Why Data Scientists Should Break Things (Daniel Parris) - KNN Ep. 163

I decided when I was 8 years old that I like watching movies, and I don't think that 8-year-old Danny had an understanding of what it was going to be like to actually work in the entertainment industry.

I think so many people are trying to be similar to other candidates with their portfolio. They don't realize that being one of the better candidates but not the best candidate almost never lands you the job.

The best asset you can have in a career in entertainment is have a network, and you have to gain a network by talking to people and convincing them to like you. That was really helpful coming into the data science profession.

Over the course of a pretty long career at a company, usually something goes wrong. Okay, how do you handle those? But also, how does a company handle if someone makes a large mistake?

When I was on the ops team, I deleted the geography of San Francisco from the DoorDash database, and so it just shut down for a day. So drivers couldn't drive, people couldn't order food, and restaurants couldn't get orders. And it was one of those things too where you looked in the database and it said like, Daniel Parris made this change. To make it even worse, the way that DoorDash would feed employees was by placing DoorDash orders for dinner. And so we couldn't place the orders for dinner, so nobody could eat food.

I can't think of a time where when I took accountability for a mistake I made, it went badly for me.

Whenever I'm having a bad day or I've done something wrong, I try to take a step back and think, is the thing that I'm upset about right now going to matter to me in 10 years, in 20 years? And if the answer is no, which it almost certainly is pretty much every time, that's always a weight off of my shoulders.

This episode of Ken's Nearest Neighbors is powered by Z by HP, HP's high compute workstation-grade line of products and solutions. Today, I had the pleasure of interviewing Daniel Parris. Daniel is a data scientist and data journalist with over 8 years of experience. He was one of DoorDash's first data science hires., and he currently invests in early-stage data products through Dash VC. He runs the newsletter called Stat Significant, which crafts data-centric essays about pop culture phenomena, and Data People, a short-form interview series with world-class data professionals, which fortunately I was featured in recently. In this episode, Daniel explains what it was like to work at a quickly growing company with an experimental culture like DoorDash, what he learned from his biggest mistakes, and why he decided to pursue consulting and data journalism after DoorDash. I really enjoyed this conversation. I know you will as well. Danny, welcome to the Ken's Nearest Neighbors Podcast. Such a pleasure to have you on.

Good to be here.

Awesome. Well, you know, it's funny, we met through a good friend of mine, Ben Weiss, well, mutual friend of ours, and he was telling me about, you know, your newsletter as well as the project you're working on with Data People. So I was really honored to be able to come on and be one of the first guests for that. And you know, I'm sure when, when that releases, I'll be happy to, to share that as well. But before we get into your newsletter, your work at DoorDash, and the, the new Data People project, I would love a bit of an origin story. Where did you first get interested with data, data science, this whole domain?

Yeah, so I had a weird, a weird path to tech, into data. I guess when I was a kid, uh, I watched like a special on America, or it's AFI's like 100 Greatest Films, and I became obsessed. I downloaded the list and I vowed to watch every single movie on the list, and I think I watched like 90 of them. And then when I was 8 years old, I said I'm gonna be a filmmaker. And then that basically carried through until about midway through college. So I was a film major in college. I actually went and I worked in entertainment. So I interned for Steve Carell and his production company. I interned for a documentary filmmaker named Kirby Dick, and I actually interned at the Conan O'Brien Show. And they were all— I mean, when you're at the lowest level of entertainment, not the greatest experience.. And so I ended up really not liking that. And so I decided to quit and I actually really didn't know what I was gonna do with my life. And at that point in time, my wife said, hey, it was at the end of college. My wife said, her then girlfriend said, hey, I want to move out to San Francisco. And I said, you know what? Cool, me too. So I just started telling everyone I was moving to San Francisco and then I had to actually figure out how to get a job. And so I started off working at sales at a travel startup, and I didn't really know what I was doing. I was terrible at sales, absolutely terrible. I don't think I actually sold a single thing while I was working in a sales function. And so I believe this company figured out that they needed to repurpose me into something where I could be more useful, and they thought I might have an aptitude for analytics. So they actually paid for me to start, uh, to basically take a course to learn Excel and SQL, and I found I really liked it. So I started furthering my own education, going to all these like General Assembly courses to learn Python, machine learning, statistics. And then I honestly just enjoy— I liked having a hard skill or a set of hard skills that I could lean on, and I found that that was going to be valuable in my career to just have that as a bedrock. And so I just started buying statistics textbooks and just reading them. My roommates thought I was crazy because I would just come home from like a long day of work and then I would just take a statistics textbook and put it out there. I think they thought I was joking. And then I also started doing all of these like free Coursera courses. You can just audit them. And eventually I moved to DoorDash. And at the time I started off in operations and I was like a data-savvy operations person. And then eventually after a year I moved to the data science side and I've been doing data science ever since. And so I don't know, I, it's kind of felt like a, I ended up focusing hardcore on film for 12 to 13 years and in some senses it feels like a waste, but in other ways it feels like a superpower cuz I, I don't know, come at data science from an unorthodox background. So I don't regret any part of my journey. That's awesome.

And you know, something I see as a recurring theme in the data domain, actually in a lot of domains, is that our expectations of what our career would be like, they don't always match reality. Can you talk a little bit about what you were expecting out of entertainment and what you actually got? And then maybe what you were expecting out of data and what it— or like your new career and what actually you ended up working with?

Yeah, I would sort of— the thing that I ultimately tell people as it relates to entertainment is that, you know, I enjoy eating hot dogs, but I don't care to know like how the sausage is made. And that is effectively how I think about entertainment. I decided when I was 8 years old that I like watching movies, and I don't think that 8-year-old Danny had an understanding of what it was going to be like to actually work in the entertainment industry. I think one thing in particular I found frustrating, and this is based on my own experience in entertainment, is that I felt like I had to basically turn my brain off for the first 5 years of my time in entertainment. Oftentimes when you're entering the industry, some of the best jobs are to be someone's assistant. And that's the typical career track. And it's not something where you can come in at a certain level based on your talent and based on some vocational training. Like you come in and it, it's actually a great job to be an assistant on someone's desk at a talent agency. And so after all of this hard work and, you know, getting a good education, it, it felt like it was ultimately gonna be a waste of my brain for 5 years. And I just found myself, for something that as a child I found so intellectually engaging, I realized I wasn't going to be stimulated for 5 to 10 years. And that's taking a risk not knowing that you're— if you're gonna actually get to write something or produce something or direct something. So that was my core frustration with entertainment. And then as far as data science, I actually found that, that same intellectual engagement that I was looking for I also found that a lot of the skills that I learned in college as it related to critical studies of film actually translated really well to a data science career, particularly like problem solving and decomposition, or like problem decomposition. When you think about the way that you're taught to analyze a film, you break it up into different parts and there's a problem and there's an obstacle a character has to overcome and there's a beginning, a middle, and end. And I started realizing that I could think that way about problem solving. And so it was this interesting combination that I really loved where you had this problem-solving element, which I felt like I almost transferred some of these skills I learned, you know, analyzing like your critical theory around texts, and then ended up learning all these hard skills that I actually had the tools to, you know, ask and answer my own questions. And I just loved it. And then being in the DoorDash environment where you have such, you know, you have so many different problems and so many different fascinating datasets and the feedback loops are daily, sometimes hourly. That was really just such a gratifying environment every day. I, I think the best decision I ever made was not going down the entertainment route and picking data science.

