove to me you do that. And that’s with the automation that becomes a no brainer and it takes friction out of the system.

Erik Boemanns: [00:09:19] Got it. And I think that makes sense. So I’ve done data development. I’ve done data integration work for decades now, going all the way back to having to handwrite these things in scripts. And data integration to me is moving data from system A right, which could be the sales system or the customer system, that people are storing all their customer data in into whatever system B, C, D, whatever other systems there are reporting analytics today. An AI system. Right. Um. And so in all those years, I had to do. Um, software development to move the data. Right? It was first it was scripts, then it was later using packaged software that that I could drag and drop shapes and then I could start to move the data. But it was still always software development. Where does DevOps fit into that model? How does it what does it replace that those traditional data developers are doing?

Tyler Johnson: [00:10:10] Well, I’ll go back to my first comments about bringing Lean and Agile into data. Uh, the idea is to look at this holistically, start with the process first, and then think about how the technology could be used to do that. And when I think about process, if you if, uh, those of you that are familiar with Lean and Six Sigma, uh, know that that lean is all about eliminating waste and processes, it’s all about efficiency. So if you can take a step of the process and building a data integration where you have, uh, a data engineer or someone, uh, with, with Eric’s skill set, for example, going in and, and building this stuff manually, uh, writing code that’s, that’s time intensive. Uh, Eric’s expensive. Uh, you know, it takes a lot to be able to do that. So one element of what we do in eliminating waste in the overall process is, is to incorporate low code into building data interfaces. And so what that does is that, uh, dramatically increases the productivity of your data engineering capability. Because you can you can take your most talented engineers and put them on value add work, uh, and then use the low code interface to take care of all the plumbing on the back end. Uh, and so you can do a task switching, uh, bringing in other folks that might not be as, as, as senior in terms of skills and talent.

Erik Boemanns: [00:11:44] Yeah. Well, that’s interesting because even though some of the products I was using were marketed as low code, right, because they had drag and drop kind of interfaces, they still weren’t. They were still you still had to code or you still had to understand what was happening. So it sounds like we’re shifting, that you still had to be a very senior data engineer to use these products, even though they were visual.

Tyler Johnson: [00:12:06] Yeah, and that’s why I said that it’s low code and not no code. Yeah. Uh, it’s all about the most critical metric with data integration is the rate at which you can build and maintain data pipelines. And that is, uh, that’s it’s unfortunate that that is not seen as a critical metric with a lot of organizations today. Uh, because there’s an implicit assumption underneath that, uh, that we’ve got to work with the processes that we have. And that’s why I’m going to always go back to starting with process and then thinking about your technology architecture, not the other way around.

Erik Boemanns: [00:12:45] Right. Well, I think the other thing as we’re thinking. Layer back in privacy layer back in security. None of those products I was talking about even worried about that. Right. That was not their problem. Right? They moved data. That’s all they think about. I, as the engineer, maybe thought about security, but even so, it’s another level up before that even gets talked about. If we’re thinking about that old way I think previous approaches that differently. Right. If we’re talking about privacy by design, security by design.

Tyler Johnson: [00:13:12] Yeah, exactly. So if, you know, for those of us that are familiar with DevOps, it’s it’s DevSecOps, right? It’s being able to shift left. And I don’t want to use too much of the jargon from the DevOps world. Sure. But but basically it’s building these things out from the get go. And, you know, I’ll go back to the the the house construction analogy. Um, you know, you want to put the, the wiring in before you lay the drywall and paint. Yep.

Erik Boemanns: [00:13:41] Exactly. Yeah. Are there other? Um, stories are examples where where you’ve seen that really work well, where taking that approach of. Privacy by design. Applying it to your data governance, the way we think about it today. Um, just some examples of where that has made life easier, better, faster.

