The Next Big Question

Drew Lazzara (00:14):

Welcome to The Next Big Question, a weekly podcast with senior business leaders, sharing their vision for tomorrow, brought to you by Evanta, a Gartner company.

Liz Ramey (00:24):

Each episode features a conversation with C-suite executives about the future of their roles, organizations, and industries.

Drew Lazzara (00:33):

My name is Drew Lazzara.

Liz Ramey (00:35):

And I'm Liz Ramey. We're your co-hosts. So Drew, what's the next big question?

Drew Lazzara (00:40):

Well, Liz, this week things get a little bit meta on the show because we're asking, how can leaders help their organizations ask the right big questions? To help us take on this subject is Matt Griffiths, vice president of data-driven transformation at Stanley Black and Decker. Before taking on this unique role, Matt served as the chief information officer at Stanley Industrial and Biogen, and also held several leadership positions at Dell. Throughout his career, Matt has recognized the potential of data as a conduit to insight and better decision-making. But as with all things data, what you get out depends on what you put in. As Matt has led Stanley on its data journey, the right starting point always begins with asking the right questions. In our conversation, Matt discusses his approach to encouraging the right questions and shares some of the exciting places the answers to those questions have led the organization. 

Before we talk with Matt this week, we want to take a moment to thank you for listening. To make sure you don't miss out on the next Next Big Question, be sure to subscribe to the show on Apple podcasts, Spotify, Stitcher, or wherever you listen. Please rate and review the show so we can continue to grow and improve. Thanks, and enjoy. 

Matt Griffiths, welcome to The Next Big Question. Thanks so much for being here.

Matt Griffiths (02:06):

Thank you, Drew, for having me.

Liz Ramey (02:08):

Great, Matt, we're really looking forward to getting to know you on a professional level and understanding a little bit more about your objectives as a data and analytics leader. But, I know that there's a really fun fact about you, and I believe that you're a Lego Master. Is that correct?

Matt Griffiths (02:28):

I have definitely been interested in Lego for many, many years since I was a small child, including coaching a number of Lego robotics teams. My first job actually after leaving college was building automated guided vehicles. And, I was fortunate enough to be able to coach a number of middle and high school teams building Lego project-related activities when I lived in Texas for a while. So, yes.

Drew Lazzara (02:59):

What are the biggest fundamentals of Lego coaching? You know, like my basketball coach would make me keep my head up when I'm dribbling and use my fingertips. Are there any good Lego fundamentals that you can share with people?

Matt Griffiths (03:09):

Well, I would say, I think the biggest one would be actually something that my first boss at that first job that I mentioned earlier first said to me, which was 'simple solutions are the best solutions.' And that's probably something that I have taken with me through my entire career and something that I have applied with the kids that I've coached through those Lego robotics competitions. You know, 9, 10, 11 year-olds tend to want to invent some very creative solutions, and there's always a time and a place for that, but sometimes just kind of keeping things simple is in itself innovative and often delivers the best results. And, so I think, you know, to answer your question, it's very much a case of channeling all of that energy that those kids have for Lego and for robotics in the right direction and allowing them to be successful.

Liz Ramey (04:07):

That's so cool. My son's school actually has a Lego robotics class after school, and he just turned nine. So he's now allowed in it, but unfortunately, nothing in person right now, but when we open back up, I'm sure he will be right there. So that's great. What's the, if you recall, what's the kind of coolest thing that you've ever built?

Matt Griffiths (04:30):

Personally? When I was a kid, I used to have all kinds of Lego kits, and I had a train set, an electric train set that ran all the way around my bedroom, which I used to have great fun with. As a coach with the Lego robotics team, we did a number of really cool projects and tasks that we have to solve through building a robot or a vehicle that would go around and be able to do a multitude of different things. And, both myself and the kids on the team had great fun building and kind of accomplishing those. It was all very programmable and really focused on technology, as well. So great, great initiative to be involved in.

Drew Lazzara (05:15):

I recently got a full, life-sized BB8 droid, a Lego BB8 droid. So, that's waiting for me. That's a project to take on very soon. Appropriate for my age, I think.

Matt Griffiths (05:26):

Appropriate for the current climate where you're probably looking for things to do inside your house.