Yeah, it seems like the data science career path is almost the complete opposite of those first 5 years in entertainment. For data science, you're sort of given a lot of autonomy. You have to figure out things on your own. And I think that can be challenging for a lot of people on the other side of that. A lot of people coming outta school, they're looking for clear expectations about their work, just like in the classroom where, oh, you have this data, we want this outcome. And for better or for worse, that doesn't necessarily come in every data science position. It doesn't probably come in most data science positions, But if you like the aspect of critical thinking, it makes a ton of sense to pursue this career path, or it makes a ton of sense why this resonated you with you when you started doing the work. I also think a lot of people discount, for example, film history, a lot of these backgrounds of study when it comes to more technical careers. But I was listening to a really good podcast. It was one of the Tim Ferriss podcasts recently. About how in consulting, history is really a good base. So I'm not sure how much people know about consulting, but there's 3 main consulting companies. There's essentially McKinsey, BCG, and Bain. And Bain, which is someone some believe to be, you know, they're all, they're all high-caliber consulting, but the guy who started Bain Consulting Group was, uh, he studied history He had no interest in business, but he was able to leverage his background into a very effective skill set around storytelling. And what people don't realize around history, a lot of people think, oh, we're just studying what happened in the past. History is a lot about thinking critically about what would happen if events didn't transpire. And that's a lot of what data science or consulting is. It's Oh, if this event happened differently or we had this new thing coming into the data, what would happen? That's a very valuable skillset of putting these pieces together and thinking about theoreticals, which you don't necessarily get in a lot of the more maybe linear STEM fields of study. And I think that that's valuable. I'm not saying it's necessarily more valuable or less valuable, but it is a valuable skillset to be able to think about these things in a different way. And it's cool that you've realized that as a superpower as well.

Yeah, when I first joined the tech world, it was actually one of my greatest anxieties. I wouldn't tell people that I studied film. I think on my first resume that I applied, I changed it to communication studies thinking that that might be more palatable for someone to hire me. Way better. And then maybe I think I actually When towards the end of college, when I realized that I had made a terrible mistake, quote unquote terrible mistake, I added a minor in business, and I was really banking on that minor in business getting me a good job. And so ultimately though, over time, I found that that was actually, you know, as I started hiring people, I, I realized as a hiring manager that you were looking for things that made people different or distinct, and that for me at least, given my— this path that I've gone down, this unorthodox path and unorthodox education ultimately make me different. Uh, one interview— so I guess one of my, like, good— my great party trick is that I can name every movie, or I used to be able to name every movie that won Best Picture since, like, 1950. That's just— I don't know, it's a childlike enthusiasm, and I just memorized it. Uh, one of my final rounds, the first 5 minutes, the person who was interviewing just quizzed me on what were the, the best picture. He would name a year and I would have to rattle off what the best picture was. And then once he was satisfied, we started the interview, but he said that that was the thing that he found most memorable. And so he wanted to talk to it about me, so, or talk to me about it. And so over time, I've actually become very proud of, of my path.

That is something I probably beat this to death, but I loved what you said there about what separates you from other candidates. I think so many people are trying to be similar to other candidates with their portfolio, with their approach to interviewing, with all of these things. They don't realize that being one of the better candidates, but not the best candidate, almost never lands you the job. So if you're one of, if you're the 15th best candidate for a job, you're probably not gonna get the job even if there's 1,000 applicants. But if, and if you do that for every single job, let's say you applied to 100 jobs and you're the 15th best candidate in every one. The odds of you getting a job are still kind of low unless 14 other people don't want the job or whatever it might be. But if in 3 of those you're a top 3 candidate and in the rest you're the worst candidate, you still probably have a better chance of getting a job than that other person who's in the, you know, 90th percentile or even the 95th percentile in all of them, which is kind of crazy to think about. But it's just what you said. It's about celebrating the differences. Because that's what people are gonna recognize. You know, if you think you like stick your hand in a bag with a bunch of smooth rocks and one spiky, that's probably the one you're gonna pull out or at least look at further and understand why it's different from the others. And, uh, there's something very special in my mind about recognizing that. And yes, you'll probably get a lot of positions where you don't fit, but the ones where you do fit and people like that, it immediately ups your chances.

Yeah. Plus, I would say one of the things I was able to develop in a career in entertainment, uh, was, was— I mean, I, I don't know if I'm— I think I had to communicate a lot is the way I'll say. A lot of— basically, you're the asset. The best asset you can have in a career in entertainment is have a network, and you have to gain a network by talking to people and convincing them to like you. And that was really helpful coming into the data science profession. There was actually a time at DoorDash where we were asked to pick whether we wanted to be an IC or manager, and I sat down and I looked like the VP in the eyes and I was like, IC. She's like, no, manager. And I think a lot of that owes to doing these reps, having to constantly communicate in entertainment, because that was effectively the only way I could think to differentiate myself. And again, Ended up being one of my greatest career assets. So it goes to the, the, to your point about soft skills being this incredible differentiator throughout the course of your career.

Yeah, I, I, I agree. And I, I think it doesn't even have to just be soft skills. It is anything that makes you different from other candidates, whether it's a project you work on, something you built. It could absolutely be soft skills. I find that to be one of the lowest hanging fruits for a lot of people. But you know, just everyone, if you look at a portfolio, a lot of people just have the same projects as everyone else, or they have, you know, similar things. And if you just do something that's fundamentally different, even if it's silly or stupid or weird, that's gonna open people's eyes. Like if it shows some representation of you, you know, like there's a chance that your interviewer really likes I don't know, whatever it might be, like anime or specific sport that you're also obsessed with, especially if you're choosing domain relatively specifically. Like if I wanted to work in sports and I had all my projects that were about unique elements of sports, that would probably be a really good talking point. I mean, I saw something recently just working in basketball. There is a massive disproportionate number of players named Jalen. In the NBA.

