Data empathy; you either got it or you don’t

Speaker 1: This is Catalog & Cocktails, presented by data.world.

Tim Gasper: Hello, everyone. Welcome. It's time for Catalog & Cocktails presented by data.world, the enterprise data catalog for the modern data stack. This is coming to you live from Austin, Texas. We love to talk about data. It's the honest, no- BS conversation about enterprise data management. I'm Tim Gasper, longtime data guy, product nerd at data.world, joined by Juan Sequeda.

Juan Sequeda: Hey, Tim. I'm Juan Sequeda, principal scientist at data.world, and as always, it's a pleasure. It's Wednesday, middle of the week, towards end of the day. We're going to take a break and chat about data. Our guest today who's here is a person who we've been interacting on Twitter for a while, and then realized," Oh, we're both in Austin." When we met for the first time at coffee at FlightPath it was like," Wow, we should have met a long time ago." Just that one hour that we talked was so much. That's Laura Ellis who's a VP of data engineering and platinum analytics at Rapid7. Laura, how are you doing?

Laura Ellis: I'm great. I'm here. We're chatting about data. We've got some great drinks.

Juan Sequeda: Yeah.

Laura Ellis: I mean, life can't get better.

Tim Gasper: Excited to have you here.

Laura Ellis: Thank you.

Tim Gasper: It's great to meet you in person, to have cocktails in person. That's so great.

Laura Ellis: Thank you. Yeah.

Juan Sequeda: Also, Laura was a speaker with her team our data.world summit, which was a really, really cool conversation about data teams. Definitely, check that out, but hey, tell and toast, so what are we drinking and what are we toasting for? Laura, you were the one who organized our drinks here today.

Laura Ellis: Okay. Okay.

Tim Gasper: For those watching, you get to take a look.

Laura Ellis: Aperol Spritz, yes.

Juan Sequeda: Some good old... Unfortunately, we don't have nice, fancy wine glasses, but-

Tim Gasper: No, we don't have fancy glasses but it's very delicious.

Laura Ellis: It is delicious.

Tim Gasper: Why is this so exciting? Why do you like this drink?

Laura Ellis: Okay. I started to like this drink when my dear friend, Blake McGregor, of IBM a long time ago, we went to a conference and saw so many different speakers and it was awesome, but we needed a break and we did ditch the conference for the day, and then, just go out and get Aperol Spritz.

Tim Gasper: Nice.

Laura Ellis: It's just always had a soft spot in my heart since then.

Juan Sequeda: I like it. What are we toasting for then?

Tim Gasper: To data, to great conversations about data?

Laura Ellis: Yeah. Getting data in the hands of the people.

Tim Gasper: Oh, that's a good one.

Juan Sequeda: Getting data in the hands of the people. I like that.

Tim Gasper: That's our topic today.

Juan Sequeda: Cheers. Cheers. Cheers. We have our warmup, funny question of the day, which is, what is the most controversial hill you'd die on?

Laura Ellis: Okay. This is a big hill for me. Documentation is fun. Not a joke. I love to write documentation, I love to read documentation. I think it is the key to unlocking a lot of power, and yes, documentation is fun and everybody should do it to help your users, and it helps your career.

Juan Sequeda: That's a nice one. It reminds me one of the first episodes we ever had on Catalog & Cocktails over two years ago. We were just figuring out what this podcast was going to be was one of my good friends and actually a colleague is Will Briggs, which is exactly on documentation, which is like nobody really loves to go documentation stuff. But I think he followed your same vibe here. It's like," Hey, we really need to go do this. It's very valuable." That's a good one not many people will agree with that. They're like," Ah, screw that." Something we need to go talk about.

Tim Gasper: That's good. That's a good hill to die on.

Juan Sequeda: How about you?

Tim Gasper: For me, when this first question first came up five minutes ago, I was like," Okay, so not a political hill to die on." I'm not going to say anything about that so what I'm going to say is I love product management and so my hill to die on is that product, product management is the best job in the world. That's the hill I'll die on.

Laura Ellis: Okay.

Juan Sequeda: I'm going to go completely different angle. Seinfeld is the best comedy show ever. Period. I will really go off on and ... I am.

Tim Gasper: You are a Seinfeld lover.

Juan Sequeda: I am a huge Seinfeld fanatic and by the way, if you're in Austin, he's coming to Austin in March. I think tickets go on sale on Friday. I'm really excited for that.

Laura Ellis: Wow.

Tim Gasper: That's probably going to go quick, right?

Laura Ellis: Very cool.

Juan Sequeda: All right, well let's just kick this off. All right, Laura, honest, no BS. Is data really getting into the hands of the people who need it?

Laura Ellis: Is data getting... I think we want it to get into the hands of the people who need it and it is getting into a lot of people's hands, but I think that it is a little bit more difficult than is conveyed to get it to everybody. Because I would argue that the people who need it is everybody who has a decision to make with their job we know that that's not happening. So even though we may never get to a hundred percent, my personal mission is to just keep pushing and pushing and pushing to keep closing that gaps we get data into the hands of the people who need it, which is everybody who needs to make a decision.

Tim Gasper: This is a very important point. You just said every decision needs to be made with data?

Laura Ellis: I think... Okay. Okay. Okay.

Tim Gasper: Well, I mean, is this a hill to die on?

Laura Ellis: Okay. No, I died, I will die on the documentation hill. But no, not every single decision, you're right, needs to be made with data I think decisions, I do think decisions that are important, you should have the ability to at least make them or make a decision whether or not to what degree you want to inform that decision with data. Not make it just based on data, but inform it with that extra piece of information

Tim Gasper: That makes sense. Why is this a struggle? Why are companies struggling to be able to accomplish this? Because I feel like we've been talking about this for ages and we have to get data in the hands of people, we got to be data driven. We've had various generations of data platforms to help people make data more self- service, more understandable, yet we're still talking about how do we get data in hands of the people. What are some of the biggest problems that you're seeing that especially a lot of the folks that you're working with collaborating with that you all are really focused on?