Drew Lazzara (05:31):

Yes, absolutely. It is a great quarantine toy to have at my disposal. Well, Matt, thanks for kind of reflecting on that. That's a really interesting way to begin your career, and we're here actually to talk about how you can help organizations ask better, big questions through data and analytics. But before we begin thinking about the future, I want to understand a little bit more about your career and how you got involved and passionate about this area. Obviously, you mentioned that first job, but how did that get parlayed into senior leadership roles in IT and data and analytics?

Matt Griffiths (06:06):

Yeah, thanks. Thanks for asking. My degree, my interest was always in engineering. I've always considered myself very fortunate that I knew kind of what I wanted to do from a fairly young age, and that was being a car mechanic. And, when I first had my taste of being a car mechanic, which was a two-week stint out of school as part of my high school education, I realized that actually, I didn't want to be a car mechanic. I wanted to be an auto electrician because those guys were using computers and dressed up in suits rather than overalls to do their work. And so, I went and studied auto electrician, auto electrics, which was also orientated towards manufacturing as part of a British government initiative back in the late nineties to really encourage companies to kind of reinvest in manufacturing education. 

And that's really where I got my passion for working in the manufacturing industry. Went and studied manufacturing engineering at university. I spent some time with Xerox in the UK and also in Rochester in New York. And it was at my time at Xerox that I got involved in this company that made automated guided vehicles. They were based in Germany. It was a four person company: the owner, his wife, his daughter, and me. And I was all things technology from managing the network and the computers and the servers and the storage and the email and the website and writing software for inventory control and writing the PLC code that ran in the vehicle -- was probably the single most fulfilling an educational role I've ever had in my career. And that's really where I got into technology and self-taught myself software development. 

Not long after that, I moved back to work for Dell computers back in the UK. That role was in Germany, spent 16 years with Dell, progressing through a number of different roles in software development and then software development leadership very much focused around manufacturing and supply chain, developing real-time control systems, supply chain management systems, inventory management systems with Dell, before leaving Dell and joining Biogen is initially leading their manufacturing technology team and then becoming the CIO. And then moving to Stanley Black and Decker as CIO for their industrial businesses. So, really kind of stems a lot from my passion for manufacturing companies and manufacturing processes. And how do we apply technology to really digitalize those types of companies.

Liz Ramey (08:49):

What was the point in your career path that you kind of moved from, there's a technology focus on how to lead the organization to a data and analytics focus?

Matt Griffiths (09:04):

I think there's always been a recognition that everything we did was to capture data, leverage data, translate that data from being kind of just data into information and into insights. I think the last few years with the advent of large scale data sources with IOT, with mobile, with sensors, with external data feeds, and kind of enrichment processes, the recognition for the power of data as an asset in the company has kind of really exploded. And that's, you've seen the advent of the chief data officer role, the chief digital officer role, which is really about exploiting not only the data with insight and organization, but the data that kind of is part of the broader ecosystem that can give you a 360 view of your customer or your operations or your employees, or your retailers. And that for me was an area I got into in more of a, kind of a core focus about 18 months ago, two years ago now. It wasn't my background. 

My background was around how to apply technology to drive business value, not a data scientist. The first thing I did was go and hire some of the brightest, smartest data scientists from the insurance, the healthcare, the various, the finance industries in, in and around Boston and Hartford area. And together, we have delivered a tremendous amount of value over the last 18 months, really about giving our business groups new capabilities based on advanced analytics that have really transformed the way we operate.

Drew Lazzara (10:54):

Well, Matt, that's actually a pretty good transition to what I wanted to talk with you about next. So, you're thinking about giving your business peers more power to make better decisions and optimize what you're doing through technology and data. So, how does that process start now? Before we think about the future, I'd love to understand a little bit about how your business defines its priorities. How is that kind of decision-making structured now, and what is the role that data and analytics play in that process as things stand today at Stanley Black and Decker?

Matt Griffiths (11:24):

Yeah. I think, when you think about analytics and data in an organization like Stanley, the four stages or four components of analytics is very applicable. So descriptive analytics, diagnostic analytics, predictive, and prescriptive analytics. And 99 times out of a hundred, an engagement with me or my team starts with a -- I need a dashboard for this, or I need a report for that, or I need a spend cube for last month, last quarter, last year. And the first step that my team does is say, well, why? What are you going to do with that data? What outcome are you looking to drive? What insights are you looking to understand, and what are you going to do with those insights? What truly is the outcome that you're looking to influence? And we really flip those conversations on their heads and take a very outcome-driven approach. 