There's— oh, I saw this.

Yeah, yeah, because of Jalen Rose, right?

Jalen Rose.

Yeah, yeah. There's, I think, 32 players in the NBA, so essentially one per team, uh, named with a first name some derivative of Jalen. And, you know, doing a project on that, that is an interesting story to tell, especially for someone that likes basketball. Like, who's your favorite Jalen? I think that's like kind of fun. Uh, but, but, you know, most people won't do something that would be quirky or different like that. And I don't think if someone was a basketball fan, I don't think anyone interviewing someone who's done that project would not ask about it or want to learn more or would not be willing to listen to the story around that. You know, I saw one of my favorite projects ever was, it was on Reddit. I don't know how real it was, but they looked at James Harden and they looked at essentially the ranking of all the strip clubs in every city that he played in, and they were able to correlate the quality of the strip clubs with his performance. And so essentially, the better the strip clubs, the worse he performed in those cities. And obviously there's confounding factors because you look at Miami, you look at New York, you look at, um, those are probably two big ones. Like the Party City players go out and there's like a, there's a curse associated with that because, you know, the, the home team players, they generally would get it all outta the system. When people visit, they want to go see all the sites. But to me, those, those are things where, ah, it's, it's such a good story. Anyone interested in those domains remotely would want to ask you more questions, or you would always come up in their mind when they're reviewing the different candidates.

Yeah, and to a certain extent, you, you kind of want to work for somebody who values the things that make you different. It's actually— when you're early in your career, getting rejected, it all feels terrible. But over time, you start to realize that if you don't get a job, it's probably because it wasn't a good fit. And if you had gotten that job, it wouldn't have been good for either party. And so that's another thing. Over time, it becomes a good filtering function. Like, maybe I did not get a master's in statistics, or I didn't go to a data science boot camp. But if someone is willing to find value in what I bring to the table, then we're going to have a great partnership.

This episode of Can Xerus Neighbors is brought to you by Z by HP, HP's high compute, workstation-grade line of products and solutions. Z is specifically made for high-performance data science solutions, and I personally use the ZBook Studio and the Z8 Workstation. I really love that the Z workstations can come standard with Linux or WSL2, and they can be configured with the Data Science Software Stack Manager. With the Software Stack Manager, you can get right to the work of doing data science on day 1 without the overhead of having to completely reconfigure your new machine. Now back to our show. It, it's funny you mentioned that, the them finding value in the way you do things. I had a very interesting filtering mechanism in the latter part of my career where my resume was very colorful. It was super integrated. It had a lot of data visualization. It had all the company logos and stuff that I worked in. And I knew that probably every Fortune 500 company would not be interested in a resume that is non-traditional like that. It was fundamentally different than almost every resume that they saw. But if someone picks it up and says, oh, this person understands design, They understand how to tell a story about what they're working on. They understand like these elements. It, it sort of created that 80/20 ecosystem for myself where it's the people that it resonated with, it would probably go to the top of the pile. The people that didn't like it, it would probably go very close to the bottom of the pile. But that's a better probabilistic system for landing a job or getting opportunities that really resonate with you than. The other way of just sending a boring old normal resume that is designed to get you past one screen. I think if you design your portfolio, your resume to get you to the final round, that's a very different story because once you get past the screen, everyone's equal playing field. If you do something that's fundamentally different early on, that can help you jump essentially the whole line. Which is a paradigm shift for a lot of people. So I wanna change gears a little bit and talk about your time at DoorDash. So something you talked to me about offline was about how you made that position shift to data scientist within the organization. Can you walk me through that a little bit?

Yeah, so I actually originally applied for a data role at DoorDash and at the time I was 23. And I didn't have any experience working in a data gig. So, but I put together this presentation and they were impressed with my attention to detail, so they actually routed me to the ops track. And so I got an operations job and I was running the— I was helping run the Bay Area market. And I kind of, in the back of my mind, was just kept thinking, how can I move into that data science role that I really want? And so One thing that was a huge problem at DoorDash, still remains probably a very big problem at DoorDash, was how do we know which restaurants are the best restaurants to be on DoorDash? So at the time, the company was pretty young. I think when I was joined, it was in the 150 to 200 range. I think now it's in the 10 to 12,000 range, something like that. And so the, the big, like, the way that the, the company was getting to scale and all the competitors were getting to scale was how quickly can we get the best restaurants on our platform and maybe get them exclusive on our platform. And so I was actually doing the job on a day-to-day basis of going out and trying to sell the restaurants. I was also still terrible at sales, so eventually they just put me inside of the HQ and they made me figure out what were the restaurants we need to go after. So literally making lead lists. And I was making these lead lists and it was based on what I thought was tasty in San Francisco, which is a really terrible way to approach the problem. So over time, I kind of became obsessed with how can we find a way to actually, in a more data-driven way, predict what is going to be good on the site before it comes on the site. And so I was taking courses at night. And so what I would do for those courses is try to actually build my own models, predictive models, to estimate how well a restaurant is going to perform once it's on DoorDash. And so I kept like honing this thing and honing this thing and still wasn't good. It still wasn't good. And then the company built this feature where consumers could actually request what was on this, like what they wanted on the site. And so this was kind of like a nice lightbulb moment where I took this, and I remember there was kind of this scrum to figure out, okay, how are we going to do this at scale? Because company was getting to a point where it was expanding to all of these different markets. And so there was a need for a centralized way to say, this is the value of this restaurant before it's on the site. And I remember I kind of came forward and said, hey, this is a problem I've been working on for some time. I've got some really terrible versions of this model, but let's like try out what I'm working on. And so ultimately I was able to partner up with a very talented ML data scientist. Her name is Dawn Lu. Shout out Dawn Lu. And we were able to work on, I guess, the first version of what ended up being DoorDash's selection intelligence model. And throughout my whole time working on this, it was, it was like, how can I develop this thing where I cannot be ignored by the company and I can eventually end up on the data science team? And so it was, I was able to take all of this information I had gleaned while in ops and my understanding of this problem, actually having to do it by hand. Combine it with some amount of data science acumen, and then ultimately working with a very talented ML person to kind of formalize it and get it across the finish line. Uh, and we were able to prove out that it was better than how we were doing, how the company was doing it when, when people were making lead lists by hand. And then we— that ended up being like the way that DoorDash, it still does, I guess, evaluates like these are the restaurants we need on site. And then shortly after, I was asked to join the data science team. So it worked out, but it was kind of like this project I had in the back of my mind that was in my— it was sort of a fantasy of like, this is gonna be the thing that gets me to the point where I want to go. And it, it ultimately ended up working out.

