From Healthcare to Banking: How Data Analytics Skills Unlock Career Mobility (Ep 011)

In this episode of the DATA SECRETS Podcast, I sit down with Noelle Butler, Vice President of Product Management at JPMorgan Chase, to uncover how mastering business fundamentals and relationship management can drive massive impact in analytics and data product development.

Noelle shares her journey from home health care accounting to leading enterprise data product teams at one of the world’s largest banks. We discuss:

• The “secret sauce” to career growth in analytics: learning the business and building strong relationships
• Real-world stories from Noelle’s time in home health, automating nurse-patient matching with early Tableau, wrestling with legacy tech, and optimizing onboarding to reduce costly nurse turnover
• How data-driven insights move the needle, from day-to-day operations to high-level decision making
• The evolution (and challenges) of data product management inside Fortune 500 organizations
• What makes a true “data product”? The difference between dashboards that deliver value and those that gather dust
• The importance of organizing analytics teams for impact and staying connected across complex organizations
• Noelle’s philosophy on change and innovation, plus her take on how AI is reshaping the data landscape

Whether you work in healthcare, finance, or any data-rich industry, Noelle’s stories and strategies offer practical advice for anyone looking to turn insight into action and make data products that actually get used.

🔗 Connect with Noelle on LinkedIn

Follow the DATA SECRETS Podcast

📬 Get episode recaps & bonus insights at allegroanalytics.com/podcast
🕺🏻 Connect with Nathan on LinkedIn
📺 Watch every episode on our YouTube Channel

Transcript

Nathan Settembrini [00:00:02]:
Your data has secrets, secrets that could change everything if you only knew where to look. Welcome to the Data Secrets podcast. Welcome to the Data Secrets podcast where we uncover how business leaders crack the code, find truth in their data, and turn insight into action. Today we have an amazing analytics leader, Noelle Butler. Noelle, welcome.

Noelle Butler [00:00:30]:
Thank you, Nathan. I'm excited to be here and excited to talk about my journey.

Nathan Settembrini [00:00:35]:
Awesome. Well, you. Okay, so you started as an accountant in home health care, counting beans, and you worked your way into business analytics and now you're a vice president of product management at JPMorgan Chase, leading the development of enterprise data products, which I'm super excited about. First of all, amazing. Congratulations.

Noelle Butler [00:01:00]:
Thank you, thank you.

Nathan Settembrini [00:01:01]:
Second of all, what's your secret sauce for our listeners?

Noelle Butler [00:01:05]:
Yeah, you know, I was thinking about things that I always share and recommend to others kind of when they're on their journey, things that, what I always focus on. What's my secret sauce is really two simple things. One is learning the business. No matter where you work, the business, the industry, how they make money is going to be different, but the principles stay the same. But I feel like in order to be successful, it's really learning the business. What do we do? How do we make money? But then the other one I always come back to is relationship management. If you have good relationships with your stakeholders, if you have good relationships with your team, then you can all come together and get things done. And I think I've noticed that that's the most important, really one of the biggest things in my career and why I've been successful and why I am where I am today.

Nathan Settembrini [00:02:01]:
That's amazing. Okay, so learn the business and build relationships.

Noelle Butler [00:02:05]:
Yeah. Yeah. Thank you. Yeah, I find building relationships, I mean, I have stories, but you know, when, when you finally find that like common ground with someone, whether it's personal or business, like you really, you really just kind of take off and both of you can work together to get things done.

Nathan Settembrini [00:02:26]:
Yeah, totally. Yeah. I mean, my father in law, he says that all the time. He was, he's been a consultant basically his whole career. And yeah, he'd always say, how does the company make money? Uh, because like, if you can figure that out, then like, I mean, the reality is that exchange of money is in exchange for value. Right. And so it's really a, it's a, it's a highlighter to help you identify, okay, where does the business create value for its customers?

Noelle Butler [00:02:56]:
Yeah.

Nathan Settembrini [00:02:57]:
And if you can identify that and then all of the people, processes, technology that go into making that happen, Then you get pretty close to the action. Um, that's awesome. Yeah.

Noelle Butler [00:03:09]:
Yeah, I agree with that. It is. It's something I struggled with relying on at first when I, like you, were talking about my career jumping from home health to a big bank. And I knew that business inside and out when I was at home health, but I really just relied on my skillset when I first joined the bank, and relying on my skillset and making relationships with people, then that knowing the business eventually came and just became an easy piece.

Nathan Settembrini [00:03:41]:
So, yeah, that's awesome. Yeah. Some industries are harder to jump into. You jumped out of a. An industry, healthcare, that's hard to jump into for other people. Financial services. You know, I worked at Equifax for four years, and, you know, it took me a little bit to understand, like, wait, so we're calculating scores on people so we can decide if we're going to. If banks are going to give them money. The banks give us their data, we score it and give it back to them and sell them for both services. That's amazing. What a business model.

Noelle Butler [00:04:18]:
Yes, I. It is interesting, but we have that in home health, too, so.

Nathan Settembrini [00:04:25]:
Yeah, interesting.

Noelle Butler [00:04:27]:
Yeah.

Nathan Settembrini [00:04:27]:
Well, I'm excited to dive into your data secrets. I mean, you have a wealth of experience, and I'd love for you to be able to share it here. And so, you know, so the Data Secrets podcast, it's all about, you know, we want to hear stories about uncovering insight and turning, like, you find something in the data and you turn that into business impact. You turn that into action. And so I am curious, like, let's go back to your time in home health first, if that's okay.