Laura Ellis: Yeah, I mean think it's hard for many, many reasons. But working with data can be hard some people may disagree, but I think generally it is one of the trickier things to do to work with data because you have to understand how to take a business problem, break it down into a data problem, you have to understand some basic analytical techniques. I'm not talking anything fancy but basic then after you have that, you also have to understand the data and how it's structured and what each individual row or column means. And you also have to know how to interact with the tools that it's available in. Even coming into a new company, I'm a longtime data person, it's not like I just went in and was like," Let me in there, I now can do everything." No, because I needed to understand what data was available and what the nuances of ... are and what it means. There's a lot of different aspects to it that you have to consider I think just throwing it in a self- service tool, it could seem like," Okay. If we build it, they should come," but there's a lot to consider.

Juan Sequeda: You just said something. It was very, very subtle, very important. You to take a business problem and break it down into a data problem and I think we overlook that because first of all, a lot of the data folks, the data teams, they know how to deal with data problems. But There is this gap which they're disconnected from the business and then the business folks maybe are not able to go translate to the data problem, the data problem, they don't get that business context. I think this for me is that critical gap right there. How are you seeing that? This, again, we talk about this all the time, I think it's been a topic every single episode now about bridging this gap between the business and the data what's your perspective about bridging this gap between the business side and the data tech side?

Laura Ellis: I actually do have a blog on breaking down a business problem to a data problem. The reason I mention that is because when I would train up analysts in the past, I try to teach them that it's no different than when you learned how to break down a word problem when you were a kid. When you did all your math, the numbers, and then all of a sudden you get a word problem and you're like,"What are you talking about apples and selling things at the market? What is this?" But you have to find the objects that you're talking about. You have to find the actions. So actually work would work with my team. If you imagine you're an analyst, you get an email probably from somebody 10 levels higher than you saying," I need this," and it's just got all this information that you don't necessarily need how to actually parse through that to get all the constraints that you would want for a data problem. What are the objects? What's the time period? What kind of calculation are we doing? So teaching them how to underline these things, circle the different elements and then start applying an analysis. I think it's easier said than done.

Tim Gasper: That's interesting. I haven't heard it articulated this way, but I like your analogy there. I have some small kids at home and so they're trying to figure out how to deal with word problems right now. In a much the same way, they're learning how to piece out what are the entities here, what are the quantities, what are the relationships between these things? Because that impacts then how you're going to solve the problem. What tool you're going to use, what math tool you've been using are learning that's going to help you to solve that problem and maybe that analogy actually does apply really well to then now that those entities and what it is that you're trying to understand what the question is." Oh okay, so this is what I need," then you can map that to what data's available? Where is it available? All that, right?

Laura Ellis: Right. You just break it down, make it less intimidating, right?

Juan Sequeda: You can see my smile right now. This is you're breaking down the problem in a way that it sounds so simple, which I think it actually is. It's listen, understand, go read, go make sure you understand what people are saying, go write it down, go find within those sentences what are the most important words, what are the most important actions, and try to understand that if you can't go ask why again. I think this is such a critical thing. I mean we do this as kids and then we come up as adults and we're working this data and we're completely disconnected. I mean I remember one of our episodes with, I think it was Parmer, he's like," We just miss a lot of the stuff that we have with kids. That curiosity, well asking why." I mean I'm glad you're bringing this up because it's...

Laura Ellis: And the ability that you're allowed to go ask, seek clarification. But gain, if you're a junior analyst and you're getting an email from somebody who's many levels above you, you might think like," Oh my gosh, I got to go back with the perfect answer," but we have to make an environment where it's okay and expected and better if you're asking clarification.

Tim Gasper: Yeah, agreed on that. I think this is an interesting way to think about it. I do think that maybe there's a gap though around, and I know a topic that's come up as we've chatted together is around the user experience around the data. So as you think about people breaking down the problem, turning a business problem into a data problem, now they're looking for the data, they're trying to solve their data problem, there's still a bunch of challenges around that. A lot of that revolves around the user experience. Can you talk about what does it mean for there to be a user experience problem around data?

Laura Ellis: Yeah, I think a lot of companies have multiple tools. You've got a data warehouse or a data lake house, your main data store, you've probably got some BI, you might have different levels of BI and self- service tools and you might have different data in there and you think," Yeah, we've got a lot of data and it's in a lot of different places so you can just go use it." But we have to remember that these people, working with data is typically not... So I'm talking if we're getting to all the people, their primary job and they might have half an hour, one hour to dedicate to solving a problem or figuring out some numbers around the problem that they're trying to solve. So we have to make it super easy for them to be able to locate the data, understand it, get access to it, and then do some analysis. Therein, lies great data structure. We've got easy data access, we've got data documentation in that, we've got data availability in the tool, we've got fit for purpose tools, content. Therein lies a lot of different pieces of the puzzle that come together to make that work.

Juan Sequeda: When it comes to understanding that user experience of the data and figuring out how these pieces of the puzzle come in. Who is responsible for figuring out what those pieces of the puzzles are and how they should fit together? In your experience-

Tim Gasper: There's a lot of work that's implied there. Regardless of whether you think documentation is fun or not. Who owns that work, who's doing that work that thing.

Juan Sequeda: Who owns the work of documentation? Who owns the work of talking to the users to figure out what is the type of tools that they want so forth?

