The Evolution of Data Science into Business Influence with Expert Lindsay Pettingill

Discover the unique pathways that can lead to a thriving career in data science as Lindsay Pettingill, PhD, and Director of Data Science at Replit, joins us for a riveting conversation. Lindsay's journey—from a Fulbright scholar teaching in Germany to shaping Airbnb's hyper-growth period—is a testament to the value of curiosity and an analytical mindset. Her insights promise to guide and inspire, whether you're a seasoned professional or just starting out in the data world.

We get to the heart of what truly powers data science: curiosity and intuition. Lindsay advocates for a holistic approach to data, stressing the need for professionals to develop insights and propose business hypotheses. This fascinating discussion also covers how personal traits, such as proactivity shown through cold outreach, are becoming indispensable as technical tasks undergo automation. Lindsay's experiences underscore the evolution of the data scientist's role, from crunching numbers to being a strategic business influencer.

Finishing on a high note, our episode focuses on the delicate dance of data-driven decision-making within organizations. Lindsay reflects on her leadership experiences, particularly during the fast growth times at Airbnb, and shares her evolved perspective on leadership in product strategy. She emphasizes the need for data teams to contribute meaningfully to their organizations, beyond technical expertise, and offers insights on how to empower data professionals to make a significant business impact. Tune in to gain valuable lessons from a leader who has successfully navigated the waves of change in the data science industry.

Follow Lindsay on:
Twitter: @iam_lpettingill
Website: https://lindsaypettingill.com/ 

What's New In Data is a data thought leadership series hosted by John Kutay who leads data and products at Striim. What's New In Data hosts industry practitioners to discuss latest trends, common patterns for real world data patterns, and analytics success stories.

Transcript

Transcribed by https: 0:05

otter. ai Hello, everybody. Thank you for tuning into today's episode of What's New in Data. I'm super excited about our guests today. We have Lindsay Pettingill, PhD, Director of Data Science at Replit, early data scientist during the hyper growth phase of Airbnb, and she also founded a company as well. Lindsey, how are you doing today? I'm doing great. Thanks so much for having me. Yeah. Lindsay, I was super excited about having you on this episode. I've been a follower of your blog and really agree with your perspectives on data science and generally about, you know, how to look, think about. Data generation and testing and things along those lines in a way that provides value to the business. But first tell the listeners a bit about yourself. Sure. Yeah. Oh, it's always a question of where to begin. I, I like to say let's see, I went to a small liberal arts school in Maine called Bowdoin college. My parents didn't go to college, which provided a lot of Motivation for me to, you know, go to college and change things. I studied German and sociology, you know, I mentioned that mostly because I mean, data people, I think usually have a range of, of backgrounds and it's always, I think, good to, to remind folks that we don't all need to study, you know, CS or, You know, the hard, the hard sciences. So yeah, I studied German and sociology, had a lot of fun in college. Afterwards, I did a Fulbright in Germany, which was a fantastic way to spend a year after college. I just I did a teaching Fulbright, which meant I taught at a German school and I could only work by law 15 hours a week and, you know, it was quite a privilege. Yeah. Yeah. It was quite, you know, it was awesome to be 23 years old and have a lot of time to, you know, explore and read and read a lot that year. After that, I worked at Deutsche Bank actually for a little bit and really got my first exposure to, to sort of the quant side of, of, of life. I mean, I'd always been super into math. I went to math camp as a kid and, that was really dorky and fun, but at, at Deutsche Bank. Use a lot of Excel and, and it was, it was a long time ago, but, you know, it was really the first quant job that I, that I had and, worked at Harvard for a bit after that, mostly because, If you sort of have an entry level job at Harvard, you get, you get to take classes at Harvard relatively cheaply, if not for free. And that was, that was super important to me as I thought about what I wanted to do for grad school. And so did that for a couple of years, ended up getting a PhD at Georgetown. In government focusing on German politics and, and really like cut my teeth as, as I guess I'd say an analyst did a lot of quant work and, got really good at Stata. I don't know if many listeners of this podcast have ever used Stata, but, I, I was yeah, I was really into it. And you know, thereafter just kind of decided that it wasn't going to, going to take the academic path. And. Had this incredible opportunistic, I guess, coincidence where I was searching for something online. And I got an ad that was like, do you want to, you know, get a job in tech? Like, are you into data? And you want to do, and