Noelle Butler [00:04:57]:
Yeah, of course.

Nathan Settembrini [00:04:58]:
And maybe give us a little, like, context about, you know, what is that business? What are some of the different data sources? How did you kind of, like, what were some experiences that you had leveraging data, you know, moving from accounting into business analytics? You know, where would you like to start with that?

Noelle Butler [00:05:18]:
The business itself was home health care, which is nurses visiting clients in their home. The model did expand to becoming partners with hospitals and having joint ventures. But the meat of it is we want our clients at home. They do better recovering from their injury, their illness, whatever it is, when they're home. The data was in its infancy stages, but also we had a lot of smart people behind the scenes building data models the way that they should be. So we pulled from Oracle. When I was there, we weren't on the cloud yet. We had, you know, a system for hr, like, workday, so that Was like our employee revenue budgeting, all of that, all the financial stuff plus employees. We had one system that, if you can believe it or not, was made from the old, like DOS system, like the blue screen with the text. That was our main healthcare system created by the owner and his brother. And I remember when I was leaving, they were still using it. They kept trying to find replacements, but it was so customized that it did everything we needed, even though it was built in this old, old structure.

Nathan Settembrini [00:06:39]:
Yeah.

Noelle Butler [00:06:40]:
But then we also had newer systems to connect to that were used across hospitals and so on and so forth. So our biggest issue that was like a day to day issue in these offices was figuring out how to match a client with an employee. And what time could they do it during the day? What days of the week did they need? Just an hour or two. This was an interesting problem to solve. There was internal hackathons to try and figure out. It was fun. But the main issue was you would think that, hey, this is a pediatric client. I need a pediatric nurse, someone who handles kids, is a nurse for kids, while that client has a dog. And the perfect nurse for the situation is allergic to dogs, so she can't go there. So it's not just skill set and location and timing. There's all these other factors that went into it. And really what was interesting that we found was the offices that employ these nurses and they had all that in their head. Right. There were some notes in the employee file and client file if they liked, if they were allergic to dogs or not or cats or whatever the preference was. But, you know, a lot of it was they just knew it. So how do we make that visible? And how do we automate matching clients and employees? So that just takes a lot of time. And it's very like person specific. What we found is we started bringing in those descriptions in the data and tried to parse things out. At the time we were using what's called Microsoft Power Pivot. Now people start to use Power Query. It was like, what's behind Power Query? That was our data prep tool. Having to bring in all of this client data, all this employee data, and how could we build a solution in tableau that these service offices could use? Because there was about, I think like 3,000 service offices at the time across the United States. And how could we help these case managers figure this out? And so I'm sure it's way past where it is today. But being part of the first iteration, they created master employee dimensions, master customer dimensions, all of the standard data Modeling, and then I was able to actually ingest that and use that in tableau. A lot of it was parsing out the location, the age ranges, and the preferences. But we started with the basics. This is a client who just needs basic needs, so they don't need an RN coming into their house. We started classifying things and started making matching based off of roles. So we had these roles of types of clients and types of nursing, and then also location based and then schedule. We started building all these things into the system. The very first iteration was tableau dashboard map, and we use specific shapes for the clients and specific shapes for the employees. And as you zoomed in on your office, you could also then filter down to your caseload and see the employees and clients on a map so they could see physically where they are. And then we could start using these matching logistics in the background to match them together so they could figure out, all right, here's this client, and these are the appropriate employees for them. And so there were dropdowns for client and employee, and then they would use their system, if they needed to, to look up the id. But generally, again, they knew a lot of this thing. So that was one of the biggest cross the entire company issue that affected the bottom line. It affected their day to day. And that was like a really cool solution to work on.

Nathan Settembrini [00:11:00]:
That's awesome. So, so effectively you have demand for, like, clients that need to be seen. You have supply of people who can go see patients. Patients. Yeah. And. And you're trying to match supply and demand at scale. And so effectively you were, you were creating tools for the local people to help kind of make those assignments. And then they would, I guess in some other system, they would then go say, okay, nurse A is going to go see patient B on this day at this time. And they kind of block that in. And now they see that nurse A has her morning booked, but the afternoon is not. And so, like, they kind of like systematically go through and they're like, okay, well what would be the next good house for them to go drive to that whole thing? And that's where the map came into place. Okay.

Noelle Butler [00:11:56]:
Yeah, exactly, exactly. And I, you know, it was very basic at that time. I think we're on like, tableau eight or nine. I don't know.

Nathan Settembrini [00:12:03]:
Oh, yeah, yeah. You and I started in tableau at the same time.

Noelle Butler [00:12:06]:
Yeah, yeah, yeah. So it's been, you know, it was just kind of that first stepping stone. I know, I've heard through the grapevine now they have applications where nurses can sign up and are doing this a little bit smarter, but you know, for what the tools that we had at the time, that was a pretty cool project to work on and really like widespread.

Nathan Settembrini [00:12:27]:
Yeah, that's very cool. Yeah. It's funny, I at the same time was doing a very similar kind of thing, but totally different application. It was employee mapping around where they work and doing like drive time analysis and seeing like, oh, you know, if we move the office to the suburbs, then all these people's commute is going to double. And so then like there's a potential for attrition and stuff like that. So it's fun. That's the beauty of analytics. You know, it's like there's a bunch of different ways to do everything and then also there's a lot of similarities between solving different problems. So.