Laura Ellis: That is a tricky question. I'll say this, personally, as the leader of data engineering platform analytics what that means is the data, the actual data, analytical data and the tooling. I would say it would be myself and my team, we are responsible if we are custodians of that platform, we are responsible for making a good user experience in that platform. But that said, the best case scenario is when you enable the entire ecosystem that uses it, that they also feel responsible. The domains feel responsible for their data as well. The different users, everybody feels like invested in this platform. But I say that not to shirk responsibility, I do think that the people who own the tools and the pipelines are responsible for the data quality and the user experience.

Tim Gasper: That's interesting. Can you go a little bit into, because I know that you're especially involved around the data engineering and platform aspect and then you sort mentioned about stewards and people in domains and things like that. Can you talk a little bit more about what your team and your focus is doing versus some of the other parts of the organization that you think about or in some of the clients that you're working with?

Juan Sequeda: I'd like to add is understand how you're balancing what we talk a lot, the centralization, decentralization. What gets centralized, what responsibility are centralized, what gets pushed out of domains and how is that being managed?

Laura Ellis: Yeah, I should say too, this is just a model that works for us. Right, so everybody's going to have a different model if it works. I think honestly, if you have a situation where you have somebody who feels strong ownership, you are lucky. It doesn't matter what their title is,.

Tim Gasper: Embrace ownership where it comes from.

Laura Ellis: Embrace, empower them, build them up, that kind of thing. Yeah my team, we own or we are custodians of the analytical data and the analytical tooling and-

Juan Sequeda: Sorry, you are a centralized team within your...

Laura Ellis: We are a centralized team, yes. But that said, so that is my team, but we have different domains that have expertise up and down the stack. So analytical expertise, state expertise, state pipeline expertise you may remember in our talk where we were talking about how we were in ticket jail, we were in ticket jail because we were essentially a bottleneck. Just the way that the team had grown over time. We realized nobody wants that and also the ecosystem of domains around us are capable. So when you look and you're not... First of all, bottlenecking groups and they're capable, it's like," Why are we doing this? Something needs to change." So we're shifting now towards that whole data mesh architecture where we can have these strong domains fully empowered to take our core models, work with some of our main infrastructure like learning, monitoring, plug into that. Our documentation, our support, et cetera, but fully plug into that and be able to make their own models, publish them till the end, own them, and then you don't have it in our reporting tools. That's what we're working towards if you check out the talk our team was talking about, we're still learning and evolving together with our domain teams.

Juan Sequeda: Got it. I think I'm really excited about hearing what you were just saying, which validates a lot of different approaches that we're also hearing, just talking to other folks is yeah, you're trying to find that balance between what needs to be centralized, what needs to be centralized and from your central team also depends on the industry, the size of organization, the culture you're trying to go change, trying to head towards. You define some model of how to go do things and you help and you want other people to go replicate that, but you also want to make sure that that's as smooth as possible. That's where that user experience comes in through. To make sure that they can," You can go do this too and we want to make this easier for you to go do," and you've defined what that should look like.

Laura Ellis: Yeah, I mean we do say, so I have an amazing, fabulous team, so great. Very, very lucky about that. We do say if we have capable people asking us to do something and they can do it, but they can't do it because our system doesn't allow it, that's a problem we have to solve.

Juan Sequeda: That's a good point. That's a really good point. If somebody really wants to go do something or take that ownership and there's a couple of issues there. Well, let's solve those issues so we can enable them to go take the ownership and go because they have the capabilities and desire to go do that. One of the things we were talking about earlier before was on how identifying all these issues and how do we make this better because we want to get data into the people, into the hands of the people who actually need it. Why isn't that happening? I'd like to go into what are the issues or opportunities that you see that we should be able to dive in to say," Oh, this is a way to figure out what those issues are and how to go address them?"

Laura Ellis: Yeah, there's a few way. There's obviously many ways for almost every question, but what we did was a few different things one is we actually is really amazing. I consider myself very, very lucky. We were able to work with an internal user researcher to go out and speak with a cross functional set of our users, our internal analytical users and find out what's working, what's not. Which is really awesome because usually user researchers are reserved for external products. Not only did we get to work with this user researcher, but this person is like, you know how we're crazy about data, they're crazy about internal user research. Awesome person. They went off and did a bunch of research and then additionally, myself and my team, we do feel very connected to our user base so anytime that I would see a presentation go out where it has numbers, I'll message that person and say," Where did you get those numbers? Who pulled those numbers?" And then make up a meeting with that person to go find out, okay, what assets did you use to get this? How was your experience? If it was awesome, let's harden that path and publicize it so that more people can have your success. If it was hard, let's think about how we might be able to solve this.

Tim Gasper: That's interesting. This really connects to my experiences coming from more of the product management side around user research and customer search, customer reviews and things like that. Obviously from being at a product company, we spend a lot of time talking to customers, talking to stakeholders, users, things like that. This is the data version of that. It's like can we talk to our internal customers or internal users? And I imagine use a lot of the same disciplines that you see in this product oriented research being done around these data experiences. I'm curious, do you even internally think of this as data products and we're doing of research around these data products? Or do you not quite go so far into the data product realm and," Hey, these are just data sets, these are whatever."

Laura Ellis: The fact that we're trying to get into the data mesh, so we would absolutely love to start going towards the data products. And we talk about it and we're like," This could be a data product, which would be very much in line with the data mesh philosophy." But that said, I don't want to over- represent, we're still in our early days of that data mesh. We're just sort exploring there I will say the internal user researcher absolutely used very disciplined techniques. Me, I'm just going out and talking to people.

Juan Sequeda: Okay, so I really want to dive into this internal user researcher because this is something I've talked about this as the data therapist and people going off, but you've been already doing this who is this person? Did they already exist? You hired them.

Tim Gasper: How did they get that role?