I was like, am I being catfished? Like, that sounds pretty cool. This is 2013, 2014. And it turns out there was this. startup, Cosla, an initialized funded called called insight data science that was a couple of years old and they had started paid ads. And, they, they basically Jake, the founder is a good friend now, but at the time, you know, he's a little bit ahead of the curve realizing that, that there's this quant sort of revolution happening in tech. And, and that he could power it. So he, he founded Insight Data Science to help people like me transition into tech. And I ended up applying and getting rejected and applying again and not getting rejected and moving out to the bay. And, Ending up at Airbnb. And so I'll stop there because you may have questions, but yeah, yeah. And that's what's so great about the data industry. People have such amazing, rich, diverse backgrounds. You know, for instance, your story, you know, you grew up a math nerd. You spent time in. In Germany, you know, working at Deutsche Bank. And I was, I think I was at the last dbt Coalesce that's in San Diego and Abi Syvasylum about it. And Abi was the one who kind of, you know spelled it out for me, which is like, you know, he's worked in multiple industries, but he loves the data industry because people have like such awesome backgrounds and, and, you know but that's so funny to hear kind of how You, you found the, the, the early innings of like Airbnb and, you know, when people were using data as a, you know, an advantage to how they were doing go to market and, you know, acquiring users and making sure users. And it was addicting and, you know, growing and providing value. So, you know, I'd love to hear, you know, your advice to other people who, you know, they may be in other industries or maybe they're super early in data or some other field. Like what's your advice to people who are aspiring? Data practitioners, data scientists, data engineers. Yeah. I mean, this is probably a thread we can pull on for a bit, but, you know, I mentioned I went to a small liberal arts school and I mentioned that because, you know, what those schools really focus on is, is training Well, it's curiosity, I think, right? Like really cultivating curiosity, maybe giving folks some mental models, helping us think kind of taking an approach that like, you know, the world's always going to change. The skills that you need are always going to change, but what's a consistent you need sort of regardless of what the world looks like. And I've always believed really deeply. And I think it goes back to, you know, why a school like Bowdoin appealed to me was you need to know how to think. Right. And, and you and I've chatted a little bit about this, but like, I think, especially with all of the advances we've had in data tools over the past, at least 10 years, you know, now with, with co pilots and even Repl. ai and things like there's a set of tools and skills that are so much more. Approachable, right? Like you can, you can just learn things and, and, and build things way faster than, than you used to be able to. And I think, you know, we're going to continue to see that, that this democratization of, of tools and technical skills, right? And so, yes, you need technical skills. I I've always been of the mind that most technical things. You know, can, can be learned, a lot easier than, than, than sort of knowing how to think. Right. Knowing how to ask the right questions. You, you had talked about, you had mentioned like Airbnb, you know, our, our incredible growth there was of course technical innovations that, that we had at Airbnb, right? Airflow was built at Airbnb. Preset, superset was built at Airbnb. You know, the, the, the, the list goes on. They're, they're still open sourcing projects. But I actually think a lot of what drove the growth, we were enabled by tech, but it was like coming up with good hypotheses, asking good questions, right? Like, Oh, we're seeing a, a slowdown, you know, over in this side of our business. If you know how to use SQL, you can discover that, right? The question is, what are you going to do about it? And how are you, how are you going to think about solving it? And I think those things are rarely driven by tech. They're driven by. You know, how we problem solve, how we think about the world, right? How, how we think about, causality and, and, and incentives, right. Which are huge in, in most businesses, but particularly marketplaces. Yeah, that's such a great point, especially now that there's all these, technical accelerators, you know, whether it's co pilot or tools like cursor that kind of help generate code and can fill the technical gaps or for people who want to learn SQL or Python, get them there much faster than they could previously. And your point is that. You know, now it's all about, you know, okay, you have these technical skills. What do you do with it? How do you ask the right questions? So I'd love to dive into that a bit more. How does someone gain that type of intuition? Man, that's a million dollar question, maybe a billion dollar question. Right. How, how do I even answer this? I think there are a couple, I have a couple of ideas and it's a bit tautological. One is curiosity, right? I'm always coming back to