Noelle Butler [00:13:09]:
It's very true. It's very true. I appreciate you calling that out because it is, it's, it is what it boils down to. It's supply and demand to your point. Right. And so supply and demand looks different everywhere.

Nathan Settembrini [00:13:21]:
Yeah. So were there any other either really like big impactful projects that you worked on in home health or do you want to move on to banking?

Noelle Butler [00:13:34]:
There was another like smaller scale story that I found interesting with home health that was related to our staff, our nursing. And we found that we had very strict training and development. So not every, there's tons of small home healthcare companies and other businesses out there. We were pretty large and we had a pretty good set of quality and standards. And so there was this robust rubric. And if you were a nurse that came in, even if you have these certifications and trainings, you have to come in and be certified and trained a certain way by us. They found that they would go through onboarding and onboarding took about 60 days and they would drop off. They wouldn't ever bill a shift. So we spent all this money training them and then they just dropped off because training took so long. That's what they were finding. And so we wanted to figure that out. We have this inkling, this onboarding problem. This onboarding program is taking way too long. What if we reduce it to 30 days? Will we get more people hired so that we can bill our clients and serve our clients? And so that one was an interesting one because there was already like a hypothesis. And so we used the data to really track and confirm that if we made this change, did it really make that impact? And so it did. I mean, it was pretty cool to watch. You could actually look at the retention trend or actually I would say like the turnover going down. So it was cool to also see proof in the data of something that, of like a change in the process. So I thought that was another fun little project that we worked on too.

Nathan Settembrini [00:15:12]:
Interesting. Yeah. So the, so the nurse training took too long. So they didn't ever actually make any money or they just got bored.

Noelle Butler [00:15:21]:
Correct, yeah. Okay. Well, they were getting paid, but not at their like high bill, high rate. They were getting paid. We were paying them. We weren't making money off of them. Cause there was no revenue yet. We were just training them. So really it was on the business's profitability. But then also we couldn't serve a client. So now we're losing business if we can't serve them. So bottom line, profitability. Yeah.

Nathan Settembrini [00:15:45]:
Was the solution to just make the training shorter?

Noelle Butler [00:15:49]:
Yes, yes.

Nathan Settembrini [00:15:50]:
Okay. It kind of reminds me of the, you know, like Facebook's like when they first launched like one of their early, like as they were growing, what they figured out was like, if, if you can get somebody to have like some number friends in the first week that they're on the platform, then they'll stick around. And so that's interesting.

Noelle Butler [00:16:11]:
Very cool.

Nathan Settembrini [00:16:12]:
That's awesome. And that's such an impactful thing because I imagine like recruiting all of those nurses to come on, spending a month or two with them only to have like all that cost or that investment in those people.

Noelle Butler [00:16:28]:
Yeah.

Nathan Settembrini [00:16:29]:
Just go poof, out the door. You probably just trained your competitor and perfect. Um, and so, yeah, that's, that's a huge impact. Um, yeah. And then also it's a win, win, win. Cause it's like the, the nurses get to make money faster, they get trained faster. Was there any, was there any downside to that? Because like oftentimes, you know, you can optimize for one metric and then, yeah, you know, you cause something else to happen. Right. There's like second order effects. And so did you, did you also, I don't know, issues that might have shown up because of they weren't fully trained or you know, that whole thing.

Noelle Butler [00:17:08]:
Yeah, I, I don't remember that. We had like metrics related to like the outcomes of maybe the patients. We really just focused on the onboarding. But I think what they found is that the skill set, like what's the minimum amount of time that they could kind of train this skill set? And so it really depended on the level too. So for certain levels they could decrease it and for other levels maybe they couldn't. So yeah, I didn't. I don't remember if they took it that much. That much further though.

Nathan Settembrini [00:17:43]:
What. What were some of those, like, key metrics that told you that the business was doing well or like that the service. So I imagine there's like, obviously there's revenue and there's profit and blah, blah, blah. But like, in the execution of the service, like, were you tracking patient satisfaction or. Yeah, like, what are some of those metrics that you guys.

Noelle Butler [00:18:05]:
I mean, patient satisfaction is a huge one. There are always like surveys going out. You can also tell based off of. I mean, there's insurance metrics that there's basic metrics saying this client should be in your care for, let's say, six weeks. Right. Based off of this surgery they had or this injury they had. So if you're going to. Over six weeks, is that an indicator of they're not getting enough care or an indicator that there. Maybe we didn't calculate this correctly. So there's always that metric of time of care. And so that was a big one to track because you get it, you kind of have these standards and you can look at the time when you start actually seeing them, how many visits you're seeing them in within that time too. So yeah, you start to look at those visit metrics.

Nathan Settembrini [00:19:00]:
Really interesting. We're working on a project right now that's a readmission dashboard. It's on the payer side. And so they care a lot about, hey, did we solve this the first time? And if people get readmitted, then it's like, maybe we didn't. So that's interesting. I remember that it's almost like adherence to some expected outcome.

Noelle Butler [00:19:25]:
Yeah.

Nathan Settembrini [00:19:25]:
And it's like, oh, are you not changing their bandages effectively or, you know, now they're back in the hospital or it took longer to recover. That's. That is fascinating. Awesome. Okay, so you, you were in health, home health care for how long? Like 10 years or something?

Noelle Butler [00:19:43]:
12 or 13.

