In this episode, we are joined by Alex Cunningham, a manager with Church & Dwight’s Advanced Analytics.
Alex and Evan have a thought-provoking discussion about how to drive analytics successes, build momentum, and develop partners. He also discussed their strategy to develop a good product and the team’s secret to integrating that process. Alex also touches on what makes the analytics team at Church & Dwight special and the most important traits they look for when growing their team.
Learn more about why we created the Mining Your Own Business podcast.
Alex Cunningham | GuestAlex Cunningham is the Manager of Advanced Analytics within Church & Dwight’s growing Analytics Center of Excellence. He leads a team of Data Scientists who partner with Executive Leaders to enable data-driven decisions throughout the company. He leverages his past experiences in Finance and IT to better serve as an analytical-translator – speaking both “tech” and “business”.
Alex further enhanced his quantitative skillset by completing his MS in Data Science at Northwestern University. He loves what he does and his passion for all things analytics is contagious!
Follow Alex on LinkedIn
Evan Wimpey | HostEvan Wimpey is the Director of Analytics Strategy at Elder Research where he works with organizations to transform deficient data into tangible business value that advances their mission.
He is uniquely suited for this challenge by pairing his professional experience in management and economics at high-functioning organizations like the Marine Corps and Goldman Sachs with his technical prowess in data science. His analytics skillset was strengthened while earning his MS in Analytics from the Institute for Advanced Analytics at NC State University.
Evan almost always has a smile on his face, which is at it’s widest when he is helping organizations use data in innovative ways to solve complex problems. He is also, in a strictly technical sense, a “professional” comedian.
Follow Evan on LinkedIn
01:43 Alex’s background and career history
04:50 How Church & Dwight structures their analytics capability
06:06 How the analytics group has integrated in the business & built momentum
11:33 How they prioritize & implement projects
19:05 The important aspects of their team
23:43 If there was a magic “analytic success” button, where would Alex use it?
Evan Wimpey: Hello and welcome to the Mining Your Own Business podcast. This is episode two. We have survived past episode one. Can you believe it? I’m excited today to introduce Alex Cunningham, who is the manager for Advanced Analytics at Church and Dwight. Alex has a pretty long tenure there at Church and Dwight, which we’ll get into here in just a moment. He has a bachelor’s in economics from the College of New Jersey and an M.S. in Data Science from Northwestern University. Alex, welcome to the show. Thanks for coming.
Alex Cunningham: Oh, it’s an honor and a privilege to be here with you, Evan. This is making my day today. I’m so excited.
Evan Wimpey: Fantastic. You are selling it. Well, sir. Why don’t you get started? Why don’t you give us a little bit about your background? You’ve spent quite a bit of time at Church and Dwight some of that in analytics. Some of that is outside of the space of analytics. So maybe just talk about sort of your career trajectory there. And actually, maybe you do a quick introduction of Church and Dwight as well. Some folks probably don’t realize that they know Church and Dwight, but there are a lot of familiar brands there. So maybe just talk a little bit about who Church and Dwight are.
Alex Cunningham: Oh. Well, Church and Dwight is a fantastic company that many people may not have heard of but have probably heard of our brands. We’re the parent company of Arm & Hammer. We own Trojan, First Response, Batiste Dry Shampoo, Arm & Hammer Cat Litter, of course, our laundry detergent as well. We make about 54 different consumer brands, most of which are sold here in the United States. But our international footprint is growing and expanding very rapidly, and we’re excited about what that means for the future of this company and what data science and analytics can do for us to fulfill those goals.
Evan Wimpey: Awesome. Yeah. Big company, big footprint, for sure. Can you talk a little bit about your history there? You haven’t always been in data science and analytics.