Juan Sequeda: How they get that role?

Laura Ellis: I know.

Juan Sequeda: Yeah, tell me more. Who is this person?

Laura Ellis: I should have invited him here. I don't know. He's in the IT team and they of course own all the enterprise applications for the company and as we do migrations from ticketing systems or whatever, expense systems, et cetera. They actually had a guy who was working on It and got so interested in this process that he went and studied user research to have that internal lens on user research with the focus of just trying to make... At Rapid7, we call our employees moose because it's the same word, both singular and plural. Trying to make moose life better is this guy's goal. He's awesome, he will talk to, again, just like we could talk probably for five hours about data, maybe everybody else would leave the party, but we would still be here talking. He is like that for internal user research.

Juan Sequeda: This person already worked in IT and he just happened to get really interested in how the business works and understand the processes and then educate himself just on user research.

Tim Gasper: And just shift it into that role over time?

Laura Ellis: I believe so. I'll say yes, and Thomas, if that's not your story, I'm sorry, but yeah, that's what I remember him telling me, which I thought was really cool because it's really cool when people go through an authentic journey to find the stuff they're into.

Tim Gasper: Yeah, absolutely.

Juan Sequeda: Now, I'm so excited to hear this because it's validating a lot of the things that we've been talking about on," Hey, this whole data therapist where we talking like,'Let's go understand what are your problems and how does stuff work and how are you doing it today? How could it be better? What's working, what's not working?'" Then it's this other topic on what we've been calling business literacy. It's like not only should we strive for having data literacy within an organization, but we should have all the data IT tech folks also understand how the business works and understand what does the flow of... How does this company make money? You pour money here, how does that go flow and everything and order the processes, who are the people involved and all the systems? Once you get that context, then you can actually ask more intelligent questions, right? This is really validating that it's happening. I think we need to harden this structure. I think we need to... And I don't think there's many people doing this. I don't know. Have you brought this up and seen other, your colleagues in different organizations? Are there other internal user researchers in other companies do you know?

Laura Ellis: I haven't seen other internal user researchers. But yeah, I think there should be more like you said, I think there should be more formalized education when you're onboarding to a company about this is how the company works, these are the basic systems, these are how things flow. And then it can be specialized to your area. If you're working on data, you should understand the financial systems and how they go together and basics around that so you understand when things come in and the different analysis cycles but the same would be for products. You should sort of know how a business functions. If you have a product, how does it get provisioned? How does sales work? Is it licensing based? Is it like all you can eat, license is a consumption based? How do you onboard customers? That stuff.

Juan Sequeda: What does the data people think about this more? It puzzles me. It is so annoying. I had to say, it's so annoying... My Twitter, my last Twitter feed discussion I'm in is like," Oh we need more technology and more tools." I'm like," No, stop it and I'm out of this room," I can't even deal with it anymore.

Tim Gasper: I feel and I'm curious about your opinion on this is that I think people are starting to think in this direction. But I feel like whereas the software world, I know we like to always make our software and data analogies here. The software world really embrace, sort of agile and lean and customer development and things like that. And so," Oh let's meet with our stakeholders. Let's meet..." And then you've got these really important roles that emerge. A product manager, a user experience designer, which is usually an external facing user experience designer and things like that. Where those things don't really exist in the data world so in the data world, you have your data scientists and your data engineers and even your data managers are often very technical and very platform oriented and things like that. So that isn't the skill set they really learned or focused on, I don't know what that means. Sometimes we're like," Oh we need new roles..." Or do we just need to change our hat out a little bit and build some new skills?

Laura Ellis: I'd be interested to see what you guys think but I think... Because I've thought about this a lot, I think one of the issues is that software development, you're making a product that you can sell. So if you're a data team and you don't sell your data, you often aren't afforded to have some of the extra... It's not a luxury because it should be a necessity some of the extra bells and whistles that a software development team that's selling software would get. So that user research, that product manager, that content writer...

Tim Gasper: Dedicated documentation people, things like that.

Laura Ellis: Can you imagine the support engineers that we had those on an internal data platform like," Yes, please. I would love this."

Juan Sequeda: If we are actually putting data into the hands of the people who need it and they're making those decisions that are making money, saving money, if we're doing that correctly, then we should be able to argue that. Should we, well... Shouldn't we say," Look, here's this data. Here's this data you're using." I'm like," Yes, I love this data." I'm going to turn it off tomorrow. You won't get anymore. If they're not complaining saying," Heck, no, I need that." Then you're not that essential, then something's missing. It shouldn't be a luxury you're saying if we're truly delivering the value of data, then it should be on par with everything else or I am I being too...?

Laura Ellis: So you're right, it's just harder that's not an excuse to stop for example, software products, they can say, We made x million of dollars last year." What we can say," We cost the company this much money." Now, we can say," With this many users you can say this..." And if we're amazing and if we are able to get out of ticket jail or actually start doing more and more proactive data analytics, data science, we can start showing the business cases that we drove this uplift in more advanced marketing campaigns or actually baking in data science to our product. That is such a heavy lift to get to that point because you are, first of all, we have to get to the basics of quality data everything with our minimal staffing. But then we do this and you can't own the product. I can't just go to marketing and be like," Use these campaigns, I'm turning off your other campaigns. These ones are more performant," or," I'm putting a data model in your product." Anyways, I'm not trying to make excuses so you're right, we should be able to have this conversation but... Or somebody taught me instead of saying but, we should have these conversations and it's a little bit, it's a barrier to get there, but we got to keep pushing.