Tyler Johnson: [00:14:03] Well, yeah. So I’ve come up with a list of 19 different, um, process requirements for employing. Uh, basically eliminating waste in the process of getting data safely from A to B and managing all of that mess. Right. And low code is one. We’ve already talked about that. Um, the other thing is, is another one is modularity. So being able to write code or have our UI, uh, do it where we’re able to automatically switch up data pipelines and move things around rapidly. And what that does is, is, is like, okay, well, yeah, that makes that makes your engineers more efficient, of course. But there’s another thing that it does that’s actually even more important from more of a business or a program perspective, which is that by having that type of approach, you have separated your different projects. It’s separated your work streams. You’ve you’ve minimized or eliminated dependencies. So I’ll give you an example. So on a Thursday afternoon we were talking with a VP of applications. So she had all the application developers and managers under underneath her. And we were the the meeting was to kick off a migration of CRM. So moving to a different customer relationship management platform. And one of the things that just came out was that, uh, she told me that, hey, you know, the next week or so is going to be hard for us because over the weekend, we’re going to release the updated version of the accounting application across the company. And so I mentioned to her that, um, well, that new version has a different database technology on the back end.

Tyler Johnson: [00:16:02] So Monday morning when the CEO comes into the office, his dashboards aren’t going to have the sales data from over the weekend. And that’s going to be a problem, because all of those integrations with the accounting data and the dashboards all will break because, you know, there’s you know, we just went from Oracle to MySQL on the back end and none of that’s going to work. Uh, so because we have these things like low code and, and, and modular architectures and data integration as code where we can move rapidly, we were able to complete that migration of those of, of pulling that accounting data, um, over the weekend. Uh, so, so the result is, is that the VP of applications doesn’t have to go to the CIO, or the CIO doesn’t have to go to the CEO and say, hey, by the way, that big accounting, uh, upgrade that we’ve been working for the last six months on, that we’re supposed to release this this weekend? Well, we can’t do it because there was a dependency we hadn’t accounted for. And, you know, you don’t usually think about data governance or data privacy by design as a way to save face with a CEO. But it absolutely does that because of that fact that it’s able to isolate those different work streams. So yeah, there was a dependency there we forgot about, but we’re so agile that we’re able to account for that and, and continue to make progress unabated.

Erik Boemanns: [00:17:38] So it sounds like basically there is a technology problem that the CEO never found out about because it never actually became a problem.

Tyler Johnson: [00:17:45] Yes. Now. You got to be careful with that. Because if you’re if you’re so good, nobody sees the problems. Exactly. Then then they’re like, well, why am I paying for this thing? So, you know, so that that becomes an issue as well. But if, if as a leader, your focus is on. The competitiveness of your organization as a whole, like you’re in it together, then that’s the approach you’re going to take every time.

Erik Boemanns: [00:18:16] Yeah. Yeah, exactly. We we want to not have problems, but at the same time, we want to make sure people know that we’re solving problems. Right. That could happen.

Tyler Johnson: [00:18:25] Yeah. And the good news is there are by by. Going down this pathway and thinking about efficiency and eliminating waste and and getting data from A to B. Um, you’re able to show a whole lot of business results that you wouldn’t have otherwise been able to show. So less, uh, less bad results. More good results. So that’s where you want to be?

Erik Boemanns: [00:18:53] Exactly. I think that’s great. The, um, a couple of questions just to get some quick, um, quick thoughts on buzzwords in the industry for folks that are that are seeing these buzzwords. Right. Um, Microsoft had a huge marketing push on the term fabric, right? They said, look at all of our fabric that we have. And I don’t know if you actually, I think you had to go three paragraphs down before you realized they were talking about data. Right? In those marketing messages. What is what’s your take on fabric data, fabric, Microsoft using that term?