Alex Cunningham: Absolutely. So you’re fortunate to work for an organization that believes in investing in those that work for them and providing opportunities to grow cross-functionally. And so I’ve benefited from coming in here as a co-op while I was still a student, which turned into my first career here at Church and Dwight working in the back end of SAP systems with the IT group, I matured from that into some, some business analytics is what we were calling it at the time. I then had an opportunity to jump over to finance, which was fantastic. I spent a couple of years over on the business side, really learning how we operate and how we conduct our analyzes and the different processes that drive the business. Having kind of gone more of the data and analytics side, making systems and processes work. And so when an opportunity opened up and Church and Dwight decided to create an analytic center of excellence team, it was a fantastic opportunity to combine my experiences in the I.T. side of the world with some of the great business insights and context that I had learned with our finance group. And so I could not have been any more fortunate to be in the right place at the right time and have been on the team since about twenty seventeen now. And we continue to grow and mature because analytics and data sciences is certainly a journey and one that we are in our early stages of, but are rapidly making great progress.
Evan Wimpey: Yeah, that’s fantastic. And I wouldn’t say it’s rare necessarily, but I think it’s overlooked how valuable it can be to come from the business and understand how the business functions into this analytics role, instead of just the analytics background and never having worked outside of analytics within the company. So I think that’s a pretty valuable vantage point to have. You mentioned the analytics center of excellence. Can you talk about sort of where that sits within Church and Dwight’s or how that structure looks?
Alex Cunningham: Absolutely. So at Church and Dwight, our Analytics Center of Excellence team serves as an internal consulting group or partner for our various business units. And what we want to do is for a very long time. We have run our business very successfully, but with a very much focus on the qualitative. And now what we’re going to be able to do with this analytics center of Excellence Group is to complement that qualitative understanding of our business with a quantitative one. And if we can put those things together, the types of business decisions and insight we will be able to generate will start to move us forward into becoming that analytical competitor right there that everybody’s chasing.
Evan Wimpey: Yeah, that’s great. That’s the goal for sure. Maybe, we try to understand a little bit how that works. You’re an internal consulting group. I work for a consulting team as well. We have to go out to people and ask, Hey, do you need this? Do you need the service? I’m curious how much within your analytics center of excellence is? Are you pulling from like the operations team or the marketing team? Or are they coming to you with the need like, Hey, we’ve got this data, we have this problem we need? We need some help with this.
Alex Cunningham: That’s a great question. And you know, it’s evolved here as our group was first becoming established and setting ourselves up within the organization. It was a lot more pulling folks in you having to go out and ask them for, you know, to partner with us and bring them into our circle. And now what’s great is as we’ve established ourselves, we don’t have to do as much of that pulling in. They are coming in themselves because they’re hearing of the successes and the wins that the organization is gaining, and they want to be a part of that too. And that’s so exciting. One of the phrases we’ve been throwing around is a P.T. Barnum expression, which is nothing, draws a crowd quite like a crowd. And in data science and analytics, we want to try and use that to our advantage. Let’s get excited about this stuff. It’s cool and it’s really useful. We don’t want to overhype it either, but you know, this is something they get pretty excited about. This is some cutting-edge, you know, forward-thinking type of stuff. And how cool is that?
Evan Wimpey: Yeah, that’s a great quote. I love the P.T. Barnum quote here. You’ve got a bit of a crowd. You’re not doing as much pulling. How did you see your relatively new team? I think you said twenty seventeen, you were established. So how did you get some of those early wins? Maybe you can talk. About I don’t know if you want to get into specifics on a project or maybe just sort of the mindset on how do you choose some of these early projects to try to get that initial crowd to try to build some momentum there?