That's awesome. It seems like there really is sort of a playbook associated with that as well that, that maybe other people could replicate is that if I'm summarizing correctly, Essentially, you take your existing work, you figure out how can data be integrated into this to make it more valuable, to save me time, to save the company, you know, effort, increase revenue, lower time involved. And then how do I build that thing, whether it's by myself or with a team to get recognized for being able to create value with data for an organization. You probably, it probably would be useful to have conversations with your manager or team to let them know that that's something you're excited about. But to me, it's really beautiful because you're still doing your work. You're just finding a way to do your work more effectively. And what company is going to turn that down, especially if you're still doing high-quality work on the job?

Yeah. And there are all these times where I, where I would say, this is the restaurant we have to get. The, you know, this, this burger place called Popson's in San Francisco is going to be the thing that, you know, explodes the market. And we would sign that restaurant, put it up on the site, and it would do zero orders. And the salespeople were compensated based on how many orders a restaurant did when it went live on DoorDash. So they would be angry. And so after a few times of me making the wrong decision over and over again and people not being compensated as a function of that, I realized that I was not good at doing the job manually and I needed to figure out a way to do it in an automated fashion. And so I wanted to basically work myself out of a job, and then also I figured this could be something that would help at scale for DoorDash. So it originally arose from that need, and then once it became important politically within the company to do it in a centralized manner and then disseminate it to all the other markets, that was when I think I was able to action on my, my need or my desire to join the data science team. And so it ended up being a marriage of— a very convenient marriage of two things that I wanted.

Yeah, I, I, I like that idea of being able to sort of align all incentives towards getting you what you want with the organization. That is a really powerful thing. Something also that we talked a little bit about offline was that with that sales team, you kind of pissed them off a couple times, even in your data roles. Can you, can you explain to me what, you know, what that means and how you reconcile something like that? I mean, I think everyone I've worked with there's always been internal conflict between teams, between individual people even on the same team. How do you navigate something like that where you do something and it doesn't exactly make everyone happy that you work with?

Yeah, so when I, when I, when I became a data scientist formally, I thought that the thing was good that was going to keep me up at night was how accurate my models were and whether other people on the data science team are going to say that I was a fraud and basically singled me out as someone who wasn't doing data science in the most, I don't know, true-to-form way because I didn't have the same formal training as everyone else. Totally wasn't the case. The thing that ultimately was really difficult is that we— so this model that we built ultimately ended up impacting the way that salespeople were compensated. So it would say, hey, This is this restaurant that's down the street. We think that this restaurant is going to perform in this manner. Ipso facto, if you bring this restaurant on DoorDash, you get paid X number of dollars. So the way that we used to frame it is, imagine— so there's like an algorithm that assigns pay to the drivers, and it's saying for this order you get, you know, X number of dollars, for this order you get X number of dollars. Imagine if you built that algorithm but it was internal to the company And anytime you made a mistake, the person could Slack you and say, what the hell is going on? I'm not sure if we can swear on this podcast, but I definitely got sweared out a bunch. Okay. They said, what the fuck? So, uh, basically we, we built this thing and there were all of these internal stakeholders and the, the first, like the, probably the worst day of my career far and away is that when we originally built the model, we built it off a dataset from Foursquare. I believe at the time the company didn't, or the, the team that was working on it didn't want to pay for data from Google. They want to pay for data from Yelp. They didn't want to pay for data from some of the more expensive solutions. So we ended up with the data for that from Foursquare. Foursquare's dataset, I don't know how off, how much the app Foursquare is used today, but a lot of the data they had, especially the data they turned over to us, was 10, as much as 10 years old. So there was some really rough data in there. And so I, again, was really early in my career. I was probably like 23, 24, and I didn't think critically about the fact that we were building a data pipeline and that there are important data quality checks that you need to do. And you need to think about systems for observability and reliability. So ultimately what ended up happening is the first day we released all of the leads to the salespeople who, you know, work within the company, it was like 100, 200 of them. They got a lead of— they got 200 leads, uh, 200 restaurants. And I think about— I think we estimate about 20% of their lead list was just like not restaurants. And so to just maximize the pain, they actually kept a Google Sheet of the worst offenders. So I think there were at least like hundreds of strip clubs that ended up in this lead list. And they're, uh, McKinsey, just like— that was actually a joke that I wanted to make, but I was just in so, so much deep water that it wasn't not worth it. Uh, I think McKinsey ended up on the lead list. Um, DaVita, which is a kidney dialysis center. Uh, it was just a huge mess. And I think all in all, it was a Google Sheet of like 4,000 really terrible leads that showed that we didn't— and people ultimately, I think, initially lost faith in what we were doing because they didn't understand there's oftentimes a difference between data quality and maybe the data modeling, or at least like the quality of the model, and then the quality of the data that we're feeding into the model. But I had to sit down with the VP of sales and we had to review this Google Sheet of strip clubs and consulting firms and kidney dialysis centers. And it was definitely one of the worst days of my life. I don't, I haven't cried in my like adult career, but it was, it was, it was a tough day. It was a bad one. And so yeah, that was the problem. And then ultimately what I learned coming out of that was, well, I learned a lot about data quality and how to build a really successful data pipeline, and that was on the technical side. But then I had to basically do damage control with 200 different stakeholders. And so it was actually sitting down with them, making them feel like I understood how they did their jobs, which was a huge component of it, because I had just replaced the way that they had gotten paid, and my first swing was a terrible swing. So I had to go talk to all these different sales reps, help them understand what this model was, and then also help them understand— like, help them understand that I understood what their job was and I understood how they thought about getting paid and going out inside of these restaurants. And so that, I would say, coming out of that was probably the best learnings of my career, uh, as far as stakeholder management is concerned. Especially if you're on the data science side and you work on anything important, you're going to be dealing with non-technical stakeholders who need assurance that what you're building actually takes into account their problems and what they're trying to do, and that you're not just trying to jam some technological solution down their throat because you saw something cool on Stack Overflow or Kaggle. And so ultimately, we were over a very long period of time and being effectively like an evangelist for this system that we were building, we were able to win back a lot of faith from the sales teams. But it, it was a pretty bad start. It was a rocky road.