Tyler Johnson: [00:19:23] Well, I don’t want to pick on Microsoft specifically. I think that’s more of a general vendor marketing. Um. Attributes of what the big vendors do. Over time, they will rebrand their technology stacks and having come from the big vendor space as as a sales person, as a engineer and as a product manager, I understand that that’s important from the vendor perspective because it gives you an opportunity to go back to your customers and say, we’ve we’ve got this new capabilities, and sometimes those new capabilities get lost in the messaging when, uh, when you’ve got the same branding for, for all of your technology. So if you look at Microsoft and their fabric branding, the fact is they, you know, they’ve got Synapse and Data Factory and Logic Apps and all these different things that they’ve been working on for all these years. So now the data fabric has become kind of a term. Um, they want to they, they want to message that they’re making progress out to the marketplace and changing that. Branding is is is a way. So like moving from Azure Active Directory to enter ID, right. Exactly. Uh, that kind of thing. And, and I can pick on any vendor out there and tell the tell the same story. Uh, what I would what I would say is that in this space, because it’s so rapidly evolving that you should avoid using jargon whenever possible. Period. Full stop. Right. Uh, here’s an exception, though, that as a customer, you want to look at, if somebody says data fabric, uh, there’s lots of different definitions.

Tyler Johnson: [00:21:07] Right. So, uh, Gartner calls it a, an emerging architecture and data management strategy. Um, you know, last week I presented at the Air Force’s chief data office, put, put a I data analytics forum together, and I presented there and we were talking about the Department of the Air Force data fabric. And so they they’ve got what they call the big six data platforms in the Air Force. And their data fabric was uh, basically they built some point to point integrations between six different big data warehouses. Uh, so their definition, you could say, well, it’s it’s a, um, you know, integration of multiple, uh, integration on top of integration. Right. Uh, and then, okay, I’ll give you another one. So NetApp. So I spent a couple of years working for NetApp. I love the company. I love the technology. Um, lots of great stories from from those days. Uh, they have what they call the NetApp data fabric and what NetApp definition of a data fabric is. Um, I think it’s it’s something along the lines of software defined hybrid cloud management for storage. So being able to use software defined methods to manage your storage, whether it’s on premise or in the cloud. And that’s great. That’s here’s that’s a third definition. Um, in fact, you can go look at what IBM Red hat is saying about data fabrics. Uh, you know, snowflake, uh, you know, Accenture, Deloitte. Everybody’s got their own little take on that. So here’s my take on what a data fabric is. Uh, it’s an assembly line for building data pipelines.

Erik Boemanns: [00:22:56] Got it. Okay.

Tyler Johnson: [00:22:57] And what that means is the focus is on process, right? Is is how how do we build and manage data pipelines at a scale that’s exponentially greater than where we are today? Because that’s what we’re going to need for I, by the way. Yes. And and you think about all those different definitions of data fabric. There’s there’s something you can glean from that. And the common thread through all of that is that each. Definers definition of data. Fabric says something about what they’re trying to accomplish, what their agenda is. Uh, you know, with Microsoft, it’s, you know, hey, we’re. Were the leaders. So data fabric. Right? Exactly. Gardner. Emerging where the thought leaders. We want you to come to us for thinking about what the future is going to be for the Air Force. It’s we want to show to Congress the American people that we’re making progress. Right. And then NetApp, of course, NetApp wants to sell storage. Right. So software. So for me, when I say data, fabric is the actual the actual fabric, not the methodology or architecture, because I’ve already built it, it’s already in place. But my agenda is that assembly line idea, which is we want to get to where we’re 1,000% faster than we are today, ten x. And the only way we’re going to do that is eliminating waste in process. And then the technology choices we make are ones that support those changes.

Erik Boemanns: [00:24:29] Makes complete sense as as we get to this point. I’m curious. What about DevOps? Do we did we miss what? What do you want to make sure we talk about today?

Tyler Johnson: [00:24:41] Well, there’s. This is an emerging space. Data fabric is, you know, beyond the. We talked at length about how overused and confusing the term is, but conceptually, the idea of having to transform the way we’ve moved data as a lifeblood of the organization from different points to maximize things like customer experience that’s new. So, you know, in the last year, I’ve met dozens of CIOs and VP’s and directors and individual contributors in this space. And and what I found is that the the enemy isn’t necessarily, um. For? For Prevost, by the way, the enemy isn’t the competition. The enemy is a status quo that we’ve already got. Uh, we’ve we’ve already got some stuff in here. And, um, you know, it’s it’s we’re kind of like where DevOps was with software development a decade ago. Uh, where, you know, ten years ago, your software teams, you know, they don’t want to hear about operations. They want to just write code and kick it over the fence and let the operations people deal with it. Right. So, and the idea that we’re going to create more automation and people are afraid of automation because they think that they’re going to lose their jobs, which isn’t true, by the way.

