Speaker 1: ASCII Anything, a podcast presented by Moser Consulting. Join us every Wednesday to find out who from Moser's more than 200 resident experts we'll be talking to and what they're focused on at the moment. Trends, security, set up, ASCII Anything and we'll give you our best answers.
Speaker 2: (singing).
Angel Leon: Hello everyone and welcome to another episode of ASCII Anything, presented by Moser Consulting. I'm your host on Angel Leon, Moser's HR advisor. Today's episode continues the conversation between Shaun McAdams and Warren Sifre, two of Moser's top data analytics experts. Shaun is Moser's vice president of data analytics and Warren is director of strategy within our data analytics group. They're ready to dive deeper into defining a data and analytics strategy for your organization and today the focus is on how that process works. Shaun McAdams came to Moser in 2015 to help establish a division focused on data and analytics. Serving as principal consultant engagement manager, director and now a vice- president Shaun has helped advance the data analytics services to over 80 clients and 100 engagements across many diverse platforms. Warren Sifre has been involved in the IT community since 1998 and has worked in a variety of industries. Warren is chapter leader for a couple of user groups and speaks at many conferences and user groups throughout the United States. Warren has a passion for data architectures and solutions. Serving as director, he is helping advance the adoption of true data strategy in every engagement. Without further ado, here are Shaun McAdams and Warren Sifre.
Warren Sifre: All right, Angel, thank you very much for the introductions. Shaun, this is a session number two for us.
Shaun McAdams: Yay.
Warren Sifre: We're here to kind of continue where we left off. In our last session, we kind of talked about the four major things that lead up to a IT strategy and a data strategy and data analytics strategy and all those pieces. And one of the things we identified was kind of what we felt was the biggest gap in pieces. And we kind of agreed. We kind of went down the path of kind of debating a little bit and wound up with process. I think what we kind of want to do is kind of bring about what is this process thing? What do we mean by process? What are some of the things? As throughout your experience, what are some of the things that you've noticed that kind of led you down the path of identifying process as a gap and the reason why it's the first thing we're talking?
Shaun McAdams: Well, I think there's a couple of things. One of them is I feel pretty strongly that if your organization wants to create a data driven culture, that it has to be done by setting the mindset of the people that deliver those types of products. I think that's one thing that's important. That's real down low, that's that those micro level. I know we'll get to that. The other thing would just be how organizations deliver these products and not understanding who has responsibility over scenarios. Things like data governance. We would look at data management principles of which data governance is a part of. And we would define data governance as the policy, the strategy, compliance that the organization wants to adopt, but who owns those? Who is responsible for applying those policy? And then who enforces them? And usually you have a lot of different units within an organization that play a part in delivering these data or analytic products and so not having those things defined, I think is one of the biggest challenges.
Warren Sifre: Yeah. It's governance is intended to sort of protect the organization from itself and from outside influences with regards to HIPAA and GDPR and all these different statutes that are intended to protect the consumer. Yeah, I could see how data governance is really a big thing when it comes to that. One that I like to reference when it comes to process is dealing with shadow IT. Those of you that have been in the industry, you've found yourself in situations where an organizational unit needs some thing and there's a priority discrepancy. They need it today, you're not able to get to it until a month from now or resourcing or whatever. There's logistics in the way that encourage that business unit to take it upon themselves to actually implement whatever solution they need because they need it now. And a lot of times what happens is that once that happens with one business unit, it starts to spread a little bit. The sentiment of, oh, you're able to get that done quickly? Well, I'm going to do it too and hopefully I can ask for forgiveness later. I got my own budget, I can spend it in a way. Why do I need permission? I'm going to hire my employee that's going to be this data analyst and in reality, they're a BI developer and they're building these solutions. What tends to happen when shadow IT comes into play is we may be brought in to say," Hey, you know what? We need to corral this. We got this proliferation of Microsoft Access databases all over the place."
Shaun McAdams: We've done it a few times.
Warren Sifre: And you got thousands of them. Now it's, all right, we can go in there and solve the problem of replacing Microsoft Access and giving you a better solution. But if we don't go down to the root.
Shaun McAdams: Why you got there.
Warren Sifre: And understand why we're there and how we got there, that means it's going to surface again, three to four years from now, if not sooner, you're going to be calling us again to help you out again and mind you, we don't mind helping you out in that way, but we like to make sure we're helping you out in a progressive nature that takes you to that next level of data maturity, analytical maturity. Gives you that ability to jump and launch that platform. What are your thoughts? Have you seen some of that too?
Shaun McAdams: Yeah. And I think going with our vision of wanting to be a trusted advisor, it's digging into, well, how did you get to this point? Not just helping correct those things, maybe putting band- aids on them, so to speak, but getting back to that particular root cause. The other thing I will say though, most companies have already started something in this space. They're doing something as it relates to data and analytics and a lot of the things that we talk about as far as how you deliver these, they may not be doing it correctly or as efficiently so it is always going to be going back and looking at how do we apply these things? How do we correct it? How do we pull back into data governance? How do we eliminate shadow IT? I would say one thing about shadow IT is you're looking at the organization's trying to get value out of data. If shadow IT is related to data and analytics, we want to get to that point. That's where we want to get to, but we want to do it in such a way where we first set those policies of how we want to act. We educate everyone on how to do it and then we allow them to do it. This democratize of data that we hear about, empower people, we're all for that, but not day one.