Tim Gasper: Right. Now that's interesting how you're thinking about all of this and I think this connects to a topic that's come up a few times on our show where we've been talking about the ROI or the impact of data teams. Especially when you're internal facing or you're not building out data products that you're monetizing ultimately. It can be difficult to actually calculate and spell out the ROI. There's no formalized sort of data accounting that you can really do around like," Oh, what were the decisions that our team made possible that resulted in us making more money?" If that report or that dashboard wasn't available, how much would a team be willing to pay to make that dashboard reappear? That's how valuable it is for this company. There's this idea of data accounting which doesn't exist, which maybe it should, at least in theory, probably literally.

Juan Sequeda: I think this is the change, this is the paradigm shift when we start thinking about we were not talking about data accounting, we should consider what that even looks like. I don't even know what that means. We should have that discussion and say," Does this even make sense with what does this look like? Let's hypothesize about this stuff." These are the discussions that we should be having more publicly when we go off and conference and stuff," Let's have this." I'm done with the technology. Of course we need technology, but it's like-

Tim Gasper: I'm done with it.

Juan Sequeda: But we need to really elevate the discussion. I'm really loving this whole data accounting. And another topic that's come up and it is something with our colleagues, with Emily Pick... here at data.world, we've been talking about this data marketing. It's like the same way we're marketing a product. I always bring up, I love my Yeti water bottle and there's a reason why I bought; Yeti he has cool marketing. I buy this. Why aren't we marketing the data that we're releasing? Because if we're supposed to go promote it, we want people to go use it. Let's go really market and promote it. Not just like," Oh, data's out there, it's on a catalog, let's just people will find it."

Laura Ellis: I will say that is something that I'm also passionate about. It is in line with documentation, is communicating the value of what your team does. Our team went through this transformation recently and we're still going through it where we're trying to have this unified experience across tools, across data sets. Our data users are our data users. They don't care what the technology, they're just like you. They don't want to talk about technology, they don't care what the technology is. They don't care about us if we give them this whole scavenger hunt that they can go on to find the data. Nobody wants that. We went ahead on this user experience project and one of the things that we got, which was very cool, was our marketing team actually made us a little logo we're the moose they made us a data moose, which was pretty cool. We're making a little bit of swag when we're incentivizing people to go through our training so we have trainings at different levels of our tools, like fewer training explorer training, admin training for our data mesh users that are able to act more autonomously in our systems. And then the other thing is we launched two other things. One is a blog, which I know a lot of internal teams have a blog, but what we try to do is if something big at the company comes out, we have a new product, we have new packages, whatever, then we'll put together a blog that talks about this is what's happening at the business level, teaching the business side. Here are some of the main data elements that this amounts to and here's where you can find it in our data lake house and our reporting solution and our self- serve and our friend solutions too, which would be our enterprise applications. Because even though we didn't make those reports, we know that they're in this sphere, they're in the conversation and we want to be able to navigate our people. We're trying on the marketing side, we also have a newsletter by the way, which is fun. We don't just write about stuff that our team does. We write about what's happening in the data ecosystem at our company. I read it on a Saturday morning, I'm like, What's the haps? What's going on with data at our company?" Maybe not everybody is like, "What's going on with the data moves?"

Tim Gasper: Who puts that together? Is that something that you're putting together or somebody on the team's doing or...?

Laura Ellis: I'll say, so we have an amazing technical program manager who is, so she's a program manager, but she's also has been a data scientist and data generalist so she keeps us all organized and then the different leaders of my team, we all put content into it and then gather content from the community it works because everybody on my team is great and is also very passionate about this thing. Everyone is excited about the newsletter, nobody's annoyed. At least they don't tell me that.

Tim Gasper: Well, this is fun. I know you said that you're not really doing data products in a formal way, but when you say that you're doing data marketing, you're doing data user experience, you're thinking about the data mesh, I'm like you're dangerously close to do a data product stuff.

Juan Sequeda: Not even close. I think you're even going way beyond what other folks are doing. I actually see and I go talk to the people who are," Oh, we're doing this data mesh with God and we've hired all these companies to go do it." I'm like," None of them are talking what you're saying." I think this is a diversity of thought that we need to have. I'm really excited about this idea of the data marketing. I mean this, we've said in talks before, when you go off and you search for data in your catalog, it should be just like that experience when you go buy something on your favorite e- commerce platform that's a lot of documentation you get out there that has to be there and which product are you going to buy? The one that has great documentation? One has crappy documentation? I mean, the documentation is a big selling point right there. I think this is something that we need to really start focusing more and people need to realize that there's value behind that but that value going, again, we really need to go strive to make connect that to the actual dollars. I think that if we don't, we're just going to keep having this conversation. People are like," Yeah, those are folks in the corner who are talking about that cool stuff, but hey, how's that making money?" I think success here is being able to go directly connect. Here is this data work project, dataset, product, whatever, and it's connected to this dollar amount and if we cut it off, people are going to go on flames. We need to get to that point, and I think if we don't, we're just going to keep doing the same thing over and over again and expecting different results. Einstein's definition of insanity.

Laura Ellis: Yeah. Yeah, I would agree. I mean it is daunting to go and measure your impact in a way that's not reaching, you know what I mean? If you are saying," I've enabled this product on using analytics, so therefore the lift of their product could be driven by analytics." It is tricky to measure the financial impact of an analytics team in a way that is either that is real and realistic or not just targeted to a few specific projects that you did. Projects we can measure if we're lucky enough to get to those projects. Projects we can measure but measuring the financial impact on the foundation, that's hard. Unless, it's not hard I would love to hear.

Juan Sequeda: Let's get into this brainstorming, wanted to have live brainstorming on how do we define success. Success of the data teams, your team. Why do you consider it hard?