curiosity. I think curiosity is the most, I have a daughter, she's just turned a year old and kids are just naturally curious, you know, like just watching her eyes, how big they get when she's like, learning something new or like seeing us do something new. I think curiosity is so, so important. It's probably a separate question, how you, How you cultivate it, how you indulge it. I think one way you indulge it is you read a lot. You know, fiction or non fiction. I don't know. A lot of people have a hard stance like, oh, you got to read fiction. And I don't, I definitely don't believe that even though I've read it. Yeah, I've read a lot. Yeah, I, I think reading broadly helps and, and, and here's why. I mean, I remember being a little kid and, Picking up the, the New York Times and, and not really understanding a thing right at first, and then you, you know, you start to sort of gather, I read publications all the time that I disagree with mostly because I get insight into how other people think, you know, what may motivate them, Yeah, and just like try to understand their world. So I think, you know, huge, huge fan of just fan of like exposure to different viewpoints, exposure to different material. On the one hand, I think on the other hand, there's something very practical, I think, particularly in data and that's. It's like actually digging into data, right? So many of us, the Titanic data set, the Irish data set, like we all, jumped into Kaggle back in the early days. And you know, I'm, I'm laughing because like everyone sort of did the work with these curated data sets. But like part of how you develop intuition is you actually go quite deep on, on something and you get to understand it and how do you do it. It's, it's actually to many people. They don't deem it worth their time. It's like you do exploratory analysis, right? Like you start cutting things by subsets upon subsets and you start to see stories. And I, I think that that part, that's actually, I think something we risk with a lot of the, where we are now in terms of like summarization tools, right? Whether you're summarizing an article or you're summarizing a data set. I actually think you have to spend time. In the weeds, trying to make sense of like what you're, what you're seeing. And I think once you've done that, you start to develop intuitions about, I mean, you have to be thinking about like data generation and things like that. But like, you got to get your hands dirty. You've got to develop a point of view. So, you know, I think to, to summarize, it's like. Curiosity is super important. You got to read a lot, expose yourself to, to varying viewpoints and, and, and start to sort of develop preferences. You got to get your hands dirty with data. And then I really do think this is hard for data folks, but like develop a point of view, take a stance, right? You might be wrong. And what does it mean to take a stance on the data set? It's like. It's saying something about like, and that's saying it could be a summary. It could be like a niche finding, but I often think data folks think like, Oh, this is someone else's job. I'm just going to show them the data particularly early in their careers. But I think. You know, back to asking good questions, developing and representing a point of view is like a massive unlock, on the, on the data side. Yeah, 100%. And I, I love that perspective. And it's really critical, especially now with everything that's going on in our industry. And, you know, one of the things that I really liked about your your blog on, you know, advice to aspiring data scientists. I recommend everyone read that. We'll have a link down in the show notes. But it was, it's also about, you know, showing your level of curiosity and interest and doing things like cold outreach for, for jobs that, you know, you want to go after. And you're sort of giving your future employees like a preview of, you know, the, the, the, the type of personality you have, because really, if you work in data science or data engineering, it's not just about, Hey, someone's going to send you a ticket to crunch some numbers, you're going to do it, blah, blah, blah, like that. That's all stuff that can be automated, right? Businesses are looking for data practitioners and data leaders. Just like you're saying, who can take a stance, who can come with a a hypothesis that the business can either, you know, prove is true or prove is, you know, negative. Right. So, yeah, you know, I, I, I think that's, that's, That's probably going to be the most critical skill set going forward in data. You know, assuming that a lot of the technical skills are, are, I wouldn't call it table stakes, but more people will have them, right? Yeah, for sure. And, and, you know, I think a common day to day thing is, you know, you mentioned, you know, linear JIRA, you got tickets coming through the worst thing that my team can do is just. run down the list of tickets. The worst thing they can do is just be like, Oh, this is my job, right? Because we're also presuming that the people asking, you know, creating the tickets, asking for things, like know what they're asking for. And this isn't a criticism about individuals. It's more that, man, you know, all of these async tools, We have, you know, kind