Nathan Settembrini [00:19:44]:
12 or 13. Wow. Yeah. And then you heard the call of banking. Talk us through your, you know, why'd you leave? Or like, what, you know, what drew you towards Chase? I mean, Chase is an amazing company. We bank with them. Big fan. Especially after the like, SVB stuff. It's like, okay, what's the biggest bank possible? Yeah, the safest place. But.

Noelle Butler [00:20:16]:
Yeah.

Nathan Settembrini [00:20:17]:
What drew you there?

Noelle Butler [00:20:18]:
The technology and the opportunities. I really. Being in health care, like you said, is maybe hard to get into. There's like smaller opportunities, not as many opportunities there. But also, you know, when you're in A technical job, you don't want your skills to get stale. And I just thought it was my time to grow and just wanted to take on a new challenge. I feel like I knew the business so well and so how to challenge myself. And so those are my biggest two things. I want more technical skills and I want to learn something new. Yeah, so that was my biggest drive. Yeah, keep learning. That's my biggest thing. Challenges don't have to be the next step up. They could be collateral and new data, new industry, new clients. So yeah, that was my move.

Nathan Settembrini [00:21:16]:
And so when did you join and what role did you step into and what was that transition experience like?

Noelle Butler [00:21:22]:
Yeah, it was. I joined in December 2020 and the experience was remote. Onboarding was not easy. I remember like the first day during COVID You're not in an office, very different atmosphere. But I will say everyone was great at reaching out on my team set up time with me to introduce themselves. We had great team engagement people who would set up meetings and just introduce ourselves and have fun. And so it was an interesting experience obviously because it was during COVID So it's just very different. But remote work wasn't hard for me in general. Like I could, I can handle that. I can, I could, you know, onboard myself and get myself familiar. So, you know, I came in as just a developer. I was excited to just be a high level individual contributor. I took over the area. I was assigned to the area for community banking, consumer community banking. So like you said, you have your bank with them. But my specific area that I'm assigned to is auto loans and leases. So not just joining a bank. I now have to learn about the auto industry. So that was like one task in the beginning. And then from there, within six months I was asked to lead the team, lead the book of work, lead product updates to the CFO and the CEO of Otto. Yeah, it was pretty crazy, but it also wasn't something I was unfamiliar with. I did a lot of at my prior role, so it wasn't that scary. Like I had that experience luckily. But it was tough because I was still six months new to the role. Felt like I just started feeling good about delivering dashboards and data and then thrown into that mix. But from there it just took flight. I got to work on like more initiatives across the department and not just for auto. And I got to. And now I am leading not just our dashboard side, but I'm leading auto as a whole. So I'm leading our data and our tech side as well. So now I Get like the full product under me. And so, you know, it's been an interesting journey and I just hit five years on Monday, so that was a shock to the system.

Nathan Settembrini [00:24:00]:
Congratulations.

Noelle Butler [00:24:02]:
Thank you.

Nathan Settembrini [00:24:03]:
That's awesome. So along your journey, were there any kind of data, secret stories where you and the team dug something up that was really interesting and kind of you followed the breadcrumbs and then made some sort of a business impact?

Noelle Butler [00:24:25]:
I don't want to say it's not as fun as home health, but I would say it's very financially driven. Driven. The team that we're in is we're serving the finance team, we're serving financial planning and analysis. And so it's definitely been an adventure trying to push the envelope and the data. I feel like in finance, especially P and A, they're looking at budgets and forecasting, they're iterating as much as they can. They're doing they know the data or they're trying to confirm things that they already know. So uncovering things has been a little bit different. But I will say that we've had some cool projects related to data for our partners because we have to provide, we have contractual obligations, we have certain manufacturers that we are their bank. And so if you go to buy, I always say Subaru, one of our longest customers, you and you don't use your own bank and you go in there, it's Chase. So your loans can be from Chase. And so what I found that was interesting about that data is there's like very specific contractual metrics that they have to follow. What rates are you giving your consumers, like auto buyers when they walk in the door? Chase only allowed certain ones. And then also just maintaining the relationship is really important with the data and how they use it. One of the biggest projects, I wouldn't say we uncovered a secret, but was giving them access to that data and those dashboards. So we had all these internal dashboards and we're just exporting them into 50 to 100 page PDFs. Oh goodness, don't we love a PDF. And so all this, like fun work going into this PDF. And then we worked with Tableau, the vendor, obviously, and they said, hey, we do this for a lot of other companies where we just turn the. You can turn the dashboard on externally. So that was a cool little project to do because we were able to find a solution, not make something too complex that was internal. We're able to add value to these customers now. So now we're not just giving you this PDF, we're Giving you this portal and you can go find this information however you want. We hope to eventually add more insights to it. And it started a whole model for the business. So no one was doing this. We were the first to do it in the business. So now we have other lines of businesses looking into it to serve their customers. So at some point, maybe it's a revenue model too. Maybe it's something they start charging for. When you have. You're giving these reports externally now, it's like an additional benefit. So there's lots of opportunities there. So I thought that was just like a fun project and it wasn't really a data secret, but I think it's just something else that we get into. And solutions we have aren't just always within the data itself.

Nathan Settembrini [00:27:40]:
Yeah, I see that. I mean, that's an amazing, empowering thing that you did for your partners. Right. And so it's like, you know, the whole, like, data democratization, you know, diatribe that everyone was chanting 10 years ago or whatever. I mean, you guys did that. Basically you. There was. There were hidden secrets that were on your side. It's kind of information asymmetry is like, we've got all this stuff over here and in order for us to tell you about it, we have to. Here's a PDF, you know, that took a bunch of manual effort and like, and actually took probably the analytics that you had, built like really cool dashboards and made them kind of like less powerful because you can't click and you can't drill down or whatever. And.