And the reason for that is because unlike traditional and other forms of data analytics, what we have tried to do is build capabilities. We take data and patterns that we've identified in the past, and we use those to build models, and we embed those models in applications and in tools so that we are truly creating a business process, a business capability that automates a decision or gives a better decision than a human could have made themselves. And so, one of the hardest aspects of our role is getting the business, our business partners to think in that different way. And there's some good examples where, you know, one of my favorite examples is around customer attrition. We had a request to help understand why customers in a certain part of our business were attriting over periods of time. And when you kind of do the 'five whys,' what you really want, the outcome you really want to achieve is how do you stop customer attrition in the future? And the fact that we can use the data in the past to predict the likelihood of a customer leaving us in the future was something that our business didn't even understand could be done. And that's the predictive and prescriptive components of what we do. And that's the transformational piece that we are really trying to drive our business to kind of think differently around.

Liz Ramey (13:54):

So much of what you just said, Matt, really kind of sits under a scope of culture to me and really getting a mindset shift within the organization to value the outcome of what data could provide, right? So, I'm just curious to know what pieces or points of the culture did you really have to get on board in order to think like that? Just to further a little bit, you know, I remember, I think we all remember the nineties idea of the IT guy always saying no, right? And, now it's moved to this, the data and analytics, where you have to have these pretty difficult conversations with business leaders to ask them, what is the outcome that you're truly looking for, right? And that takes some cultural acceptance of data, right?

Matt Griffiths (14:53):

Very, very much. And, it's funny that you asked that question. I have a slide in presentations that I use around this topic within our organization. And the slide includes a quote from a man named William Edwards Deming. And I'm a manufacturing guy, I work in a manufacturing company. Edwards Deming was recognized as the forefather of the lean kind of practice in manufacturing. And one of his quotes is, 'if you do not know how to ask the right question, you discover nothing.' That's the culture component that we're trying to drive, it's getting the business to ask the right question. But I think every company is different, but in my experience, the best way to get the business to ask the right question is through the use of examples. I gave an example a moment ago around customer attrition, but we have 30, 40 similar use cases that we have created over the last 18 months. You know, everything from predicting the performance of a given promotion of a new tool at Home Depot to predicting the lead time of how long it's going to take to ship something from Asia to North America to predicting which of our facilities are most at risk from a COVID-19 impact, and which ports are being used to import our product into the country that are also at risk from a COVID-19 impact. So, you know, and everything in between. 

And, when you start to use these examples of the things that can be done with data, and people -- you can see the light bulb go on. It's things like -- you can predict the likely cost of ground transportation in six months’ time based on the number of truck driver jobs being advertised today on monster.com, right? Or the price of copper based on a whole bunch of hedge funds indexes that you know are looking and tracking research in those particular spaces. And, when you start to use these examples, people understand, 'Oh, well, if you can do that, can you do this?' And, that's when you start to see people ask the right questions. I would say, it's taken us, in certain groups upwards of six months to get people to start thinking differently and start thinking in the -- what will happen and what can I, or what should I, do about it, that kind of foresight view versus the old way of -- what has happened and why did it happen, which is the hindsight view. And, it's definitely been a fascinating cultural transformation to observe and watch Stanley go through that.

Drew Lazzara (17:58):

So, Matt, in listening to you go through that, I'm curious about the mechanics of how you spread that culture because it sounds like if you go example by example, team by team, it's kind of a guerilla culture, an enculturation movement, where every time you get an opportunity in front of a part of the business, you can make a little bit of an impact a day at a time. But for companies that are sort of behind where you're at in this culture and data culture curve, what's a good starting point for making a larger impact on the way the entire business looks at data. Is there something that you can do on a macro level to help spread that faster than just the guerrilla style you described?