of put requests first and not sort of exploration. And I think they sometimes take the place of having a conversation. That's like, Hey, I'm thinking about this thing. Here's what I think I need. Right. And, and often coming from, you know, maybe someone in marketing, who's of course like doing the best that they can, but they actually sit down with the data person and, and the data person discovers like, Oh, wait a minute. You know, you actually want this thing over here. That's so. You know, it's quite different. And had I just answered your ticket, like I would have been able to check it off, but would it have gotten us any closer to what you're actually trying to solve, you know? And, and I actually think that's a huge differentiator in people's careers as well. The, just if we bucketed like early. Early, mid and late, particularly between early and mid career is younger folks, they take it literally, right? They take it, something in a ticket, like quite literally, and then, and often miss an opportunity, right? Miss an opportunity to understand what, what their business partner, you know, really wants or, or needs. So yeah, there, there's some interesting tensions there. Yeah, I wanted to ask you specifically about another awesome blog you had titled being data driven is not a product strategy, and I think it kind of speaks to a lot of what we're talking about now, but you have this one quote in there. That's your data definitely shouldn't be driven by data. What did you mean by that? Yeah, yeah, that was, that was an interesting one because it had been brewing for years and I, I was kind of hesitant to, I kind of overthought it for a long time and then finally I was just like, I gotta, I gotta write this. Yeah. I think, you know, taking a step back, there's all, As a, as a data person, you're often, you know, we kind of simplify things and it's like, is your work data driven, you know, and data person would be like, I'm only going to work for a data driven org. Well, what does it actually mean to be data driven? Right? There's probably some nuance, but what we often simplify it to is that, like, well, data is going to make. Decisions as if data can make any decisions on its own, right? Like data is not sentient, right? Like it, it, it can't, we make decisions using data. And so, you know, I, I've been in lots of orgs where this is a big meaty topic, right? It's sort of like, there were times at Airbnb, for instance, I love, you know, I love Airbnb. My wife is actually still there. We met at Airbnb. We're still heavily invested in the company, you know, I wanted to do very well. But, we had so much success with our experimentation framework that we had this period where, like, people were afraid to make decisions until they ran an experiment. And it just induced this culture of Of, you know, local Maxima and, and, it was, it was really interesting because, you know, you'd be like, Hey, what are, what are you going to do about this? Like signup login flow. And, and there was like no point of view. It was more like, well, let's try these 18 things. And why shouldn't your data be driven by data because you're presuming that data, like that there's like a right or wrong, or there's like a progression in data, but it's like. You can easily be led astray by, you know, quote unquote data, like again, taking the example of, of A/B tests, right? You can do a sequence of a, A/B tests, that make a lot of sense in time, right? But you, so you run these things sequentially and then, okay, now we have this, this like holistic experience, but there is no holistic experience. It's just like, you did a bunch of small things that actually don't add up. Or what they add up to is not greater than the sum of the parts is actually less than, and so, yeah, that was that, that one sentence was like, You know, it really, to my mind, tried to capture or tried to represent that frustration with with the local maxima that that is an unfortunate consequence, I think, of, of, like, the success of a lot of the successes we had with experimentation. But then also thinking as a data leader, right? A big part of my job is. At least the way I think about my job at Replit is like, you know, one, one part of my job is building data core data tables that allow us to understand our business. And like, there, there is a way where I could figure out my roadmap. Like entirely based on data, but it needs to, I need to take a step back and actually base my roadmap and how I think about enabling my partners and the company that's entirely rooted in my stakeholders. Right. And how they think about the business, what their needs are separate from my, like, you know, little data team desires and maybe the, the, the, the like world. That I envision you know, the little private world that I might envision for data that, that, that could potentially be disconnected from the business, which would be a very bad pattern. Yeah, it's, it's super important for companies to understand that there's overhead and, you know, technical debt associated with everything related to generating a new report or running an experiment experiment, right? And even when you that's such an interesting point that you brought up about how, you know, at Airbnb, people were, you know, Hesitant