Noelle Butler [00:28:27]:
Yep.

Nathan Settembrini [00:28:29]:
And I've been part of these product projects before where it's like, okay, we have our server in house and so actually I do want to go like for just a minute for our tableau people, I am just curious, like, how. How you did it. Like, were you guys on prem? On server. On prem. And then you decided, hey, we actually, we need a separate server. Did you separate site? Did you tablet cloud? Like, what?

Noelle Butler [00:28:55]:
Yeah, jumping into, like the specifics we used, we were in between on prem and cloud. We're now on cloud. But when we started this two years ago, we were on prem. So in order to do this, they had to be on cloud. So so now we had to sign up for cloud and we had to use it. We did have to create a new site. So we have internal and external. So that was my. I was worried about that because are we now publishing to two areas? Do we have to worry about, you know, if I published here but didn't publish here, and we have Solutions for that to make it seamless now. But they have their own site, we have a little portal. So if you're external, you put your name, you put your work email address and then it'll go to our tableau team internal to permission them for the right brand. We had all of those standard rules that we use internally and did the same thing for external. Yeah.

Nathan Settembrini [00:30:04]:
And the content, were you. So were you partitioning it by partner where it's like Subaru has a project or Subaru. Are you doing row level stuff where it's like if I'm part of Subaru, then I only see Subaru's data. But I'm looking at the same dashboard as gm, so.

Noelle Butler [00:30:21]:
Good question. These dashboards are very specific for each partner. So they all have their own projects. They may have similar metrics, but their thresholds and everything are very different. And then they like you, the data behind it is like so scrutinized and very detailed and very different for each partner. So unfortunately we couldn't just build one dashboard and have a drop down from our level security to make it easy.

Nathan Settembrini [00:30:48]:
Yeah, I mean the three reasons why. I mean you checked all three boxes. It's like the people want different information. The way we present it is different and the underlying data is different. So therefore a data product, you know, for to serve everyone is not really on the table. That makes sense.

Noelle Butler [00:31:10]:
Very true. Why make it easy?

Nathan Settembrini [00:31:12]:
That's awesome. So that's actually a great segue.

Noelle Butler [00:31:15]:
Yeah.

Nathan Settembrini [00:31:16]:
I do want to talk about data product management with you because, you know, that is your title. It's a fascinating concept. Yeah. Because I don't think a lot of people do it. Well, some people don't do it at all, you know, so. I see. And actually we didn't talk about your data horror story. I do want to talk about that too. But so in the data product management world, it's like on one end of the spectrum you have order takers. So it's kind of like, hey, just the business says build me this, build me this, build me this, build me this. Right. And you're just like reacting. Right. And so this is kind of like the traditional IT model which, you know, it works in some situations, but for analytics, you know, I'm pretty passionate about the fact that we, we should not be order takers, we should actually be partners with the business, you know. So on the other end of the spectrum you would maybe. And I don't know what spectrum this is, but yeah, you could imagine a world where I, as the analytics product Manager just feel like I know everything about the business and I just give them stuff. One end of the spectrum, I'm giving them exactly what they want. I filled out their tickets. On the other end of the spectrum, I'm giving them exactly what I think they want. And really neither is correct. Right. It's kind of like there's this middle spot where it's like, yeah, and maybe I'm stealing your thunder, Sorry, but.

Noelle Butler [00:32:52]:
No, you're fine.

Nathan Settembrini [00:32:54]:
I feel like there's this middle area where we collect requirements and we understand what the business needs and we work with them to create the right solution. It's kind of like one metaphor I've been using recently is if you take a chisel. So a chisel is sharpened on one side and it's almost like a blunt instrument, right?

Noelle Butler [00:33:18]:
Yeah.

Nathan Settembrini [00:33:18]:
And so like on our two ends of the spectrum, it's like we've got, you know, the chisel on one end or that it's a chisel on the other end, where we're just kind of like doing one or the other, but like in the middle, we sharpen the blade on both sides and we end up with like a way better, more effective tool.

Noelle Butler [00:33:37]:
Yeah. So that analogy, maybe using it soon.

Nathan Settembrini [00:33:44]:
So I'm curious. We'll say there's a data horror story for after this section, but. So stay tuned for the data horror story. But let's talk about data product management for a second. How do you define it? How do you think about it? What's your philosophy?

Noelle Butler [00:34:02]:
My philosophy is I'm stealing this from a colleague. But we're not delivering just data, we're delivering solutions. And we're also building something you want people to rely on. If they don't rely on it, what does it mean? What are you working towards? So those are my biggest philosophy and items I think of and what product management is. You want to help every level make a decision. Whether you're an analyst, a financial analyst, and you're trying to help your boss answer a question about why a certain line of business wasn't profitable, or you're the CEO asking, should we renew this partnership with this auto lender, Those are big questions to ask and need different levels of data. And if it's not reliable and we don't deliver a solution for them, that's just how I think about it. We want to reduce uncertainty. You want to have that decision support available. And so that's what I think about data product management. Obviously, there's a lot of very, very technical guidelines and different systems built when it comes to having official data product management. And that's what I found the difference going from healthcare to Chase, we have those guidelines, we have those roles and responsibilities. And before I didn't have those, but I was trying to operate like I did. So to your point, there's that struggle of who's building this and if you don't put some ownership on the stakeholder, are they going to use it? Are they going to care about it? Are they going to have other people use it? Are they actually going to make decisions based off of. It's kind of like a long winded way of saying it. But yeah, I think, I think to your point, you need ownership of the data, you also need ownership by the stakeholders and you guys need to work together. And I think that kind of brings me back to relationship management from the beginning and why it's so important.