Laura Ellis: Wow. Because I think you have to make a leap. I mean, I'm trying to think real time. How would I measure the success? Okay, this is where my mind goes. Okay, we can poke a ton of whole incident. I'm like, well could we say that employees that use data, these employees, these products drive their decisions X percent more with data and since that have seen an a Y percent lift on sales and then therefore it has on average this financial impact and then that times weekly active users. I don't know, I'm throwing things out there, but it's hard. Yeah, maybe you guys have ideas.

Tim Gasper: Well, I mean maybe just rolling with that idea you threw out there, maybe that's a survey and you're surveying the folks that are leveraging the data to create these different things of," Hey..." I mean some of it might be connecting it back to user experience, more around the user experience that they have, but maybe some of it is around," Well, what was the value of the dashboard that you created? Was it how important this was this for decision making? Was it must have? Was it nice to have?" I don't know. I'm brainstorming on top of your brainstorming here on how if you can't get super quantitative about the monetization of it, then maybe you can at least talk about the impact of it.

Juan Sequeda: Then go the user research, we should be able to go have the user research on to understand the experience and the process people are using data. We should be able to go have that type of user research to understand the value of this data. Let it be quantitative, let it be qualitative, and I think-

Tim Gasper: And that's a good point.

Juan Sequeda: I mean, look, I'm asking this out now. Everybody who's listening, are you doing surveys for example. Are you going off to your executives, to the leadership team, to the sales team? Are you going off and serving them, saying," Hey, we've delivered this data. Are you using it or not?" Or get that qualitative, have those interviews because at the end of the day, we don't know, we're not even asking these questions and I think we live in this vacuum so separate, we live in our bubbles. I mean it sounds so... I don't know what the answer is, but I think a first step is to go start talking to people and even do it a survey.

Laura Ellis: I think it is hard, and we know data platforms are important, period. Internal data platforms, we need to get data in the hands of people. Everybody makes people traditionally make better decisions when they have data access, whatever decision that they end up making. I do think if we push ourselves to try to measure the monetary aspect, we could even probably do more. If I knew that I had to measure my team's financial impact, then we'd probably be out there actually pushing more financially impacting projects.

Juan Sequeda: Because that would help you prioritize. At the end of the day, you're like," Well, I've already figured out and we realized that we're working together and then with that team, without us, they're going to go fail on their projects and it's going to effect the company financially. If for another team we don't know maybe that forces the other team who doesn't know to figure out what is going to be the financial impact. Maybe you realize I'm going to just start working more where I know that I'm giving a financial impact." I don't know maybe at the same time we're over indexing too much on the financial impact. That may be a case too, but we don't know. There's so many unknowns right now.

Laura Ellis: Yeah, and some people have done, when you request time with the data teams, if you request a data team project, you have to put in the financial or whatever metric impact, which I think is nice. It's a nice way of having stakeholders also be bought into partnering with you to actually get those metrics, right? Because we can release the data, but usually it's empowering people in a system that we don't own. Empowering them to change something in the product that will impact product sales or something as I said, marketing or whatever if we create that as something that we're asking for front, then maybe that would set us up for success.

Tim Gasper: This is interesting. I think this provides a nice framework to think about. And I like that we connected it back to the user experience research too, because in user experience research, there's value research, which is more, is this a valuable use case, what's the impact? And then there's more solution research, around more of the is this the right solution? Is the user experience correct and things like that there's a few other kinds of user experience research. I like how we took data accounting and user experience and all wrapped it together. Before we go into our lightning round today.

Juan Sequeda: Wow. All ready? So fast.

Tim Gasper: I know we're having an amazing conversation time is flying. Before we go into our lightning round, I would like to bring it back to user experience very squarely again and just understand, so we talked about, you do these interviews. You have a user researcher internally, which is awesome. What other sorts of recommendations would you give to some of our listeners and some of our audience around how can you take more of a user experience centric approach to data?

Laura Ellis: I am famous for saying the devils in the details. I implore everybody to be detailed, to make sure that you have a healthy dose of details at whatever level you're at. Because when you dive into the details, that's where you find out how things are really working. Beyond the clean architecture, diagrams, beyond the operational processes, you find out how things are really going and that can inform your strategy and your vision and also keep you honest about again, what's really happening. My advice is to dive into the details and then even more than that, hear it from the people. If you see people doing great things, struggling, even just using your product, just take a minute, grab a coffee with them virtually, ask to hear about their experience it's no better way to be motivated to fix something than to hear somebody struggling.

Juan Sequeda: This is the empathy right there. Yeah, this goes back. I think this is a theme throughout the last months and a shout out to our guests is one of our guests who brought it up, empathy and curiosity. This is what we really need. I think that user researcher is curious to understand and having that empathy that this is the critical aspect missing. Wow. Time flies. Ready? Want to head out to our lightning round?

Tim Gasper: Yeah, sure, let's do it.

Juan Sequeda: All right, so we're moving to our lightning round section, which is sponsored by data.world, the data catalog for successful cloud migration. I'll kick it off first, so should we pick someone in our extended data organization to wear that data UX researcher hat? Or should that be some outsider?

Laura Ellis: Oh gosh. I say if there's anybody who can wear that hat, hold onto them tightly and empower them to do it because it's a real luxury to have it. Take whoever you can.

Tim Gasper: I love it. I think every organization could benefit with somebody with that hat that's probably not done enough. Really somebody who's focused on that. Next question for you, so you talked about documentation and that was the hill that you would die on is the documentation can be fun. Is making documentation smooth and easy good enough? Or do you got to do more? Like you got to gamify, you got to incentivize? Is smooth and easy good enough?

Laura Ellis: Yes. Okay, so not only will I die on the hill that documentation not can be fun but is fun. Yes.

Tim Gasper: Thank you for the correction.