Matt Griffiths (18:35):

It's a great question. And we definitely do... And I'll talk about other mechanisms that we have other than that guerrilla style, as you mentioned. I would say that that one-on-one group type approach is more effective than it may appear. And the reason for that is because culture change typically happens from the top down. So if you educate and inspire the senior leadership, then the rest of the org, they start asking the questions of their teams. And those teams therefore have to start say, 'Well, okay, that's a great question. Let me go and think about how I can answer that question or how I can rephrase the questions that I'm going to ask in the future.' So that kind of top-down filtration of culture is really important. But, I was once told, and my experience is kind of consistent with this, is it takes six months for every level of an organization to achieve culture change. So, if you've got six levels of your organization from the most senior leadership to the sixth layer, then you're talking three years before you really permeate that cultural trends throughout the organization. 

And to your point, three years is a really long time, especially in this age. So, we absolutely have other delivery mechanisms for helping build that culture change. The first one is the use of the Workplace tool by Facebook, which is our internal social media tool. It really has replaced our traditional intranet. For those of you who aren't familiar with it, it really is Facebook for internal organizations. It's been massively adopted and embraced here at Stanley Black and Decker. And, we have a number of groups focused around digital and data and IOT and industry 4.0 and other kind of key initiatives in the organization that we use to share examples, share white papers, share training material, talk about success stories, talk about failure stories. And that group has grown from about 200 people back when I first joined the group in the end of 2018 to over 1200 people now. And the reason that that's so important is because my view around data and the future of analytics in companies like Stanley is we're not going to have this center of excellence model forever. There's going to be an uprising of citizen data scientists. We're already seeing that. Data should be absolutely a shared asset across everyone in the organization. The whole kind of democratization of data is something we're pushing very hard. We want to get the right tools, the right education, and the right understanding into the hands of the people that are really using it on a daily basis and building training programs and education programs and communication programs is a core part of what my team are responsible for.

Liz Ramey (21:49):

That's fascinating, Matt, 'cause I had written down a note here, just curious, at what point as a leader, do you know when you can kind of switch that light on, right? Or I guess break the levy, right? Because everybody wants data. All of the business leaders out there for years now have been saying, just give me access to the data, and I'll figure out how I want to use it or what question I want to ask. How are you as a leader in that digital transformation place, how do you decide, when is it okay for the levy to break and for me to allow this kind of democratization of data or even analytics capabilities?

Matt Griffiths (22:38):

You know, I have that conversation a lot because I think there's just a lack of recognition as to just how much data truly exists and truly can be used. Everything from marketing data, sensor data, mobile data, operational data, ERP system data, Salesforce data, and there's still a mindset, which is outdated now, that it is possible for a human to pull all of those data sets together and draw any meaningful insights from it. And, is it possible? Absolutely. Okay. Is it still the case? Yes, it still happens. But what I, the conversation I usually end up having is, there is so much data out there now, one approach we could absolutely do is go off and spend 12 months identifying and acquiring and ingesting and cleansing and enriching all of that data into a data lake. But the risk is that you end up with a data swamp, which is pretty meaningless and very hard to find any kind of value in. And a much better approach is to start with the outcome, right. Rather than start with -- I just need access to all the data and Tableau or ClickView or Power BI tool to get access to that data. The conversation, we try and change the conversation to be -- what outcome are trying to achieve. What's your hypothesis, what's the value. What's the outcome that you really want to influence. And then let us work back to identify what are truly the datasets, both internal and external, that you need to be able to achieve the outcome. And what is the algorithms and the models and the machine learning and the patterns and all of the other technology aspects that have to be created to enable that outcome. And that, even though we still absolutely get a lot of conversations today around just give me the data. I would say the vast majority of the time, the solution ends up being a much more focused, outcome-driven solution that really drives quite often millions or tens of millions of dollars of value.

Drew Lazzara (25:00):

Matt, we've been talking for maybe 12 minutes, and you already thoroughly convinced me just in speaking to me for that amount of time to ask questions that are outcome-based. So, given how intuitive you make it sound and how kind of obvious you make it sound to frame things that way, why do you think that business tends to ask the wrong kinds of questions in the first place? Why is the instinct to not focus on the outcomes when they're speaking about data, whereas they might do that in some other aspect of the business?