It seems like from this conversation and many others I've had, sort of the low points are also the best learning points, especially when you recover from them. Um, can you tell me about maybe some other low points or some major screw-ups and what the takeaway— I mean, over the course of a pretty long career at a company, usually something goes wrong. And I'm interested in, okay, how do you handle those? But also how does a company handle if someone makes a large mistake? Because I think there is so much fear about making a mistake in your work that a lot of people don't take initiative. They don't build things that would be really valuable. And nobody really knows what the real downside of making big mistakes is because very few people talk about it.

Yeah. I think there's two funny mistakes I made. One where I think it's hard to surmise any sort of learnings. It's just funny, so I'll tell that one first, and then there's one where I think there's some good learnings, and maybe I'll elaborate on that in more depth. The, uh, the first one is that when I was on the ops team, I deleted the geography of San Francisco from the DoorDash database, and so it just shut down for a day. So drivers couldn't drive, people couldn't order food, and restaurants couldn't get orders. And I got And it was one of those things too where you looked in the database and it said like, Daniel Parris made this change. And the— this was also when we were physically in an office in San Francisco with 200 other people. And so it was this thing where all of the like— and then, uh, actually to make it even worse, the way that DoorDash would feed its employees was by placing DoorDash orders for dinner. And so we couldn't place the orders for dinner, so nobody could eat food. It was a huge mess. And so it was kind of randomness. I just like played with some variable in the database, deleted it, and then over some period of time we had to actually like rebuild the, the geography and then ultimately bring all of the restaurants back online. And it was embarrassing. And I think even like the CEO was involved at a certain point asking what happened because he opened his app and it just said, hey, the company you built does not exist where you are.

So the app didn't say Daniel Parris brought the whole thing down.

He did message the channel that was like Northern California operators and he's like, hey, you know what happened? And I had to chime in and explain what happened. Fortunately or unfortunately, in that, in the early days of DoorDash, that stuff happened a lot. I'm not saying it was forgivable, but it was a rite of passage to make mistakes like that just because it's such a complex system. You have these three-sided marketplace for, if you include support. And so there was just, and it was still a startup at the time. And so there, it was built originally as like a college project. So there were all of these different things that you could break. And ultimately, like, you know, I was, it was a very bad day, but I felt fine because I had seen a lot of my other peers make similar mistakes. The one where— the other mistake I made where I actually did learn, and I think there's, there's actually learnings that you can draw from the experience, was I was given the task of trying to make drivers more efficient with their routes. And so if you're a driver, you want to do as many deliveries in a short period of time. And so I came up with this incredible idea that I was going to just cut the radii for all of the restaurants by like 30%. And so the radii is for, for like a restaurant is when you think of a consumer or you think of a restaurant, there's a certain radius, uh, and so a consumer has a radius where they say, okay, if I am in this location, I can see restaurants for 5 miles around me and I can order from restaurants within 5 miles. And so I decided to slash that radius by about 30% because then the drivers wouldn't have to drive as far to drop off the food, and then they could do more deliveries, make more money, and everyone would be happy. We wouldn't, you know, they could make more money and maybe DoorDash would have to pay less over a longer period of time. So I made that change, didn't really tell anybody, uh, I was just like, this fun little project I was working on. And then the entire Northern California region shrunk by about like 20 to 30% because consumers could no longer order restaurants that they were used to ordering. And this was like a totally, uh, again, there was a lot of experimentation in the early days of DoorDash, so people were experimenting with things like Radii all the time. And the sort of like crowning achievement of this failure was that one of the most important customers that DoorDash had in its early days was Andreessen Horowitz. Uh, they loved to order like a giant thing of sushi from this place called Fuki Sushi. Every day, and it was this known order where we always had to make sure we got good quality. And apparently they inbounded and were like, we are no longer able to order from Fuki Sushi. And obviously this went straight to the CEO, who then forwarded it to my manager and said like, what is going on? Why is basically like our best customer angry at us? And so I got pulled into a meeting, I think with the COO and a few other folks, and I had to explain how I just like, you know, for a fun science experiment, just cut everyone's selection by about 20 or 30%. Uh, and obviously it— I got pretty heavily chastised for it. I— my manager, I think, stood up for me at the time and said, hey, he was given this very specific goal of trying to make things more efficient for dashers, and this was his solution. But what I learned is I obviously took a very myopic approach to solving this problem, because DoorDash is a three-sided marketplace. It matters the only way that things actually really work out is if all sides of the marketplace thrive. And so solving a driver problem by doing something that's bad for consumers is not like an altruistic way to actually grow the marketplace. And I think that hyperfocus on like one specific metric and the myopia that came with that ultimately was a huge learning for me. And then throughout the rest of my career, I think one of the things I always try to take into account is if we can pick one metric, How do we pick a metric that incorporates a bunch of different perspectives from— if there's multiple consumer groups or multiple groups that we need to take into account? And how do we make sure we're not being hyperfocused on something that isn't going to yield the result that's best for the company? Yeah, I—

from the outside looking in, it seems like those are pretty big mistakes, right? Yeah. And I, I think it's important for us to sort of Look at those from the company's perspective though. So if you have an employee, they shut down, you know, California for, for a day or San Francisco for a day. What does that say about the systems that you have in place? Should someone on the operations team be able to shut down San Francisco or delete all the geo reference data? Probably not, right? There's a big lesson in operations. That the company can get, and that probably saves them potentially millions of dollars in the future, even though they probably lost a good amount of money during that day. And so to me, a lot of it is like, yes, you made a mistake, is a stupid, dumb thing, but the lesson learned for the company and for you is probably more valuable in the long term than, for example, firing you for, for doing something like that.. And I think a lot of people think that, oh my goodness, I'm the problem. I messed this all up. But if you make a mistake as an employee, or there should be checks and balances in organizations and making the mistake helps the organization create better checks and balances in the future. And I think that that's something that's really important for people to realize is that, you know, especially early in your career. You're actually better off making mistakes because one, people don't expect a ton from you. Two, you get learnings that compound over time. So let's say you did that later in your career. Let's say when you're 26, 27, 3 years later, you wouldn't have had all the knowledge about what you gained from that experience for those 3 years. You probably wouldn't have been as good an employee, as good a worker. You probably wouldn't have evaluated problems in the same way, especially for the second one. And, uh, you know, to me, I, I can't stress enough that your work, your employability— unless you do something that is truly horrific to people, you know, you're like murdering people with a push of a button— most things are recoverable. And they also make really good stories like we're sharing on this podcast, you know what I mean? I also think that makes DoorDash look really good, that they didn't fire you you for these things because they saw you as significantly more valuable than just, you know, one error or, or one mistake ruining an entire good body of work.