Laura Ellis: Sure. But also, yeah, I'm very particular about it. It has to be quick. Nobody wants to read your thesis. I know we're all proud of the new data that we brought on, but nobody wants to hear the journey that you went through. Right?

Tim Gasper: That's true. The journey is often... As a data scientist or something you want to say," Well, here's everything I did..."

Laura Ellis: You want to say it, but let's lead with the answer what we need to know then you can have all your appendixes and a little five minute video is like the chef's kiss.

Tim Gasper: Nice. For the loom fans out there, you pop up your loom and you make a little video, right?

Laura Ellis: Yeah.

Tim Gasper: Awesome.

Juan Sequeda: Can we teach data teams to be empathetic or is that just...?

Laura Ellis: It's like can you teach people to be empathetic is the question, right? I think, yes. Can you teach people to be empathetic? I think, it is easy to not feel empathetic when you're in ticket jail or you're just getting tickets and they're like not people asking you things, you're not seeing how, you're not feeling it day to day. I think if you make sure that people are faced with the reality of what's happening, good or bad because lots of times things are going awesome and you want to show them," Hey look, share a slack somebody sent you. I love this new data that you release." Screenshot that, send it to your data engineers. So yes you can help them feel it, but can you teach them to care? Well, that's like a human thing.

Juan Sequeda: That's a very good distinction right there.

Tim Gasper: I like that, all right.

Juan Sequeda: Last one, Tim.

Tim Gasper: Last question? People process technology, right?

Laura Ellis: That's right.

Tim Gasper: Is data user experience really mostly a people problem?

Laura Ellis: Oh gosh. Well, I think it's all three because it's also about the implementation, which has a lot to do with the process and the technology, but technology doesn't drive itself, neither does process. At the end of the day it usually is about the custodian and in the custodian team. Unless you don't have the money to buy appropriate technology or buy in, which is also a people problem to get the process, it's typically a people problem. My opinion.

Tim Gasper: So it's all three. Tech's a big piece, but it doesn't drive itself. The custodian team matters and buy- in matters.

Laura Ellis: Yep.

Juan Sequeda: All right. Tim, take us away with takeaways.

Tim Gasper: Oh my goodness. Takeaways. Many exciting things that we talked about here today. you had discussed that when we want to get a really good data user experience and if we want to get data in the hands of everyone, that's a big ask. It's been a big challenge for a long time. We may not get to a hundred percent, but we have to have that mission to get data into the hands of everyone. Almost every decision should be made with data. The decisions that are important need to have the ability to use the data. It's really tricky to work with data, that's like one of the main R reasons why we have these challenges today. It's because you have to do a lot of different things, one of the biggest things is you actually have to translate a business problem, as you were mentioning, into a data problem. That requires and you give this example of when children do these word problems and they have to understand," Okay, so there's five apples and there's five oranges. And there's Fran and there's Sammy and there's some exchange happening here and then they're doing something with those objects and then they cut one in half." You need to understand the problem and what the objects are and what the verbs are, this is knowledge. This is the knowledge and the understanding of it and the translation from the business to the data and probably vice versa. That's tricky. That's hard. You have to understand the rows and columns. You have to understand the tools and the technology. You have to understand the data that's available and the people that know things about this data. There's a lot of moving pieces that have to come together in order to make this a success. There's a lot of moving pieces there one of the biggest things you really highlighted and what the central points of our discussion today was around user experience of data and how important user experience is to try to tame some of those challenges and really make it so that people can take advantage of the data and the knowledge to be able to do work more effectively. You said that we have to make it very easy to find the data, understand the data, access the data, things like documentation, search, fit for purpose tools, content, context. All of this has to come together to make that work it's important who owns that work. The data engineering and platform team should take ownership to create a good experience for the platform, but then embrace ownership wherever it comes up and if you're thinking about this data mesh oriented model, then you can really think about where are the domains, where are the pockets of expertise that you can be empowering ownership around. And you mentioned that are embracing more of this data mesh approach because you were experiencing some of the bottlenecks that come with the team having to do everything right that's really hard so that was a big shift there. We see a lot of companies now shifting to, at least in inspiration by data mesh and in many cases actually implementing data mesh. You had mentioned that there are a lot of opportunities on how to address data user experience. You mentioned about having an internal user researcher and obviously a lot of companies don't have internal data researchers, so maybe you got to think more about a hat, somebody can wear that hat. If you can get somebody like that, great. It sounds like it's had a huge impact for you all and that person can come from anywhere. For you, all that person came from IT, they had a passion about user research and they focused on making the life better of people around them. Culture and empathy expressed in all of that and then they shifted into that role. I think you can also think about this as an evolution of people within your company and maybe they're wearing a certain hat today, but maybe they're going to grow into a different hat, a more of a knowledge hat, more of a user research hat. There's a lot of opportunity there. And so much more I'm going to pass the baton over to Juan. Juan, what were your big takeaways?