to make decisions without experimenting at first you know, running an experiment. And of course, when you get into that, you know, you have to scope for the experiment. You have to think about the return on time invested. Maybe that's sort of a Self fulfilling prophecy where, you know, you, you, you don't make the decision because the experiment would take too long to run and you're just not very confident in it. And, you know it's, it's just one of those interesting things where companies have to figure out really like, what do they want to do? Like, what is the business objective here? And then the data team can work with you on how to achieve that business objective rather than, you know, throwing in a report over the fence. Like for instance, you might get a requirement from, from the business. Like, Hey, I want like a a map chart of, you know, all our, you know, Users by where they are in the state, and it's like, well, what are you going to do with that data? It's like, Oh, we want to do product launch and prioritize a region, right? And that's a much better way of phrasing the problem, right? Then saying, Hey, I need a new report. Yeah, yeah. And, and, you know, I think this topic is so fascinating because I've always been drawn. At least in the workplace to like really strong leaders. So, you know, Brian Chesky, a generational leader at Airbnb, right? Like he is, is just, he's so rare. And he's, he's. You know, grown tremendously. He's been running Airbnb for a while. And Brian will talk publicly about this moment at Airbnb, right. Where, where he was just, he, he can tell a story better than I can, but, you know, my understanding of it is he was very much like things were kind of out of hand. Right. Like, And COVID, of course, played a massive role, right? Business evaporates overnight on the cusp of a potential IPO. Like, most people would have not, not made it. But, you know, Brian did and, and, and continues to. And I think a big part of that was him kind of saying, like, we're going to do things different. And, you know, Brian makes a lot of decisions at Airbnb, right. Continues to, and, it's, it's interesting because I think there is something at least personally appealing to me about like centralization of, of decisions. It doesn't work for everyone. But strong, strong leadership, I think, is a really rare quality. I think particularly strong leadership, as it pertains to product and product strategy. Yeah. I mean, the one, the one thing I'll say there that kind of, I think completes it is like, you know, Brian will talk about this, this period where, where, you know, there's a lot of delegation and, and there may still be delegation, but it's, it's a little more centralized and it's, it's something I think really interesting to consider. Cause what does it mean for data teams? Right. A lot of my friends, a lot of us have left Airbnb for various reasons, but there was definitely conflict right over. And it's even odd sort of talking about it. Like I helped build the culture of experimentation. And I'm also maybe because of that hyper aware of what the limitations are and it was You know, there are lots of, there was a lot of tension around like, okay, we're not gonna experiment. What? Like, how can a data scientist possibly say, you know, we're not going to experiment on this major launch, but you know, I'm at a new company now, Replit, and it's, it's been really interesting for me to sort of come out of the other side. And like, of course, I want us to experiment. I don't want us to experiment on everything. There are plenty of reasons not to experiment. There are plenty of times. When, you know, as you noted, you want to think big and experimentation can can sometimes be at odds with that. But it's, it's interesting to, like, reflect on my own journey from, like, being such an evangelist to now being. You know, it's probably just maturity, but being a little more practical, like I'm constantly having conversations with engineers and I'm like, I, we don't need to experiment, let's just do it. And they're like, wait, what? That's the best type of experience, honestly. So yeah, the, knowing when to be practical and knowing when to sort of just, you know action is better than inaction, you know, especially for, for, for companies that have to move fast to, to, to get, you know, Product market fit, product adoption, et cetera, you know, totally. I think that's so, that's so well said. And, and, you know, I think about it in terms of leadership. It's like, you don't, you don't have to make the right decision all the time. It's about the speed of decision making, right? Like, man, being wrong is great. As, as long as you're fast to, to acknowledge it and, and move on, I'd rather be like wrong more often and making more frequent decisions than, than the alternative. Yeah, I love that. That's that, that should be quoted and put up in the wall of every data engineering room, assuming they, they have an office, if they work remotely, then, you know, pin it to the Slack channel or something. Totally. Yeah. So, so you've had, you've had such great experience both as a individual contributor to data science teams and as a leader. So I want to get your take on how the data scientists or data engineer can up level, you know, just from, you know, thinking about everything is like a technical problem that needs to be solved. And then Using those technical skills to unlock value in their company. Yeah, good question. I think context is always helpful here. And when I say context, I in particular mean like the context of the data team usually or data engineering team. I think, I think we're seeing a shift in the way that data teams think about their identity, like their ID, who they are, and their value. I think that, again, this is very based on my experience at Airbnb. 