Alex Cunningham: Absolutely. That’s a great question. One of the things that we have spent a lot of time thinking about is how to prioritize our analytics projects and efforts. And you know, it’s interesting that although how we source them right, that that push for pool has changed a bit, how we go about prioritizing them has not. And I think that’s been a consistent theme. So the way we think about it is kind of a triage system. The first aspect is, are there any major fires going on right as a new group, right? We were very fortunate that they were not right, so we could skip that aspect of the triage. But I do mention that because, you know, events like COVID, right really came in and affected not only Church and Dwight, but, you know, the world. Right. And so when there are those kinds of major events like that, you know, obviously that that real sort of. Adjusts our priority list, so to say, however, barring any sort of major global phenomenon, we start to prioritize our work based on the outcomes that day that those experiments are going to provide. And that’s what we call the work that we do. We call them experiments. And the reason we phrase it that way is that sometimes they don’t work. These are new areas that we’re investigating, where we’re lifting these little rocks and trying to see what sorts of nuggets or insights we might find. And, you know, unfortunately, right, sometimes we turn over a rock and there’s not anything there, and that’s OK. So long as we are, you know, sort of failing fast so to say, and not trying to do too much. And that’s another aspect of how we do that prioritization. So when we look at those outcomes, we also consider the complexity in deriving that outcome because if we find too much complexity, what we try and do is borrow from agile methods and break that complexity down into some smaller components that can get worked on a little bit more succinctly because one thing we hear from executives a whole lot is we do not want to boil the ocean, right? And in data science, that can be easy to do, especially in the big data type, environment, or world. And in fact, some data science teams want to boil the ocean. And so we got to kind of make sure that, you know, to prevent those types of things, we’re prioritizing the work that has the most meaningful outcomes and those outcomes can be dollars. But those outcomes might also be things like efficiency or future learnings that set us up down the line to do bigger and better things. And so that’s a little bit about how we think about prioritization when it comes to analytics.
Evan Wimpey: Okay. Yeah, it’s great. And I love the terminology of the experiment. It’s data science. You need to conduct these experiments. That’s a good mindset, especially if you can. If you can get some synergy, you get some agreement. This is an experiment and there’s no certainty that we will achieve this outcome. We’re exploring here. So let me ask you if when you lift the rock and something is interesting, you can improve a process or save some money or have a better forecast there. Can you talk about the steps that your teams take to try to get that implemented when once you’ve got, you’ve experimented and you have some good data science tool or model or product? How do you get that into production?
Alex Cunningham: Oh, that’s a great question. And, you know, that is the hardest part of data science. I believe your previous guest, Gerhard had touched on this just a little bit in that the modeling is the big secret here. You know you don’t want to say the quiet part out loud. Too often, modeling is the most straightforward piece of all of this. The change management and the implementation of that solution and embedding that into the process embedding that into systems is where a lot more of the complexity and time and attention is spent. And so just to talk a little bit about that, you know, we’ve started to try and break that any process we do from an experiment standpoint down into a couple of critical components. The first thing that we do is always ensure that we have the data we need at the onset sometimes, right? It’s really easy to just jump right into and get excited about some problem statement and then you get into your data understanding following that good old Crisp D.M. framework and you realize that, oh, I don’t even have the data there to answer this question. And so, you know, having learned some of those lessons early on in our experience, we wanted to make sure that we always start with a data assessment and having that data first. And one of the ways that we do that is that we ensure that not only do we have a data set, but we ensure that that dataset is accessible in a self-service fashion via a dashboard or other methodology to the business units and analysts themselves. And why is that right? Because we want to be a prescriptive group. We don’t want to necessarily be looking backward. So it sounds a little counterintuitive. But what it does is by giving analysts that data upfront, they’re able to ask deeper questions. If I can quickly, as an analyst, answer who, what, where, and when looking backward did something happen, I started to ask myself why. And those are why questions are what lead to great data science experiments or models. And so we start with the data access. We grant that access via dashboard or visualization or some other self-service means that sets us up now to start doing those data science experiments. And once we create some sort of data science asset, write a model or some sort of analytical framework for answering a question, what we then want to do is take a bit of a pause and treat that solution almost as a semi-finished good right coming from a manufacturing site. That’s right. That’s good data is our raw materials, and the model is a semi-finished well. Well, the thing about a semi-finished good right is it is the opportune word there. We need to take that to completion and to do that, to make our finished product what we’ve started to, to call that phase is the operationalization of our models. And that operationalization is really where we define the processes in the workflows, you know, in a lean sort of sense, developing standard work and training the analyst community on how to conduct that standard work or what these new processes would be. Figure out the on-ramps to integrating this solution with existing systems and architecture, and we treat that really as a whole, wholly separate effort from this experimental piece that we conduct upfront. And it’s really because it does take so much time and attention and is such a different animal that, you know, we treat it as such. And so, you know, that’s just a little bit on how we start to think about implementing the things that we do so that they don’t turn up and be self-aware, right? Nothing is sadder than when you create a great solution and it just gets put up on a shelf up here. And it’s interesting and it might be insightful. And it never makes its way back off that shelf. And we want to try and make products that are, you know, data products, right, that don’t end up on shelves but end up in users’ hands.