Matt Griffiths (25:30):

Hm, that's a fascinating question. Honestly, I think it's just the transition that we're all going through between, I mentioned earlier, the descriptive, diagnostic, predictive, prescriptive. Most people in the business, and matter of fact, most people in technology even, don't recognize the capabilities of predictive and prescriptive analytics. They don't recognize that a machine learning algorithm is far more capable of identifying patterns in data than a human could be. They don't recognize that in many cases when we ask or request a dashboard or a report, what we're really doing is looking for data to subsequently make a decision. And, actually you could automate that decision in real time and really kind of create this completely closed loop, this closed loop system. 

A couple of years ago, Stanley spent quite a lot of time thinking about competing on the rate of learning, which is I think a McKinsey philosophy, although it's been around for a while. But the theory is that the faster that you can take data and learnings from your operations, your customers, your employees, and turn those insights and closed loop control those insights back into your operations, the faster you can compete, the faster you can win. And there's a recognition that companies of the past, maybe like a Walmart or a Dell or a Ford, who they were competing primarily on volume. And that's how they were able to really kind of grow their businesses. And companies like Apple and Tesla, and they're competing on innovation. Both of which are important, but there's now this third dimension, which is competing on the rate of learning, which is taking data from your operations, from across your business, and driving that data back into the system. Amazon, Netflix, Kimberly Clark, even Coca Cola. So, if you've been to one of those restaurants that has those vending machines, that you can literally mix and match any particular Coca-Cola product that your kids typically make a beeline for and start adding all these weird and wonderful flavors into their cup. Cherry Sprite came from monitoring the usage of what people were mixing in those machines, and Cherry Sprite was created because of the insight that came from those vending machines.

Drew Lazzara (28:17):

So, data's not all good is what you're saying, basically.

Matt Griffiths (28:23):

So, and it's that philosophy where you say, okay, what is the outcome I'm trying to drive? I'm trying to drive new products, or I'm looking for white space opportunities, or I want to understand the effectiveness of my operation. How do I take the data, acquire the data, clean the data, and most importantly, drive insights from that data, and then deploy those insights back into operations in real time. There is no room for a human in that closed-loop process. And that's the shift that we are really trying to get people to kind of embrace. And I think they are. I just think it's going to take time before you see that mainstream across every company and every leader.

Liz Ramey (29:14):

Since we're on this topic really around the kind of, data economy or analytics economy, I would say. I'm just curious if... 'cause you have been talking about even with Coca-Cola, right, these ways in which they can create products based on the data that they're capturing through even consumer behavior and whatnot. But, I also like to think about it as what developing internally to help you make those decisions. You can also then package and essentially sell to someone else who is trying to save that, or overcome that same challenge, right. So, have you had success in that sense? Are you taking some of these models that you've built and turning around and getting them out into the market? So, then the business sees your organization as a revenue driver?

Matt Griffiths (30:08):

Yes and no. We've not marketed or sold a specific algorithm or model or mathematical solution to a given problem. We have, however, partnered with our various product teams to understand how we can apply those same concepts in our products, right? So, for example, Stanley is a very broad, has a very broad range of products, but most people don't actually really understand just how many kind of different industries we're in. And one of our major business segments is actually in security, specifically, electronic doors and video and audio surveillance and monitoring systems. And those doors, if you walk into Home Depot, for example, you'll see Stanley Access Technologies or Stanley Doors on those automated doors that you walk through. You see them in hospitals a lot, as well. Those need monitoring, they need maintenance. 

Today, typically what happens is something may go wrong with those doors. That means the door owner has to make a service call to one of our field service technicians, and then we dispatch someone to come out and fix that. And in many cases, it could have been something that either could have been prevented from ever happening in the first place through regular maintenance. It could have been something that could have been easily fixed by someone onsite if they just had the technology and know-how or have the know-how skills. And rarely is it, actually not rarely, but it is maybe 50% of the time, it is something that is something that has to be fixed by a service technician. We don't know that 'cause we don't have the data about the failure in the first place. 