10 years ago, you know, there were so many projects that we did that were like, they were cool, like cool data projects, right? But like, were they really moving the business? Like they were cool, maybe technical challenges. They were interesting questions, but like, There was a lot of navel gazing, I think around like, I'm a data person. I have these great skills and I'm going to build this thing that like, you know, had you done the TAM analysis of who's actually going to see it in your product? Like an example, right? Like a, a very small feature model, like in the Airbnb product that like, you know, 10, 10, 000 people would see the feature, like 10, 000 people at Airbnb scope is like nothing, right? And there'd be months of work on this feature that you were like, wait, what? Like this, this thing pops up when you're 12 pages deep and you actually made a mistake, you know, like, and I'm not, you know, it's funny. I'm not making fun of anyone, but even though it may seem that way, But we like weren't really asking, you know, the question first, which is like, who's going to see this and, and like, what's the impact. And I think today we're, you know, whether it's the end of Serb, like, I don't know, but the times have definitely changed in terms of like, My sense is that data teams now are much more about empowerment of their partners, right? Like unblocking of their partners. It's, it's not like these, I'm a data scientist and I'm building this cool thing. And I think that's healthy. I think it's more than necessary. And, and, you know, I mentioned that because I think if you're on one of the teams now, which should be more rare, that's like, you know, more inwardly focused. It's at odds with your question. And it, and these companies exist. These teams exist where it's just sort of like something's not working. And it's it often. Has to do with how the data team sees themselves or how the org sees the data team, right? Because if the team, someone had a tweet recently, that's like, oh, your data team's bored. Like I've never met a bored data team, you know, because if anything, they're like doing too much. And I was just like, I was bewildered. I was like, what does that even look like? But, but you can imagine it in these situations. So ideally you don't end up in that situation. Ideally you're in a, in a situation where like, okay, your value, your impact on the business, if it's not direct revenue, it's unblocking your partners so they can, they can have. more, more direct impact to revenue. And, you know, I think more often than not, you have to sit down with your partners, right? You have to understand, you know, if you're working with someone on a, on like an ops team, a business team, Like literally, what is, how do they think about the business? What is the problem they're trying to solve? One thing we've had a lot of success with on my data team at Replit is, you know, no one likes filling out linear tickets and assigning tasks. You know, like I, I really don't think many, well, someone probably does, but there aren't many people. I think so many people want. They want to be empowered, right? They want to be able to solve questions, solve problems. The challenge is the tools, right? They don't have the tools. They're not engineers. They can't build them. The tools don't exist. We can't buy from the vendor. My team has done a lot of work around like actually building tools for our partners in business, in sales, in finance. We use Hex. I'm a huge fan of Hex. Super talented team, great product. One of the reasons I love Hex is that it allows me to empower other people. An example. This is such a trivial example, but it's been such a big unlock for my team. Research saying, Hey, I've got this list of, user IDs. Can you get me emails? And the first couple of times you're like, sure I can do that. Join on dim users. And then like five times in, you're like, I. I, I can't, I don't want to do this. Like, it's not that it's hard, but it's, well, it's hard and it's easy and my brain is dying. So, you know, with both Replit, actually we can do this with Repl. it and, or with Hex. It's like, you can templatize a query with Jinja, right? And like, give them some dropdowns or allow them to upload a CSV. And boom, all of a sudden they're getting the emails on their own. It sounds trivial, but you know, all of a sudden the researcher is like, I can do this on my own. I don't have to bother the data team. They feel better about it. That's a very trivial example. But we have countless examples like that. I have this like little app that I built. I still very, you know, I lead a data team, but I still like to build a lot. And it it's a, we call it the B2B B2B report