Tyler Johnson: [00:26:09] Right. Um, they’re afraid of that. Uh, so what you’ve got to think about is, well, what is the effects that transforming the way that you’re managing data in your environment? What are the cultural effects? How to how does that affect, you know, do you have the like the the right leaders in in in place, right. Yeah. Uh, do you have the right partners in place? And I think one of the things that kind of gets lost in the message with DevOps is that even though we’re a technology platform and, you know, I built this whole methodology around, you know, eliminating waste and, and data management. Uh, what’s really the most important thing is, is, is mapping out that strategy and that roadmap and thinking about, uh, what kind of waves that could potentially make in the organization. Yes. Because, you know, it takes a while to change. It takes much longer to change culture than it does to swap out a tool.

Erik Boemanns: [00:27:13] Great point. Yeah.

Tyler Johnson: [00:27:14] And and it’s unfortunate, but my prediction is that, um, that since since data is the lifeblood of any AI strategy, uh. You know, people are not paying attention to getting the data into AI as part of their strategy. If you look at the if you look at the some of the research today and look at, you know what? What are the most important elements of your AI strategy? Data governance is like at the forgotten, forgotten. But it’s like if if AI is like a light bulb. Um, you really should take care to make sure that your switch works and your wiring works and all your electrical works, because otherwise you’re not going to be able to turn the light on. Yes. So that’s really where I think that, uh, DevOps can help and, and Eric can help, uh, with, with thinking about how that affects things holistically. And then, of course, tying all of that at the end of the day to the bottom line. So revenue and profits.

Erik Boemanns: [00:28:23] Good, good idea to tie that back to the actual revenue, right?

Tyler Johnson: [00:28:27] Well, yeah, you have to do that constantly. Exactly. And that’s that’s one of the things people don’t understand about agile is that and a lot of times we’re agile fails is because they don’t do that. They take a big waterfall project and put it in chunks and say, okay, we’re agile. It’s like, no, at the end of two weeks, you need to be able to show your customers and your other stakeholders and improvement in their life and an improvement, and that improvement needs to happen every two weeks.

Erik Boemanns: [00:28:54] Exactly. Yeah. Strongly related to continuous improvement. Right.

Tyler Johnson: [00:28:58] So absolutely. Yeah, CI is a big part of lean.

Erik Boemanns: [00:29:01] Exactly. So how can people get in touch with you if they want to learn more about DevOps?

Tyler Johnson: [00:29:07] Uh, well, fortunately for us, we trademarked the name a while back. So if you just type DevOps, it takes you right to the website.

Erik Boemanns: [00:29:15] That’s easy.

Tyler Johnson: [00:29:16] That’s first entry for DevOps Priv ops. Uh, I’m also on LinkedIn. I post every now and then. Um, and, you know, I’m also a blogger, so I’ve got a lot of if you’re interested in more in depth information around this, I’ve got several blog posts I can point you to. So, you know, and then also, I’m always happy to have a conversation so free. Feel free to reach out to me directly. You can reach out through the website or through LinkedIn. My email is Tyler Johnson at devops.com. Uh, so yeah.

Erik Boemanns: [00:29:53] Okay. Great. Any parting thoughts?

Tyler Johnson: [00:29:56] Uh, no. Just that it’s great to see you, my friend.

Erik Boemanns: [00:30:00] Yeah. Good. Good talk. I enjoyed having you here today.

Tyler Johnson: [00:30:02] Yeah, it’s been fun.

Erik Boemanns: [00:30:03] Yeah. All right. Thank you, thank you.