Evan Wimpey: Yeah, that’s great. That’s for academics. Let them do the interesting stuff, just to be sure. I think it’s a really useful tool to try to give that analyst that is sitting within the business, the dashboard to try to give them the descriptive information and help let them sort of ask the why questions, let them develop the business question that you need to answer. And presumably, that helps on the back end. So once you’ve got the semi-finished goods and you’re ready production-wise and it’s going to change. The way that analysts do their job or that business makes a decision. Well, they were helping on the front end too, so they’re more prepared to implement the change on the back end.
Alex Cunningham: Absolutely. Absolutely. You know, and the other thing with that that I think is an important aspect is it makes our business partners true partners, right? Nobody wants to, especially in this space. Data science and analytics can be scary. Folks can be concerned that they’re going to be replaced by some algorithm that they’re not right. That is not a good way to do business or think about this. You know, folks might be a little bit intimidated about their skillset or, you know, lack of knowledge, perhaps or experience in math or statistics. But that’s what becomes important in having this group be that real partner to say, Oh, you don’t have to worry about those aspects. We’ll take care of those pieces. You provide the context, provide that subject matter. Expertise provides how you’re going to use this to make an actual decision and will worry about some of the mathematical intricacies or other aspects. And so to let people know that you’re here to provide them insight and not provide them a, you know, the direct answer or, you know, some crystal ball that’s going to do everything for them is a really important context in dealing with these groups.
Evan Wimpey: Yeah, that’s great. That’s a really important context for sure. So let me ask you then about your team specifically so you’ve got your relatively new analytics center of excellence. You’ve got a handful. I’m not sure if you’re OK to mention how big the team is right now that sits in the analytics center of excellence that needs to build this rapport and help deliver the dashboards and then do the experimentation and push this semi-finished good out and focus on both the technical and this change management aspect. That’s a lot to focus on when you’re hiring for your team or you’re looking for talent for your team internally or externally. What are some of the things that you try to key in on four for that team?
Alex Cunningham: Absolutely. It sounds a little counterintuitive, but the most important aspect that we hire for is humility because as we just mentioned, you know and a lot of people are shocked when they hear that right because you expect our R Python or some tool or technology, right? And those are important as well, right? Don’t get me wrong, however. Empathy and humility. Right. Two closely related traits are what allow us to become those good partners to folks in the business in that internal consulting type capacity. We are humble enough to admit that we’re not the experts here. We might be pretty good at some of the math and some of the modeling, but when it comes to making business decisions, that is what these folks do for a living. And so we want to be respectful of that. And being humble in that approach is what allows us to be that good partner. And you know, the close cousin of humility is empathy. And that empathy allows us to understand when folks are having hesitation because they’re uncomfortable with the context or some aspect of the experiment. It also allows us to do that kind of really good planning for operationalizing our data asset once we’ve we’ve created it because by understanding where they’re coming from, how they want to use it, the ways that they are thinking about these decisions and putting ourselves in their shoes, that allows us to better design those sorts of end pieces or end states if you will. And so those to me are the critical components of this team. Everything else is going to come second to that. Again, important some important aspects, right? We have to also be fairly good with and mindful of our math and is applied math, you know, using computers at scale. But you know, those human characteristics in those human traits are really what is most important to us here and what is going to allow us to grow and grow and develop as an organization?
Evan Wimpey: Yeah, that’s great and generates a lot of that buy-in and that teamwork that you shoot for getting the business to come to you to ask, Hey, help us solve this problem. And I don’t know that this is fair to say, but my intuition is that humility makes for a better technical data scientist, too. When you’re exploring data and you’re willing to question your intuition or question what you’re seeing in the data and you’re a little bit skeptical of maybe your results or what you’ve seen, just an exploratory analysis, I feel like those two things would go hand in hand.