finder. And, you know, you enter a business domain and there's all these template inquiries behind the scenes, you know, that like, that's not beautiful. It's not like career defining work. And that is not the point. The point is it's what the partner wanted. So yeah. Yeah, yeah, that's that. That's definitely awesome. And yeah, shout out Barry McArdle. I just met him last week in Vegas at Google Cloud Next. So cool to hear the Hex shout out and and also the amazing work your team's doing at Replit.. And yeah, I think it really all comes down to empowering your your team. You know, your partners in the business to better make their own decisions and really understand what they want to do. That's always my first thing. When I hear someone in either go to market teams or product or engineering come to me with a question or a task, I'm always like, well, what do you want to do? What are you trying to achieve? Right. And then that's usually where. The real requirement comes up and yeah, I think empowering people to work in their own tools is really great too. And you know, for instance, you know, we, you know, we use you know, our, our, our product stream for a lot of our ingest pipelines, but we use high touch to sync the data and to, you know, things like HubSpot and Salesforce. So I don't have to onboard like, you know, a sales person into like a data tool or like a reporting tool or big query. For them to, you know, know how many credits their customer is using, right? That data is just there for them ready to use in the tool they're familiar with. Right. And I think this is a, and, you know, this is, yeah, we're, we're shouting out a lot of cool vendors, but, you know, at the same time, it's really that, that kind of intuitive thinking about making things easy for your your partners in the business to, you know, make, make the decisions they need to make without too much. You know, overhead of tooling. Right. So, and I think that this is going to be different for every data team, you know, depending on what your, how your company operates, some companies need that reporting layer layer, and they need that, you know, semantic layer, whatever you, whatever you want to call it because you're, you're, you're just too, too, too massive to, to deal with like sending data to like different pools and stuff. So, but yeah, it's, that's, that's the fun part of the journey here. Totally agree. Yeah. Yeah. It's super fun. And the, the tools we have today compared to, to 10 years ago, you know, are just. I was reflecting on this, chatting about this with a colleague the other day, like, when I started at Airbnb, making a shiny dashboard, you know, was like, you had to sort of walk through hell to do it. And most of the time you actually couldn't be successful just because of security and things like that. Right. But like, I think about it now and I remember doing work at Airbnb on our superhost program, right? Like fantastic. Retention lever for airbnb is is is this super host program, which acknowledges, you know, hosts who who offer excellent service. And, you know, we're thinking about revamping it a little bit and, you know, It was so funny. I just discovered like, there were like very small changes I could make to, to sort of the code that, that, that just like, Oh, all of a sudden 400 more people are eligible for this because we made a rounding error or something. Which was, you know, it's always fun when, when you find stuff like that. But ideally I would have created a dashboard that I could give to a partner and the partner could, you know, move a slider and see the impact, right. Of, of different, you know, levels of a certain metric. Like today, that's just a no brainer. Right. But, but 10 years ago I actually, you know, we are really slowed down in terms of decision making and it's really fun that, that tools are fun. I think tools are really fun. Yeah, absolutely. The tooling is, is fun these days. Yeah, I, I agree. Especially, you know, 10 years ago compared dealing with things like Hadoop and HBase and figuring out how to program MapReduce jobs. Now, so much of that is, it's SQL and these nice, shiny SAS tools that, you know, feel a lot like, you know, the you know, business applications you're, you're used to using already. So, but, you know, of course there, there, there's still the overhead of needing to know how it works under the covers to get the full usage out of it. Yeah. So, yeah, absolutely. And Lindsey Pettengill, PhD head of data science at Replit. Thank you so much for joining today's episode of what's new in data. Where can people follow along with your work? You can find me on Twitter. I don't know my handle, but you probably do. I'll put it down in the show notes. And then I have a website made with Replit, lindsaypettingill. com. So check it out. Oh, so fun. You, you made your website with Replit. I, I got to check that out and see if I can, I can put my portfolio together with that. I use Replit AI. No, no, no need to know CSS or HTML. Oh my gosh. Okay. Well, thank you. That that's, that's definitely super great. I I'm excited to check that out. Myself for my own portfolio page. Lindsay, thank you so much for joining and thank you to the listeners for tuning in. Thanks so much. It's a lot of fun. otter. ai