So, now we're looking at how do we embed sensors in those doors to provide a bit more insight into what's gone wrong so that we can quickly address and resolve that issue faster than we can today. We have other products that are related to steel recycling, very large attachments that you see physically on the end of cranes, crushing steel girders, and tearing down buildings. You've probably seen them. We lease those out or we rent those out. And today, they are typically rented on per hour or per day or per week basis. Some customers may use that tool a hundred times an hour. Other customers may only use that tool 10 times an hour, but typically, they're paying the same rate. Now, we've embedded through IOT sensors and RFID and RTLS tags, the ability to capture data around the usage of those products so that we can actually kind of pay on a more of a consumption basis. And now you pay every time you use it, not how long you have the tool for, and that's, so we're starting to really kind of open the eyes of our business partners to different ways of going to market, different ways of revenue generation, maybe a more consumption-based or as a service type model. But, again, it's something that's going to take some time before we get that widespread.

Drew Lazzara (33:25):

Matt, there's a lot of exciting potential that you outlined there. And it's clear that it starts with building that culture of asking the right questions and focusing on outcomes. But, I want to pivot a little bit to what happens next for organizations because one of the elements of analytics that you've touched on a few times is this idea of predictive. And I would imagine that asking the right questions is just stage one. And the second stage would be trusting that the data is telling you what the future is going to be like. So, how do you get the organization to go from asking the wrong questions, to asking the right ones, seeing some results and having that permeate through the culture, and then go to, we need to go to the data first so we decided what our strategy is for the future? How do you get to that truly predictive place, where everyone is looking to the data at the same time and understands its potency?

Matt Griffiths (34:15):

It really is about demonstrating the effectiveness of the model, firstly, by taking historic data sets. So, the normal process for developing a predictive model is take a historic data set. So, let's say we take the last three years of demand forecast from Home Depot, as an example. We use the first two years to train our algorithms on fluctuations, seasonal variation, the impact of building permit restrictions, weather conditions, whatever, whatever different variables that we want to include inside that model. We used the first two years of data to train the model. And then we use the third year of historic data as a way of validating that what the model is outputting is actually what happened. So, that's kind of step one in gaining trust in the model. Step two is then to kind of go into a parallel path, which we have some examples right now of in demand forecasting, in freight management, where our model is running alongside the old business process. And so now we can start to see true, future state kind of performance, but we are able to compare it to a) what actually happened, and b) what the old mechanisms may have kind of indicated as a potential result. 

So, let's say demand forecasting, we're a JDA user for our demand scheduling. we've built algorithms that take into account a lot more variables than what JDA is capable of taking into account. And we're seeing a 30% improvement in demand forecast accuracy over what JDA was providing. But we're running those two in parallel. We're still running the JDA demand forecasting model, which is as an input into manufacturing, and we're running the algorithms and the models that we've built, and we can see the differences. So, you start to build up trust. And then the third phase is much more around adoption, and we've... I've quite often, again, another slide that I have in the presentations that I typically pitch is that only 30% of the success of a program is down to either the technology or the algorithm. And that 70% is really down to business adoption. So, I have folks dedicated to driving business adoption, to measuring success, to identifying any barriers to why a particular new way of working is not being embraced. And that means that we need the right stakeholders in place, the right sponsorship in place. There's a whole change management aspect around these types of changes. And it's definitely some of the most challenging aspect of our role. 

And, in fact, we've had programs that have failed because of business adoption. We had one in France last year where we were doing potential new customer identification, where we had developed a machine learning algorithm that profiled what a typical Stanley Security customer was in France, in this particular case in Paris. And we had purchased some lead information from a third-party vendor. And then we had kind of overlaid the machine learning model template with this data. And we identified all of these potential customers that were likely candidates for Stanley Security. And what was interesting in that data was that it turned out that hair salons were twice as likely to become a Stanley Security customer than any other type of business. And, which was an interesting insight. Problem was our sales compensation model was much more orientated towards medium and large size companies than it was around smaller companies. And so, the algorithm or the results of the algorithm just really weren't used, even though there was a massive amount of potential there because the sales folks weren't compensated as much for that type of business as they were for larger businesses. The algorithm wasn't really used as well. So, that's an example of how you really have to kind of marry the technology and the business process together to drive the right level of adoption.

Drew Lazzara (38:58):

Matt, I had a conversation just this week between a CFO and a CIO, and both of them were talking about this increased appetite for data in the organization. They were talking about a lot of kind of the customization of data that we've touched on in our conversation. But both of them thought that the other person's unit should be the owner of data in the organization. So when you talked about the center of excellence model can't last forever, if you're trying to be a truly data-driven organization. So, what are your opinions on where data ownership should live within that governance structure to make sure that you're helping all of these different areas in the business get the precise kind of data that's customized to them? Do you have any thoughts on maybe a prescription for the right kind of governance in your point of view?