Juan Sequeda: Well, one of the things we did was validating a lot of the points that we brought up before about this business literacy combined with the data literacy. I think there should be some more of the formalized onboarding process of," Hey, how do all these systems work together? Where are all the business cons? Where are all the products? How do they all work together? How would we even onboard customers with an organization?" I think that's something that's missing within organizations and we need that more, help us generate that empathy of how things work. If we're treating data more like a product, maybe there is a gap here between the luxuries that the product engineering team may have. They have product managers, they already have user researchers and stuff like that. There's a gap there and why do we have that gap? Because I think commonly, we think the data in the IT,"God, that's a cost center for the organization." So something there needs to change, I think we've had this brought up, this whole data accounting, there is no formalized accounting. Maybe there should be. What does that even mean? I don't know but let's actually start that conversation. I think that's a really fascinating point that we just made there. We brought up the whole enablement and data marketing. This is, again, documentation continues to be a theme. It's all over. Documentation is key for this. You talk about how you guys created some swag and logos to encourage people to do the training. You're promoting the work that folks are doing, I think you want to celebrate those wins. You have a newsletter, right? Not just what the platform team is doing, but what's happening within the ecosystem you have an a TPM who's actually, who's doing that work right there, who's generating that newsletters. This is really cool culture that you're discussing there. Talk about how do we define success here? Let's go brainstorm into doing more than just," Ooh, okay, yeah, we have some users and so forth but how do we actually connect the data to the actual money to the financial aspects directly? How do we understand what that value is?" We should just start doing surveys. I ask everybody here," Are you even considering doing surveys?" Do we even... Because we don't even know what we don't know right now. Let's start asking people around this. And then a good point you brought up is if you are requesting time from a data team, you should be providing the financial impact of what that's going to have. I think those are small steps that we can start doing. And then additional best practices that we discussed at the end about data UX is be detailed. When you dive into the detail, that's how you figure out how things are really going, not just look at that pretty neat architecture. You got to hear it from the people. If you hear people are struggling, go talk to them. Go understand what they're struggling, this goes back always to empathy. And if I were to summarize our whole discussion today, it's user experience research for two things. One to understand the business and the processing within the organization. And two, a huge potential is to understand the true value, maybe even economic financial of that data within organization. The two words takeaway is user research.

Laura Ellis: Yes.

Juan Sequeda: All right. That was a long take away. How did we do?

Laura Ellis: I love it. Yeah. Great. This is awesome.

Juan Sequeda: All right, so we'll throw it back to you. Three questions. All right. What's your advice? Who should we invite next? What are the resources that you follow?

Laura Ellis: Right. Advice, this is life data advice. So my advice is to build people up around you. Build up your team, build up your peers, build up your peer pluses, build up your friends, your family, build people up around you, invest in them. There is more than enough sunshine for everyone. Important to know that. Giving to other people and making them feel good will only give back to you. It never takes away from you that's everyone advice.

Tim Gasper: I love that.

Laura Ellis: Second will...

Juan Sequeda: Who should we invite next?

Laura Ellis: Oh, yes. Who you should invite next is a dear friend of mine who I've talked to you about before, Caitlin Hudon, and she is a principal data scientist, longtime amazing woman in the data field. Talks about all sorts of things data, @beeonaposy at Twitter, and also has a blog as well. She is a dear, dear friend of mine in the data field, but also in real life, so...

Juan Sequeda: Awesome.

Tim Gasper: Awesome.

Laura Ellis: Yeah.

Juan Sequeda: Finally, what resource do you follow? People, blogs, podcast, books, conferences...

Tim Gasper: Also mention about your blog as well. Yes. Some folks were adding comments saying you mentioned some of your blog posts. How do they find your blog?

Laura Ellis: Okay my blog is littlemissdata. com. So you can find me on Twitter @ littlemissdata, and my blog is littlemissdata.com and I talk about all things data. Some stuff is weird, but it's all data related and I find it fun and that is just me. I'm not going to apologize.

Tim Gasper: That's cool. It's real, right?

Laura Ellis: It's real then I like to connect with people. I'm more of a people learning person, so these conversations, community stuff, I like to be involved in communities. I have two favorite... Two of my favorite communities will be a shout out. One is Our Ladies. That is where are my first taste of user communities. It's also where I met my dear friend, Caitlin. So I love Our Ladies that was founded by Gabriela de Queiroz, who also founded AI Inclusive. And then also, this is for the ladies out there, both of these women in analytics, Women in Analytics community is a hybrid community in person and online, founded by Rehgan Avon, who's also a friend of mine. And then just so any communities like this, and then talking to the people in person, on Twitter, we like to talk on Twitter. And then I find from there, you know about events going on. Once you connect with people, it's like a spiders web.

Juan Sequeda: Also shadow, you organized an interesting conference that you brought up a while ago.

Laura Ellis: Yes, yes. Actually-

Juan Sequeda: Talk about this.

Laura Ellis: Yes, Caitlin's going to laugh that I'm bringing her up so many times. That's very classic, our friendship. Caitlin and I run an event called Data Mishaps Night, which is super fun, if you like data, it's super fun. Every February, this will be our third one coming up this February, where people share their data mistakes. And last year, we had 13 speakers and it was like... I think we had maybe 500 or 700 people online. It was awesome. It's cathartic because you hear crazy mistakes that we've all made. It's funny, but it's also empowering that we all make mistakes.

Tim Gasper: I love that. We have to check that out because I feel like every conference you go to, every community, part of it's always the perfect architecture great. The perfect technology. We made a billion dollars, right? It's like," Oh, what about the times where we screw up?" And that happens more than when we expect, right? And that's the road to iterating to getting better.

Laura Ellis: Yep. Yep. It's a good time. You guys should join.

Juan Sequeda: Join. All right, well, next week we have Loris Marini, who is from the podcast Discovering Data. We're really excited. We were on his podcast. He'll be on our podcast. That's going to be a great show. With that, Laura, thank you so much. This was a phenomenal conversation because we ended up in a place which I did not expect that we had. We do a little prep, we know where things are going to go, but we ended up with this whole user research, and I'm super excited. I hope everybody who's listening is thinking right now How do I get user researcher internally? Not just external, but internal for data. With that, Laura, cheers.

Laura Ellis: Cheers.

Juan Sequeda: Thank you.

Speaker 1: This is Catalog & Cocktails. A special thanks to data.world for supporting the show. Karli Burghoff for producing. John and Brian Jacob for the show music. And thank you to the entire Catalog & Cocktails fan base. Don't forget to subscribe, rate, and review. Wherever you listen to your podcast.