Yeah, I know I'm, I'm making myself seem to be the worst employee of all time and DoorDash to be like a fly-by-night company, but so I think it's just an early— to your point, an early feature of DoorDash is that it was actually a really decentralized company. And it was mostly operations, and you had people running markets. And one of the really cool things about that period of time is that everyone would experiment in their own market. And so yes, it seems like I made some really bad mistakes, but I also, through that same spirit of experimentation, found some really awesome stuff. And once somebody— you would hear about all the mistakes that the other operators make and you would share them, but at the same time you would also find out about a really cool mistake— or sorry, the opposite of mistake— really cool breakthrough that somebody made., and then that would be shared. And so that was heavily encouraged. Obviously also these mistakes that I made, like an example of Reddit, they were pretty short-lived, um, and we were monitoring it as an experiment. It wasn't something where we just made the change and then totally forgot about it. But yeah, that was— experimentation was just a huge component. And I think it was one of the things that made the company special in that period of time was you have to be willing to fail. Uh, and actually one thing to That was a huge thing that Tony Xu, the CEO, would preach was, "It's fine if you fail." I'm not sure if this is an actual quote, but I'm just sort of paraphrasing his collectivism. "It's fine if you fail, but then you have to broadcast it." He was really big on making sure that people would talk about when they did something bad and that we could all learn from that. Because it was really— I think the worst thing would be if somebody made a mistake and then just never told anybody about it. And so it was just a great time to be able to play in this pretty substantial sandbox with enough data to be able to have this continuous feedback loop so you could learn and iterate, but at the same time have a little bit of the freedom to fail. And then yes, for somebody who is early in their career, I was able to learn some very valuable lessons and I'm all the more better for that. And when I started managing folks, whenever they would have a bad day, I would tell them about other terrible things that I had done and then they would feel a little less bad. But even on a smaller scale, Obviously, most people aren't going to shut down an entire market or make Marc Andreessen personally angry, but you just know these things, someday you'll look back on it and it will seem kind of funny. And so it will pass, you will learn, you'll be okay. People really don't get instantaneously fired in tech. And if you make a mistake and you don't learn from it and you don't grow from it, you, you could potentially be fired. But if you make a mistake and you're on top of it and you understand what you did wrong, then you will be all the better for it. And showing that you actually have control and mastery and ownership of what's going on, I think, is ultimately more important than some random mistake that you might have made in the moment.

Yeah, I, I can't think of a time where when I took accountability for a mistake I made, it went badly for me. I would argue that when I've made a mistake and I've taken accountability and did the right things to fix it, it's actually made me look better than when I didn't make the mistake in the first place. You know, a lot of the things that I read, they talk about— I actually read a book recently called Die with Zero. It's something I highly recommend. Our other mutual friend, Jeff Lee, recommended that one to me. And he talks about how early on you should be taking a lot of calculated risks because they have disproportionate payoff in the long run, you know, all of the things you described, you could sort of understand what the downside was. It wasn't, it wasn't something that was completely risky compared to what the upside would be for the company or whatever it might be. And as long as that calculation is there, as long as you can articulate that, most people understand that in order to make progress, you have to risk a little bit. You have to be able to Understand what the downside is of an experimentation that it doesn't work, or, or you lose a little money in the short term to get longer term returns. And that's something I think a lot of people can apply to their careers, to their lives. I mean, I, I've been applying that a lot, whether it's experimenting with new content, experimenting with, with different models in my work, because how else are we gonna make really good aggressive progress if we're not doing something fundamentally different than than what we were just doing before. That sounds really boring to me where we're just trying to do the exact same thing over and over again. And to me, that sounds like a very exciting company culture to begin with. Can you talk a little bit more about maybe some of the working in an exciting experimental startup at the time and maybe some of the fun and interesting things that happened?

I guess what direction would you like to go with this? Because it— I mean, especially when we were in person, it was, it was just crazy. Uh, I mean, the company I think was doubling or something in size, like in the actual order volume every year. And so there were just so many different facets of the experience. I always say that it felt like I worked at 15 different companies because I got to see it in its startup phase, and then I actually now even still consult for them and I see it as a public company that is much bigger and is ultimately trying to navigate like life on public markets. So I guess, is there any specific area you would like to dig into?

Uh, maybe, maybe associated with growth and scale and, and work, or anything that you think is particularly interesting associated with that, like managing a, a quickly growing company? Like, what are some of the, the growing pains or the funny things that happen associated with that? Obviously, we talked about deactivating San Francisco. I think you'd mentioned offline you sent an email to, to quite a large number of people accidentally, or things along those lines. But it doesn't have to be those, just like company-wise. I would imagine that there were a lot of large changes that happened relatively quickly, and those often have unexpected consequences across the board.

Yeah, I mean So I've— there's, there's definitely some humorous stories, and I think there will always be. The thing that I guess I found to be the most impressive as the company reached a certain scale, I guess taking a bit of a different direction, was its commitment to experimentation, particularly like building out— and this is getting a bit more technical— but building out an experimentation framework and experimentation platform at scale and making a commitment to that. So one of the things that's really difficult about a company like DoorDash is that you have no idea if you made a change, did the thing that you made a change to actually make a difference? Because there's a lot going on in a given day for the company. If it rains, as an example, like the company will grow, supply will suffer. And so you'll have no idea if you happen to run a free delivery promotion that day, was that the thing that made it grow or was it the rain? And so making, being on the data science team and ultimately building out that commitment to experimentation, saying, hey, we're running an experiment. We forced ourselves to send an email at the onset of an experiment that said, hey, we're running an experiment. And then you had to send an email at the end that said whether it was a success or failure. Building out that muscle and ultimately getting the company, it got to the point too where any initiative, like a product initiative, an ops initiative, had to go through the lens of experimentation. There was effectively a gatekeeper. So one of the things that really always impressed me, just looking back on the company at that time, I thought was really special was its commitment to understanding the ground truth and having that be a gatekeeper for whether it actually rolled out a certain feature or took a strategic path. And I found during that time, as we were growing, we could have made a ton of mistakes. But the fact that there was an experimental rigor to the way that we ascertained whether a strategic change actually manifested what we wanted, I think was a huge strategic advantage relative to our competitors.