Nathan Settembrini [00:36:17]:
Nice. Yeah, yeah, that's awesome. Okay, so Data product Manager, Data Product Owner. I've heard these titles before. Maybe for someone who's brand new to this concept, what would you even define as the thing that we're managing? What is a data product?

Noelle Butler [00:36:40]:
Oh wow, that is very funny. We've been doing something in my current role defining data products, I'm like, it just sounds so whimsical. Right? What is a data product? Yeah, what I found is it, it's like a curated set of data or you could even just say like knowledge through data. And you're trying to like this podcast, you know, it's the case file, you're trying to solve a crime, right?

Nathan Settembrini [00:37:14]:
Yeah.

Noelle Butler [00:37:15]:
So I think managing and ownership related to that is really, you're building, you're building this case file, you're understanding what the problem is and then from there you're trying to figure out do we have the data and how can we use it? What's that feasibility of all the technology that you have. Can you solve this problem, this crime? I find it to be ownership and management. What is a data product? I think of, I work in the auto line of business. So we have auto data, but we have auto financial data, we have auto competitive data, we have data related to rates that we give our customers, we have data related to profitability. So when we've been talking about data products, it's really auto as a whole, but then within auto, what's your decision that you're trying to support or what problem are you trying to solve? And kind of look at the different types of data you have and package that together so that the user can solve their problem.

Nathan Settembrini [00:38:27]:
Yeah, I think if you. Yeah. So like, if we, if we think about what is a. What is a product? Right. Not even like a data product, you know, so like Google says, it's. It's, you know, there's a market and it's basically anything that goes into that market to help solve a particular need. Right. And so in our case, our market is our stakeholders, and we're building things, dashboards, data processes, data science models, data apps, those kind of things. Right. That basically help our stakeholder achieve some goal. Right.

Noelle Butler [00:39:11]:
Yeah.

Nathan Settembrini [00:39:13]:
And so, yeah, I think the interesting thing is if you don't take that mindset with the dashboards you build, then you end up with hundreds of dashboards that get one or two views and then never get used.

Noelle Butler [00:39:30]:
Yeah.

Nathan Settembrini [00:39:31]:
Which. That's the biggest bummer ever. Right. You spend all this time and effort and it's like you build something that sits on the shelf. It's not fun. Yeah. I see data product management as an active way of helping prevent that from happening.

Noelle Butler [00:39:48]:
Yeah. Yeah, that's interesting. I think. Yeah, it's very true because our products can be a lot of different things. We happen to use Tableau, but we also have other reporting tools now, like Thoughtspot and Databricks. And so we've gone from, like I said earlier, we're not just dashboard team anymore. We're like a reporting solution team. Like, what's your need? How do we solve it? Maybe we don't just need to use Tableau. Maybe it's something pretty simple. And so, yeah, the products are definitely changing too.

Nathan Settembrini [00:40:26]:
Yeah. How do you guys organize from a team perspective to kind of like you're a data product manager. You're like, you know, so you're kind of like oversight. It is. Are there little pods like you all over the business or is it kind of a hub and spoke or. And then what does your team look like?

Noelle Butler [00:40:50]:
Yeah, it's. My team looks like we are. I'm the product, like owner, manager. It's kind of used, interchangeable of auto. And I have a team that is called our tech team. They build the data pipelines. Like you said, they're very, very technical. They're using scripting language, they're building the data marts. They're doing all of the very technical work before we even ingest or consume it in reporting. And then within that team, we also have business analysts. So they have the technical skills, they have the business skills. They can translate. So they help our tech team, they help our stakeholders and try and refine their requirements and translate it between the two. And then we have our what used to be just dashboard team, it's now business intelligence team. Very specific on the reporting like solution, like the end result. So taking the requirements that the business analysts worked on with our stakeholders, working with the tech team to get that data available for us and then building the actual dashboard. So like the developers of our reporting and so that's very systemic across our line of business. Our line of businesses, I should say that structure. Yeah, yeah. So like I have that in auto and then you go to another product in Chase and they have that same structure and then it just keeps iterating. And so we used to be, we call that product aligned. So everyone's under the auto product and they roll up to the auto product. So that's what we call product aligned. We used to be skills based. When I started I was only in the team for two months when that changed. But instead of me rolling up to and being aligned to everyone that does auto work, so like even the auto finance team, who are our stakeholders, I was in the dashboard development team. So all of us who develop dashboards were in the same team and we were kind of contracted out. Like we would come back together on team meetings and stuff. But then you would go to your stakeholder meeting for auto and so again I was only in that for a few months, but. And now it's been product line. So that's how we're structured.

Nathan Settembrini [00:43:18]:
Yeah. And I guess the thought process behind that change was hey, you know, right now everybody's centralized and we're kind of, we're sending people out and they're coming back and you know, there's all this, if we move the team inside the business line then now they're closer to their stakeholder and the products that they develop are going to be more aligned with what that team needs.

Noelle Butler [00:43:42]:
Yeah.