Alex Cunningham: Oh, very much so. You know, and it also lends itself to a little bit of that, that natural inquisitiveness that we also tend to find in the data science group. These are folks who for a very long time have always wondered why something works or how does that work? And, you know, might have tinkered as a little kid, for example, taking computers apart and putting them back together, toasters or anything you get your hands on, right? That kind of engineering mindset is also something that you know is important and allows you to, you know, as you’re being empathetic around what have I not considered or what might I not know myself? You have that natural curiosity to go be that lifelong learner, that self-starter who then goes out and says, Well, I’m going to go ahead and read up on that just a little bit.
Evan Wimpey: Yeah. Perfect. OK, Alex, I want to ask you, I want to ask you one last question. You’ve talked a bit about how projects get prioritized, and how you help production allies or operationalize a data science project that has been built. Let’s put that to the side. It’s just the Alex Cunningham show. You’ve got your easy button and everybody is aligned to your vision and whatever. Whatever your focus is, the business problem, the thing that you want to solve, everybody’s onboard and they have the exact same vision that you have. So given that unrealistic scenario, what’s the type of problem you want to tackle? Where do you put your data science effort?
Alex Cunningham: So that’s a really interesting question, and the way I think about that maybe might be a little bit different in that. I believe if I could snap my fingers and have any kind of behavior change or project come to fruition, it wouldn’t be this. It would be creating time or more time, I should say, for our data science teams to learn to be comfortable with providing time to do that type of self-learning. You know, one of the things that Gerhard mentioned that I think is important is that that idea of being a lifelong learner and good data scientist and good data science groups embrace that, that learning aspect. And so one great thing is to devote real-time, real resource allocation to just true learning. And you heard a really interesting fact at a talk the other day that you know, do you know how often the average academic paper is read?
Evan Wimpey: Two dozen times?
Alex Cunningham: Well, maybe by their family and loved by a parent who’s trying to support them, right? Well, you know, the sad thing is that on average those articles are only read about three times. And so think about the amount of learning out there and sort of the reinvention of the wheel, if you will, that is going on because stuff is out there and just not being consumed or read or talked about as much. And so, you know, imagine you could take a data science team and, you know, bring that number up from three articles to six. Let’s not go too crazy here. What might that do for an organization? That, to me, is a bet worth making. And something that we here at Church and Dwight and on this team are going to try and start to devote more time to because we feel that by doing that, all these other projects and things will flow from it.
Evan Wimpey: That is a fantastic answer. And I, I feel fortunate for your team. I hope you’re able to carve out at least sometime, probably not as much as you want, but I hope you’re able to carve out some time that they’re able to do that and get that readership and citation up on some of those academic journals. Alex, thank you so much for taking the time out of your day to come here and chat with us. It’s been super useful, super helpful. Alex Cunningham, thank you very much.
Alex Cunningham: Hey, thank you. It’s been an honor and a privilege and I’m looking forward to doing more in the data science space and helping to be a good partner to transform our organization, an organization that I very much love into our future analytical competitor selves.
Evan Wimpey: Fantastic. Alex Cunningham, thanks so much. Thanks for listening, everyone. That’s all the time we have for today. Be sure to like, subscribe, multiply and divide, and we’ll see you next time on the Mining Room Business podcast. So back with Alex Cunningham for a quick joke that he hopes to tell. Alex, you’ve got a statistics joke for us.
Alex Cunningham: Oh. Evan, you’ve asked me a lot of questions, I have a little bit of a question for you. Why was four afraid to ask five out on a date?
Evan Wimpey: I couldn’t tell you.
Alex Cunningham: Well, unfortunately it was too squared.
Evan Wimpey: Oh, Alex. Some numerical humor. There we go.
Alex Cunningham: Hey, I’ll keep my day job. But OK, we always got to have a little bit of fun with it, right?