I think that that is embodying, it seems like that's what embodies a true data-driven organization. I've been in a lot of companies or I've talked with a lot of people where their companies are making the decisions and then they want data to back it up. And that's kind of the opposite of the approach that I think you should be taking. And it's refreshing to hear about organizations that are truly living data first. And I think that that's probably more normal in companies that are newer. I know that companies that have been around a long time, particularly non-tech companies, for example, I worked at GE for a while, they don't quite do that as well. And there, it's a lot of confirmation bias rather than focusing on ground truth. And also data can be unbelievably biased based on how you've made decisions in the past. So if we have a methodology for, you know, assigning drivers that was done by ops people and humans, and then we build a model using the labels that humans created historically, it's not a really good model because you're just replicating what the humans were doing when they were making those matches or whatever it might be. And so I really like that philosophy. And I think that that's something a lot of people can take away with. How they do anything, how they apply to jobs, how they approach their own work, or they can evaluate if their companies are making decisions like that as well. You know, I am interested, I mean, obviously you're consulting for them now, but you decided to part ways, you know, what went into that decision that you wanted to pursue more of your own things and, you know, how did you navigate that?

Yeah, I mean, I'd been there for 6 years. And so I just felt like my time, a lot of my friends were starting to leave. I don't think there's something inherently wrong with a larger company. I think it's always just a series of trade-offs, but I had joined a company at a very specific moment in time when it was a startup and I was deactivating entire geographies. And when I was getting ready to leave, it no longer felt like the same company. And so I also was just looking for new problems. I really wanted to go out on my own and start, especially I think one thing that was, I wouldn't say tough because I actually did enjoy being a manager, but when you get into a bigger company, the things that spell success are usually around how well you can maneuver through an organization. And I was finding, especially going down the management route, that I was A lot of what defined success for me was about if I could, you know, if I was a good operational— or sorry, if I was good at, you know, maneuvering through the organization versus whether I was, you know, running this experiment that was going to give us an insight into how we should think about selection, or running an experiment or figuring out, oh, here's an insight that no one's ever found before and this is gold for us and I'm going to disseminate it to all the other teams. And, you know, those, those really wonderful lightbulb moments that I loved at the beginning. So I sort of decided that it was time to go, and then I ended up, I guess, going out on my own. I don't know if you want me to talk about like what I'm currently doing. Yeah, yeah, please do. Okay. Yeah, I mean, I had always wanted to understand, I guess, like what it would like be like as a consultant. You know, a lot of the way that I worked at DoorDash was in a sort of consulting manner. You— there was a centralized data science team, and then ultimately you would be assigned to projects. And so it actually wasn't much of a leap to imagine what life would look like where I was doing consulting work outside of basically being on my own rather than being in a corporation. So since leaving, I've just been doing consulting work, and then I've started, I guess, like a few newsletters. Uh, one is a newsletter called Stat Significant, where I finally— I've always found that I can't really fully quit film and my love of pop culture. And so it was this fusion of this data science skill set I have and this interest in statistics. But what if I combine that with elements of pop culture that I also am just endlessly fascinated by? So some of the things I've written are, you know, why do people hate Nickelback? And can I find a statistical reason why people hate Nickelback? And actually, there was some crazy stuff that I found. But I like more than just that they are bad. That was an interesting find looking at the career of Nicolas Cage and just seeing if you tracked it via data, is it as crazy as Nicolas Cage himself? And the answer is yes, his career is bonkers. And so just being able to, I think, sit down in a given day and write about what I'm interested in, combining all of these disparate skill sets and these interests that I've had throughout my career. And I've actually found that people have been interested in that work, whether it's to co-syndicate what I've written or having me write research content for them, like commissioned pieces. And so I've been able to actually carve out work through that, and that's, that's been an absolute joy because it's combining all of these things that I love. And being able to write and being able to actually produce content that other people find interesting is, is a great thing. Um, and then yeah, I consult DoorDash and a few other companies and I enjoy that work. It's just nice to be assigned a few different problems and to be able to bounce from thing to thing, and it's always testing different sides of my brain. So I enjoy the consulting life. And then recently launching a new newsletter called Data People, which is interviews, transcribed interviews with world-class data professionals. And that's just fun too. I think I enjoy talking with smart people. Who doesn't? And can you take a lot of the insights from those conversations and put it into a different form factor and basically build bits of microlearning for people. So if you could take the collective wisdom of an accomplished AI researcher, or you could take the wisdom of somebody who's in the data governance field, and you could condense it into 4 to 5 minutes that somebody can read so that they can understand what that person does and what are the ways that I can learn more about it, that's effectively what the purpose is of that newsletter. So It's definitely been a fun road. It's one of those things sometimes where my wife has to ask me like, what did you work on today? And I'll have to explain, well, I worked on this thing and then I like worked on that thing. I definitely am scattered across a few different things, but unfortunately I do like having variety. It's just I have to be very rigorous about how I plan my work and make sure that I'm not scattered between too many different things.

Yeah, I mean, the path that you're taking, it seems like it's becoming a lot more common and more generally, I guess, like societally accepted where you can work on a bunch of things. You can pursue a little bit of passion, but also autonomy and freedom. And tho— those are things that I'm very excited about. You know, it's, it's funny. One of my favorite concepts is Dunbar's number. So if you're not familiar, you know, it's right around 150 and that's, the, the most people that people can keep in their active working memory or keep in good contact with, whatever it might be. And something sort of magical happens when companies pass that threshold. So you got in at DoorDash, it seems like around that number. And as it continues to grow, the incentive system starts to change. So people go from trying to do what creates the most value for the company to focusing on, oh, what gets me promoted more effectively? It becomes less about alignment around a value or a goal and more towards individual pursuits that you were talking about a little bit with, you know, playing a little bit more politics, or there's things that are outside of the value you're creating that go into promotion decisions. And I see consulting, contracting, the newsletters you're producing is very fundamentally going the opposite direction where you have complete control over your own vision and your own goals. And for me at least, that's something that's very refreshing. I find it very difficult to not necessarily see the fruits of my labor. And in what you're describing, you can directly see the impact of everything you're creating. And I, and at least for me, again, that's something I feel very powerful. I feel is very powerful. Um, and I'm, I'm really excited to, to see what comes out of these. Again, I obviously, I was honored to be on the, on one of the early data people or to be one of the early data people, technically.