Warren Sifre: And ultimately what you're kind of describing there is that evolution from analytics as a service, to data as a service and that evolution piece. Because organizations, when they find themselves wanting to develop an IT strategy and they're trying to eliminate shadow IT and bring data governance and some of the other facets that go along with that, they find themselves wanting to control everything. And in a way you need to because you need to fully understand what are the requirements? What are the needs? How are things going to work? Build the processes in place so that internally you understand what the work and how the work's going to be delivered, what people you need, what are the gates of checks before things go to production at UAT? And how is data certified? And made sure that it's available and secure and tagged. All those pieces. And then once you have that, kind of maturing to the next level where you can promote the intent behind shadow IT is to give the business units the ability to run at their pace. And they're just taking it versus us giving it right. And that's the change. That's a philosophical change that I think this framework that we have comes into play. Share.
Shaun McAdams: Yeah, basically, when we look at how we deliver data and analytic products, we want to look at that macro level. We did that the last podcast and we talked about these four pillars. Process was one of those. Still staying at the macro level, how do you deliver these products? And then you apply the mindset or the approach, which I know we'll get to in the next one. We look at delivery systems as being called insights, engineering and platforms. Insights we would say is anything that consumes data in order to produce information. We're not really picky about the maturity level of that particular insight. If it's a file you send to an application or a person or a tabular rapport, infographic, visualization, if you're using some type of predictive modeling. We don't care about any of that. What's important is we understand that we have a delivery channel that's taking data, turning it into information. Engineering, it's data engineering. It's how do you acquire information from all these operational systems? And then what process are you going to follow? Or what stages do you want to introduce on data so that you can increase the level of confidence? You can introduce data quality. Where are those things are going to take place? Ultimately to get into the ability to serve it into this insights delivery channel. And then the last would be platforms. You obviously need the tools on which, whether people are doing insights work or engineering work, they're using particular sets of technology. What is that? Where is it located? And so all of that work, we would classify as platforms. Again, boiling things down like we did from the last podcast to what are the things you need to do this well? We would want to map the delivery of data and analytic products across insights, engineering and platforms.
Warren Sifre: And that's an interesting formula because in most organizations, they have a BI team or they have a data team, a DBA team, an infrastructure team, all these different teams within IT And they need to be working together in some way. And most of them do. They pass tickets back and forth using some tool or hey, you know what? I'm done with my part, it's your turn. But how do we do that from a conceptual standpoint with regards to the data? And I think this insights, engineering and platform concept kind of gives you that logical framework to be able to start working through what that looks like. Because I think this is a platform that if you're able to logically create that segmentation and handoff of work amongst that, then what can happen is that you can truly have and preserve some of the segmented teams that exist. But now we can say," You know what? Here's how work's going to get delivered." The insights team, they need a report. They go out there and they try and see what's available. And they say," You know what? I want to build this report, but I need data from this ERP system and I don't have it." That would mean someone in data engineering needs to get that work and be like," All right, how do I bring this in?" And it's like, oh, guess what? We're hitting a piece of technology, we're trying to pull data from a proprietary data format that we don't have built into our platform yet so we don't have the tools to be able to bring that in. That goes down a platform and then a platform says," All right, we need to find the tool to meet this need." And then they go through it. They figure it out whatever evaluation process they do. They find the tool, they bring it into the ecosystem. They make sure it's compliant, they make sure of the security, they make sure it doesn't infringe on any of the governance and existing processes that are in place. And then it goes back up the stack. And as an organization starts to mold their mindset to this handoff concept and these three different pillars, now you can truly have one team own it all. But if you philosophically understand that that's what's happening and you're able to put in process that does that, when you find yourself that you've outgrown this one team and you need more insights people, there's more requests for reports, there's more things to happen than what your team can handle, you can start handing it off to the business. And this kind of takes you to that progression and that big thing that's been happening over the last five years, that's framed or termed self service BI.
Shaun McAdams: Yeah, self service BI.
Warren Sifre: And that's a natural progression. You own everything and over time, you're going to want to not own it. And self service BI in a way, shadow IT.