Nathan Settembrini [00:43:43]:
And like the downside is maybe you're less connected to other people in the organization who do what you do. And so is, is there a mechanism to like make sure that everyone who does like your team does, that you're aligned on best practices and tooling and all that kind of stuff?

Noelle Butler [00:44:01]:
Yeah, that was the biggest concern and still is at times. We have a weekly what we call like cross lob calls. So all of the, all of the like dashboard BI developers are on this call from the different lines of business and we kind of stay up to date, share best practices. If there's something that's going to affect, if someone sees an issue or finds something that affects them, hey, I think this might affect other people. It's kind of like that communication. So that has been the biggest one to stay engaged. And then we use teams and things like that just to chat during the day or send emails. Like, hey, I saw this. I don't know if this is going to be an issue for you, or, hey, I did this cool new thing. Check out this dashboard I made. Like, I'm really excited to share it. So that was tough at first, but it having that weekly meeting and the team's channel really, really helped foster that a little bit better.

Nathan Settembrini [00:44:58]:
Yeah, yeah. Because what you don't want is duplication of efforts. So, like, if someone builds a data pipeline from some system, some loan system. Right. And you figured that out and you've got the data piping in and it's perfect. And then some other team is like, we need to get data from this loan system. You build it twice. Right. And it's like, yeah, yeah, what a waste. Right. And so that's good that you guys have. You have mechanisms at least in place to try to sync the team up.

Noelle Butler [00:45:30]:
Yeah.

Nathan Settembrini [00:45:31]:
All right, so let's hit your data horror story. So what's an example of a time when things didn't go perfectly?

Noelle Butler [00:45:42]:
Yeah. One that comes to mind is when I was in healthcare, we published data externally. So who are. We would get questions from different newspapers or, you know, like digital news and different forums to highlight our business. And our business is in a lot of places. So depending on where we were, what state were we in, what newspaper was requesting it, what was the context behind it? There was one instance where we published it was. It's usually like the number of clients, number of employees and revenue. And you kind of truncate the data so you're not giving the exact number but a roundup round down. So it's not exact because we weren't public, we were private. So we would give this snapshot out and I remember running the numbers and I remember it being wrong. Right. So this is getting published. Think it was slightly higher. It was higher. And you don't want to inflate anything. And so I think, like, thankfully, where I worked for, who I worked with, it wasn't. It was a big deal because it was published incorrectly, but it wasn't like, oh, my gosh, the world is ending this job. You're never going to go anywhere. It really just set the tone for how, like, how do you handle something when it goes wrong? And really it's just accepting the mistake, identifying what you can do better. And so that was that actually create. Helped us create, like, a standard model for this intake. And how do we define a client? How do we define an employee? Questions to ask so that if we don't know what they want, what are our standard questions to ask? Like, what state do you need this for? What county do you need this for? Very specific. It wasn't just a fire drill. Every time we would have a rubric of questions, we'd have a checklist, and we created defined ways to do all of those things. I remember we had a spreadsheet at the time, and we'd be pulling from our data sources and these exact filters to use for certain instances and things like that. It was, you know, I was, like, very scared. I remember at first when that happened, but the outcome was, like, just made the system better. So that's where I try to learn from and teach everyone else, too.

Nathan Settembrini [00:48:25]:
Nice. Yeah. So it sounds like you almost. You created some processes, and it's almost like you created a data product for that to start. Nice call back.

Noelle Butler [00:48:38]:
Yeah, yeah, we did. We did.

Nathan Settembrini [00:48:42]:
Yeah. Cool. All right, so let's jump into our lightning round, our metric mystery theater. I love this section. First question. What's the weirdest metric that you've come across?

Noelle Butler [00:48:55]:
Maybe this is boring, but I really just want to say, like, number of dashboards. You know, we have a dashboard for dashboards. What does that mean? Yeah, what's good about that number? Am I. Are more dashboards good?

Nathan Settembrini [00:49:10]:
Were you measured by number of dashboards?

Noelle Butler [00:49:13]:
Yeah, users and, well, users as well. But we do look at the number of dashboards. It's very interesting.

Nathan Settembrini [00:49:21]:
Interesting. Yeah. I mean, that's. If our listeners want to go back to episode nine and hear from Nadeem Von Haydebrand. He's. He's another data product manager, kind of evangelist for that whole mindset.

Noelle Butler [00:49:39]:
Yeah.

Nathan Settembrini [00:49:41]:
And, yeah, he's. He's big proponent of outcome over output. Yeah. So, like, what's. What is the outcome that we're driving?

Noelle Butler [00:49:51]:
Yeah.

Nathan Settembrini [00:49:51]:
Not. Not the output, but that is. That is a weird metric to be measured by because not all dashboards are created equal. I. I say that once a week, I think, because a dashboard could take 20 minutes, it could take two weeks or longer or more. Just depends. But. Yeah. Okay, so next one. What's a vanity metric that you've come across that you either secretly love or hate? Vanity metric being a metric that it makes you feel good when you look in the mirror, but it doesn't really drive the needle.

Noelle Butler [00:50:33]:
Users, customers, subscribers, whatever you want to call it, It's Just a number. What I love or hate, I just like it because I think it's something to have a pulse on. But usually it's a denominator for something else. To understand your cost, your profitability, I think it alone doesn't always tell you how your business is doing or what your outcome is. I like to use the example of I pay for cable and Internet and a phone. I don't have a landline, but when I called they said it's cheaper if you do the triple package. All right, yeah, it's cheaper than me just doing WI FI or cable. I, of course I'm going to do that. Why wouldn't I? I love to save money and then. But what's really interesting about that, that people don't really think about is why are they doing that? Well, now they can count you as not just one customer for WI fi, one customer for a cable, you're a third customer. You're doing three products that they are selling. So that's why I think it's about any metric. I think it's just something to try and like praise or celebrate, but it's not really telling you what it means or what the business is doing.