Matt Griffiths (39:42):

Well, I'm not sure I have a prescription. I think that's one of the evolving hot topics. You definitely see different CFOs and CIOs and chief data officers from different industries, I think, typically have different views on this. I think in industries where data has always been a core asset, like finance, insurance, healthcare, to an extent, I think there is a view that there should be a lot of centralization. There should be a lot of center of excellence type models. Governance is owned by a single group. All the data is owned by a single group. That's not my perspective. I think tools like DataRobot and H2O and Alteryx, and very powerful machine learning and advanced analytic and computer science type, data science type tools are going to be as prevalent in the business in the future as Excel is today. And that's definitely my vision, and I'm trying to get those powerful tools into the hands of people that can really use them. 

And, what's interesting is -- the software vendors in these spaces, like an Alteryx is a good example and DataRobot is a good example. They are increasingly making those tools easy to use by citizen data scientists or end consumers or people that have never written a line of Python or R in their lives, or don't know what a multivariate analysis is. And, so I think the technology itself is also moving more and more towards the end user, the consumer of the data. And so, if you think that that all of this data and analytics and predictive and prescriptive analytics is all going to be democratized across the entire company, then the data itself has to be very accessible, does have to have a strong security and governance model, but maybe it's owned by the business process owners, and that there's a much more federated approach to data governance. And that's definitely the approach we're taking. We do have a centralized group who are responsible for defining what the governance model is, but the actual ownership of the data is very federated across our business groups. And I think that's the model, my personal view is that that's the model that gets data into the hands of the people that could really do the most with it.

Drew Lazzara (42:31):

Matt, before we let you go, we always like to ask two questions about the future, as well. One is a big question from a previous guest. In this case, it comes from Jacqueline Welch, who is the chief human resources and diversity officer at Freddie Mac. And her big question for you was -- what does the future of work look like? And I would be curious to understand how a data-driven person like yourself looks at the optimal future for the way people work?

Matt Griffiths (42:57):

It's definitely something we are spending a lot of time thinking about as many other companies are, as well. I think COVID has been a great example of the power of data. I mean, all of a sudden, the world is much more aware of statistical analysis and the importance of being able to use data to make decisions. It's in everybody's life now. It's not just in the lives of people working in a place like Stanley, even my children are aware of the percentage of positive test cases in the local area in the last two weeks. So, even my children are using predictive analytics now to make the decisions around should they go out with their friends or not, right. And, so I think the world is going to be much more data-driven. I think the computing power, the algorithms, the advent of data science, the amount of talent that's now available to help make good decisions, the tool sets that are becoming more and more user-friendly to help drive insights from data are all going to mean that organizations are a lot more kind of using the data as part of their daily decision-making process.

Liz Ramey (44:16):

So, if you were to able to ask an executive in any role a question, what would your next big question be for them?

Matt Griffiths (44:25):

I think my next question would be, or my big question would be, how leaders who maybe don't have the practical understanding and insight around the power of data are going to really drive their organizations to think differently. I mentioned earlier that this is very much a top-down culture. And, I think if I'm honest, when I took this role 18 months ago, even I was a little bit skeptical as to how big is this big data and machine learning and kind of initiatives? Is it really everything it's cracked up to be? And I definitely have formed my opinion. But I guess I would challenge leaders of organizations to really ask themselves and their organizations the question, how are they going to become a data-driven organization?

Liz Ramey (45:25):

That's great. Well, Matt, we really appreciate all of your insights. We've learned so much from you today, and it's exciting for the future to be able to hear from leaders like yourself think about the future and think about how they're going to impact that. So, we really appreciate you being here today. Thank you very much.

Matt Griffiths (45:45):

Thank you for having me. It's been great to talk with you.

Liz Ramey (45:50):

Thank you, again, for listening to The Next Big Question. If you enjoyed this episode, please subscribe to the show on Apple podcasts, Spotify, Stitcher, or wherever you listen. Rate and review the show so that we can continue to grow and improve. You can also visit Evanta.com to learn more about our C-level communities. Network, share, learn with Evanta, a Gartner company.