Shaun McAdams: Yeah, it is. But if you can do it following the paradigm we're talking about, I wouldn't classify it as shadow IT because everyone's not out doing their own thing in their own way. I look at it, if you were to take kind of a reverse triangle, where I would want to get to, if you were to take a reverse triangle and can kind of cut it into three sections that lower bottom of that triangle would be platform. You don't need a whole lot of human capital to take care of that. It's not going to change super often. It's going to change, like you said, as particular needs manifest themselves for folks that are doing engineering, for folks that are creating insights. Engineering based upon the amount of information in the organization, how that's being curated and following a process that I know we'll get into for just the engineering, going to be a little bit larger of a team. May still just be one team. Kind of depends on the organization. But I think a lot of organizations want to get to that top layer of that triangle, that insights and there are a lot of teams, there are a lot of consumers of it and that's great because usually those consumers are very business focused. They have that particular set of domain expertise, they're trying to use data to solve a particular question, answer a question. And putting the tools in their hands is great, but putting the tools in their hands without how they use it, I might as well go and give my vehicle to my 16 year old and say," Hey, have at it." I'm not even going to train them or teach them or do anything and expect that they're going to be just fine. They're not going to have any accidents. It's just not realistic.
Warren Sifre: I take it your vehicle's stick, right?
Shaun McAdams: Yeah. I kind of look at it that way. And that does follow what you talked about, these as a service models. Analytics as a service, data as a service or platform as a service. But I agree with you a 100% in the philosophy rather than kind of going from the bottom to the top. Create a platform, then a data as a service, then an analytics as a service model, you really have to go kind of top down in this case because you want those data governance principles and policies and best practices. You want knowledge management to exist somewhere so that you can educate and handoff analytics to the business and not just sort of here's the keys to the vehicle and then go. Approaching it as, let's do everything analytic as a service and then let's back off into a data as a service and then maybe for some organizations, it also makes sense to back off to platform as a service.
Warren Sifre: And this goes back to trying to begin with the end of mind or inaudible. You find yourself, what do you want to be when you grow up as an IT organization when it comes to your data strategy? Do you want to be the one that's constantly responsible with building every single report? Always under the gun, always having to do all the BA work and going back and forth and trying to navigate all the resources and prioritizing and anything else. And then of course always being looked down on because you're not moving as fast as the business unit wants to. Or do you want to eventually find yourself, hey, you know what? I want to be able to decouple this and give the business units that have chosen to take that responsibility and the ownership, given the opportunity to do that. And maybe still have a team internally that can help those that don't want that. Because different business units have different needs. If you're talking about a supply chain environment, you may have some people that are really big on analytical statistical things and they want to build their models. They want access to the data. They just want to run at a 100 miles an hour and build these predictive models. And you've got the other side, you know what? I just need to make sure I got my material then. I just want to make sure people clock in. I want to make sure how much overtime I'm being used. I don't need any range or sophistication. Again, this goes back to, depending on the type of environment, you can have that progression, but you got to understand what is the target that you're trying to hit in that three to five year plan? So that as you are putting these processes in place and as you are finding the right people to help you with these processes, you can make your way to that and be able to build off of itself versus constantly refactoring because we didn't think far enough ahead.
Shaun McAdams: Yeah. And most of the clients that we get involved with I'd say, they're in that frame of mind. We really haven't thought about what we want it to be. Not with enough knowledge to understand how to even define that. The other thing I think you hit on, which is important, as you matured as an organization and you're looking to add analytics into the organization, there are dependencies. You alluded to this, there are dependencies that exist between insights and engineering and platforms. And particularly between insights and engineering. As you're acquiring data is into this environment, how are you going to use it? What type of models are you going to? Over time, that churn is going to be reduced because you're going to have most, let's say, of the data within the organization, into the environment, most of it flowing up through our stages. Most of it refined into some usable data models. The requests from these insights teams into engineering for different types of data to go through the entire process, those are going to start to become less and less and that's also a good key indicator when the organization is ready, to allow different parts of the company to kind of become more self sufficient in this BI as a service.
Warren Sifre: Yeah. It's an evolution point. It means, hey, you've reached a particular maturity level organically. It wasn't one of those things that you're going to do this now. You do it over time. And this may seem like a big undertaking and in essence it is, but like anything else, we're going to give you the tools over the next few podcasts to kind of how to be able to break this down in a way that makes it manageable, what are some approaches that make it something feasible? And how to get started.
Shaun McAdams: And financially responsible. That's the other big part of it. You look at the strategy we're talking about and we're saying," Hey, small investment first. One team. We're doing these things the right way. We're getting more people more involved over time." It's not, let's bring everyone in the organization in, let's get these tools in. Everybody, here's your BI tool and then later you're trying to figure out all of these things because you have these collisions that happened. The challenge, I think though, in what we're talking about, where stuff is macro level, is kind of going down. Going down to the micro level. And when you're presenting it to the organization, the people that are going to do the work, whether it's platform, engineering or insights, what's the approach? What's the mentality they need in order to deliver these? You even have to take a step down and I know that's what we're going to get into next podcast.
Warren Sifre: Exactly. Exactly. Thank you for tuning in. This is Warren.
Shaun McAdams: And I'm Shaun.
Warren Sifre: Thank you. Have a wonderful day.
Angel Leon: And that's our episode today of ASCII Anything. We hope you enjoyed Shaun and Warren's conversation on data and analytics. We will be back next week with another episode of ASCII Anything, presented by Moser Consulting, your technology partner. Until then, so long everybody.
Speaker 2: (singing).