Yeah, I think you're going to be the first data person. Oh, wow. What an honor.

You said you had a lot of smart people, so, you know. Yes, there's a—

That might be an overstatement for me, but— I mean, chronologically, you will be the first installment of data people, not you're like the top.

You're not at the top. No, no, no, no. That would never be the case. I was gonna say in aggregate, I'm sure you'll have a lot, a lot sharper, hopefully less interesting than me though. You know, that's the, that's what I got going for me. But, but no, you know, to me, I'm interested not only in the upside of the consulting and contracting. I mean, that's something that's important to me and I talk maybe too much about from going off. More on your own like this, maybe what are some of the downsides to doing that? Like, what do you have to deal with on that front?

Yeah, I mean, it's one of these things where you're your own boss. When you're your own boss, like you are your own boss, like it's sort of, there's, it's a two-sided coin. And so I would say some of the things that I found is you don't get as much, I mean, it's obviously different in remote work, but one of the, think about all of these different social networks you typically, the typical person has in their life. There's family. Maybe there's friends that they've made through college or through other community groups that they're a part of, sports teams, etc. And then there's work. And so when you decide to go out on your own, you effectively forego— unless you're, unless you're able to build up that community on your own terms through your consulting practice, at least when you start, like, you lose that social stimulation. So many of my days are just me on my own. Obviously when I'm working with clients, I talk to them, but that's as a client, that's not necessarily as a peer. So you lose that peer-to-peer relationship that a lot of people can get working at a company. And so I would say that's, that's one thing. And then the other thing is you're not necessarily gonna be the first thought on somebody's mind. I've had people reach out to me who are excited for my consulting services and then ghost me, and it's very confusing. Because I did not ask for them to— I did not ask for them to solicit my business. And at the same time, there is then a hypothetical amount of money that I think I am entitled to that is now gone. And so there's the Daniel Kahneman theory of loss aversion. And so throughout that whole interaction, you kind of get built up and brought down simultaneously. So you just have to, you know, again, it's pros and cons. The pros are you do have that independence. You do get to make your own decisions. You can really build something and feel that ownership over a long period of time and focus more on the things that you find interesting. But at the same time, you might not have some of the structure and some of the community that's built into a traditional 9-to-5 job.

Yeah, I would also add there are two other benefits where you can diversify more effectively. So if you have one job that's making all your money and you lose it, you're kind of screwed. If you have 10 consulting clients, probably not gonna lose all of them at the same time unless you get canceled on the internet or something along those lines.

Yeah.

And I, I think, well, I completely forgot what my other benefit was, but that, that to me is, is, is really a major one that, that I've been harping on a lot lately. I mean, as we move into, no, and it, the unlimited upside is also another one where in theory you could, you could have 20 clients at the same time. You could have 30, you might have to scale or do whatever it might be. But that is a fundamental difference, especially as we move into more uncertain times that I think is valuable. You can move quickly to new opportunities, whereas if you lose a job, if you're looking for another traditional job, you have to go back and interview and go through this whole process, which is unbelievably time-consuming. Whereas if you're network and you've built a funnel, these types of things are in my mind, significantly more resilient, although less consistent. But I'm again, really excited about the newsletter. I'm excited about data people, and I'd also love to be able to share more of your stories and help people connect with you. What are the best places to learn more and also to potentially reach out and ask you questions if that's something you're open to.

Yeah, I'm absolutely open to that. I guess I will— you can always reach out to me on LinkedIn. I believe my name is Daniel Parris on LinkedIn, so it's a picture of me sitting at the stump of a tree looking off into the distance. It's a really pretentious LinkedIn photo, but I think it's pretty solid. And then you can email me at— I've got too many email addresses. That's the other thing when you're starting different sort of entrepreneurial ventures. I now have like 4 email addresses. So maybe the best one is daniel@askdatapeople.com. You can shoot me an email, I'll respond. Maybe just say that you listened from, you know, you found out on Ken's podcast versus, so I know that you're not spam. But yeah, I'm happy to chat about anything. I don't know, I feel like I've lived a lot of data life and so I'm happy to just chat about whatever's on somebody's mind.

Amazing. I'll also, in the description as well as the show notes, I'll link all of the resources for your newsletter for data people when it does come out and your LinkedIn and other relevant resources. So Danny, those are all my questions. Do you have any final parting words of advice for our listeners here?

Yeah, well, I guess we spent a lot of time dwelling on all the things I've done wrong in my career. I'm actually perfectly fine with that. I feel like I lead with those stories anyway, so I'm not shy about talking about them. There's a quote from a Bruce Springsteen song that I like. I think I actually said it colloquially throughout this conversation, but it's in this song called Rosalita, and he says, someday we'll look back on this and it will all seem funny. And that's sort of like a way that I think about things throughout my career is whenever I'm having a bad day or I've done something wrong, I try to take a step back and think, is the thing that I'm upset about right now going to matter to me in 10 years, in 20 years? And if the answer is no, which it almost certainly is pretty much every time, that's always a weight off of my shoulders. And so yes, you're probably living, you know, your day-to-day working on projects, making mistakes, feeling like you had a tense interaction with your manager. It's always important just to think about, is this thing ultimately going to be the single thing that determines success for the rest of my career? Most of the time it's usually no. And so you're— you, as long as you learn and you pick yourself up and you don't let it become ingrained as a part of you and you don't, you know, turn inward and become a bit negative, you can grow from it and you can just keep moving on and everything will most likely be okay.

I love that so much. I think that without a doubt, almost universally, we all learn more from the things that we did poorly than we did from our successes. They're also unfortunately more memorable, but they're probably more memorable to us because that lasting impact also helps us to course correct over time. You know, if I made the same mistake twice, and I've, trust me, I've made my fair share of mistakes, but fortunately I made many of them early, even before I entered the workforce. If those weren't still with me, I probably would've done some really stupid stuff that would've probably cost me money, cost my clients money, whatever it might be. And you know, again, I'm really grateful and appreciative that you're candid and willing to share some of the low parts of your career so that many people can see, wow, this person who came on is very successful in spite of any of the things that might've happened, but they can also sort of learn from the things that you did and take those lessons away so they don't have to go through the, the, the pain and suffering that we went through in our, in our early mistakes to avoid that. So again, I am super grateful for you being willing to, to share those things with us. And I know that, uh, the, I'm confident this one will have a, a lasting impact on, on everyone listening.

Awesome.