Nathan Settembrini [00:51:53]:
Yeah, that's funny. All right, so how about a more important metric? So what's your North Star metric? The one you'll never stop tracking?

Noelle Butler [00:52:04]:
I know, I hate to be boring, but profitability is just such a big one because there's revenue, there's, that's like the finance person in me. I know, I know it's boring, but I think profitability talks, tells you a lot about the business. And I think that talking about the last one, the vanity metric, like you could have users grow, but your profitability stays flat because you made a big investment and you had a big expense that year and you're not making as much money as you did the year before with less users. So I think profitability is just huge. It's really only nothing else. Shouldn't say nothing else matters because I came from healthcare and we cared about our patients outcomes.

Nathan Settembrini [00:52:46]:
Right.

Noelle Butler [00:52:46]:
But the business can't survive if there's not profitability. We can't pay our employees, we can't pay the people that are out there doing the work if it's like a service type job. So I don't know, I just always feel like profitability is important and everything else can maybe distort what is really going on in the business.

Nathan Settembrini [00:53:09]:
Yeah, yeah. I think profitability is interesting because it is one of those metrics that has two Sides of it that kind of lead you towards the right direction. It's like, are we creating value revenue in a way that's sustainable costs? Right. And yeah, it's a good one. It's classic.

Noelle Butler [00:53:37]:
Classic.

Nathan Settembrini [00:53:38]:
All right, so let's, as we kind of move into our last section here, to land the plane. Yeah, I need another metaphor for that, by the way. I say land the plane every time. But, yeah, as we, you know, enter our final chapter, our final case file of our conversation today. Okay, so AI, obviously a big deal. Some people, especially in, like, the data community, some people are like, I've. I've seen clients basically kicking the can, thinking AI will solve everything for them someday. I've seen people spend tons of money and get no value from it. You know, I'm talking about, like, large deployment type stuff. Like, yeah, obviously ChatGPT is my best friend, but it was my IT guy this past week that was really helpful. But, like, what are your thoughts on, like, on a bigger scale, like, the impact of AI on the data? Like, the people like us who live, work, breathe in the data community.

Noelle Butler [00:54:48]:
Yeah. My take on AI is usually a little bit more positive. I think something that's inherent in me is my adaptability for change. And I'm very flexible. Anytime my manager has changed, the organization has changed, the way we do, things have changed. I just have this attitude of, all right, well, it's changing. Just learn it and teach and use it and move on. You just got to adapt. So I am very lucky that that's a personality trait of mine. It's not for everyone. It changes really hard for some people. But I think it's. No matter if it was AI, it could have been like a new technology. For example, we use tableau. There's Power BI out there. There's other systems out there. What if the organization was all of a sudden like, hey, we're not going to use Tableau anymore. Everything's moving to Power BI because they're using their money to invest there. Well, that's what you got to do, right? So I feel like no matter what the change is, you just got to think about it on that scale with AI, like, just going to be a new tool in your tool belt. You need to learn it, and if you don't, you may get left behind, especially in a technical field. So it's just. Don't be scared about it and just learn. Listen to podcasts, because there's so many different points of view, whether it's extreme peril, like our world is going to not be here in five years or really positive, like it's gonna solve world hunger, cure cancer, so on and so forth. So I feel like my biggest takeaways are just change and adapt. And the one, only one example I'll kind of leave with is something I find. My mom's a nurse, and she. Well, she's retired, but she. I remember her saying. And this stuck with me, and maybe it's why I have this point of view, is she went from being in an operating room for years to then teaching and training and actually going into data. So she started doing data quality metrics really to do it. But she was out of the operating room by the time she retired for, like, 14 years. She's like, I keep my license just so I have it. She does a certification every other year. But she's like, I couldn't go in that operating room. It's not just the technology. It's the technique that changes, it's the guidelines that change, it's the legalities that change. Like, number of people that should be in there, what instruments they use. So I think no matter where you are, there's going to be change. And we just got to treat AI like a new tool and just change that we got to get along with and we got to move, you know, use it to our advantage, too.

Nathan Settembrini [00:57:28]:
Yeah. Awesome. Yeah. I love a good, positive outlook.

Noelle Butler [00:57:33]:
Yeah.

Nathan Settembrini [00:57:34]:
So thank you for that. You know, I agree, and I feel similarly. Yeah. I'm excited to see what it does for us. If it means that we have to make fewer clicks to build something that's beautiful and functional and, you know, impactful for the business, then, yeah, you know, I'm all for it.

Noelle Butler [00:57:57]:
Great.

Nathan Settembrini [00:57:58]:
But, yeah. So, Noel, how should folks connect with you if they want to reach out to either learn more from you or, you know, just connect?

Noelle Butler [00:58:08]:
Yeah. I'm on LinkedIn, so you can look for me, Noelle Butler, and My company is JPMorgan Chase, so that's usually in my tagline.

Nathan Settembrini [00:58:18]:
Okay, cool. Well, thank you so much for your time and sharing your experiences with everybody. And, yeah, for all of our data sleuths out there, keep searching. The answer is out there. You just got to keep digging. All right, that is the pod.