This is the transcript of my conversation on @FromSourcePod with Annie Flippo, a Head of Data Science and a robot enthusiast. We talk about her job managing a team of data scientists, from employee retention to vendor relations.
This has been edited for clarity.
Michelle: Annie, can you tell us your current job title and how long you've had that job?
Annie: Yes, I have two job titles. In my organizational chart, it’s Manager of Analytics but really, I’m the Head of Data Science. I've been here for about a year and a half.
Michelle: What does an average day look like for you?
Annie: I don't really have an average day. Some days are very meeting oriented where I do a lot of planning with the stakeholders. During those days half or most of my days are in meetings and trying to get the requirements of what they would like us to work on. Other days are more heavily on working on insights, building models, building a self-service platform for our in-house sales team.
Michelle: What has been your favorite long term project?
Annie: Where I work now is an AdTech company and we work with location data. We wanted a way to target advertising to people in a more succinct way so we're building a persona product. Not a persona like soccer moms, but it's a persona based on location. We want to know where people have visited in the physical world and we call that geotypes. That has been a long time coming. Trying to get the data, vet out all the different data vendors and finding a day to make this product, it's been about a year in the making. Finally we launched it about a month back so it's a very long term project and now the sales team is trying to figure out how to sell it. Now it's more of an educational endeavor for us and our clients and also our sales team.
Michelle: Do you think what made it exciting was the novelty of the project or because it's been in the works for so long?
Annie: I'm excited because there's nothing like that in the market place currently and I think location intelligence is very big right now, it’s the next big thing in digital marketing. I'm very happy to have been part of it and learned all the pitfalls and all the benefits. It's been in planning a long time and I worked on it for a very long time and now we actually have a physical product. We did have several vertical as a POC (proof of concept). We did grocery, restaurant and retail and this way we can try to test the market to see if our clients are interested.
Michelle: So I go to Starbucks pretty much every day in the afternoon, would my persona be the woman who walked to Starbucks every day and then at one o'clock you'd be advertising Starbucks to me?
Annie: No, well, so we do this anonymously so we don’t actually know who you are but track your visits over time. We can see maybe you’re a mom because you go to schools a lot or maybe you are an Uber driver because you know you go to lots of different places and go to airports and things like that. Based on where you visit, we can group people on their commonalities. For example, we have one that's called affluent savers. This is someone who's actually affluent but they like to save money because they go to discount retail stores a lot such as the DollarTree. It’s not typically what you think of when you think affluent people. You’d think they would go to very high-end stores, like jewelry stores or expensive restaurants. But there are some segments that are not like that. So we can’t treat everybody the same, instead, we group people based on where they're actually going.
Our advertisers are interested in millennial singles who go out at night or maybe folks who have kids, it all depends on the use case.
Michelle: It's interesting how you were able to use the data to actually dispel myths and find groups of people that you can't just think up. You wouldn't normally think of an affluent person going to the dollar store, but there are plenty of people who are affluent because they go to the dollar store or because they're penny pinchers, but that’s not the first thing that comes to mind. What is the most boring but essential part of your job?
Annie: That is a large part of my job. It’s working with vendors. A vendor will have an unexpected change in the data format and not tell us in advance. This happens quite a lot. Our ingestion jobs break and then we have to go chase down the sales rep or accounts manager. “Hey, you didn't give us notice so you can't just change that.”
Then we go going to a big meeting to figure out what happened. You know sometimes the accounts manager didn't even know that was changing. Someone behind the scenes in the engineering teams made some decision because of performance or whatever. So the data changes a lot and you can never rely on data staying in the same form at all times. That is essential to the job because without data we can’t do anything. We have to have a steady stream of reliable data and then you have to do a lot of vendor management.
Michelle: Do you clean up data, or is that more your date engineers on your team?
Annie: That's more the data engineers. The raw data that comes in is never in a format that you can just right away analyze. You have to create more features, clean up the data. you Dates alone and timestamps have like fifty ways to clean that up. You have to figure out what makes sense, how you do it and what is a reproducible way. You have to understand what's coming through the pipe. Server logs always come in UTC but we don't in UTC. Let's have a meeting at eight o'clock UTC time, nobody talks like that. It is always based on local time. People have lunch at noon at their local time, people go to a movie theater on their Saturday nights. So we have to understand what is the local time, what's happening in that region.
Michelle: That sounds like a frustrating part of any technologist’s jobs. Time zones and time stamps and everything that goes with it. Would you say that's the most stressful part of your job?
Annie: The most stressful part of the job is my high performing team members giving notice. You have a good team and everything is working well, but it is such a hot market right now. Recruiters are always calling them, whether they are looking or not. They'll contact them through LinkedIn or through a friend of a friend and then they’ll say, “Hey we have this great job opportunity would you be interested in knowing more?“ At first, they say no, not really, but then they hear a few of these and think maybe I'll chat with them it can’t hurt. That’s how you lose people and it's hard to find another person because the job market is so hot. It takes a really long time to get somebody new coming in and get them up to speed. Every time you lose somebody it's more like a three to six month process to get an equivalent person back.
Michelle: How do you manage the stress?
Annie: Well I try not to focus on it because you can't really focus on people leaving. All I can do is try to give them interesting projects and mentor them as much as possible. Some of them are Data Analysts/Business Analysts that want to become more Data Scientists. I always give them suggestions on projects they can do outside of work. Then we talk in the one-on-one on how they can improve. We slowly build up a rapport and sometimes I even help them with their side projects.
Your job is your job. There's an ebb and flow, exciting parts of your job and boring parts of your job. You can't be a hundred percent exciting all the time. I don't try to think about it too much. I just hope that I'm giving them the best experience while they're here and then if they go it’s because they had. It’s because they had found something different and interesting. That’s why I go to a lot of meet-ups and I help out with UCLA. I like to know a lot of students who are graduating, about to graduate or have graduated recently so I have a resource to tap into.
Michelle: It sounds like It's actually a very positive experience because you're helping others grow, whether it's at your company or outside your company. You’re making sure they’re always learning new things, which is very important in tech.
Annie: Yeah, it is very important in tech. I think that if you don't grow your career is pretty much over. Right now everything is moving so fast. The minute you come out of school, within two years you'll be behind again. You always have to learn and I do that myself as well.
Michelle: Which skills you find the most essential on a day to day basis?
Annie: A couple of skills. One is a willingness to learn, just like I said earlier. The second skill is empathy and that is the hardest one to learn. It is also one of the hardest ones to interview for because you have to be able to see from the other person's perspective. Let's say your executive wants a specific product report and it's a really difficult thing to do and you're trying to understand why they are asking you. You have to go down the road and ask how can I help you find a solution to your problem. So those two things are the hardest and most essential skills. All the other skills are much easier to learn. You can find an online class or a YouTube channel, there are many resources to learn.
Michelle: Have you ever tried to teach someone empathy and has that worked at all?
Annie: (laughing) I think I always try. I don't really teach, I just say what do you think those guys are thinking about, why are they asking this? I like to pose more questions back to the person who's having issues or having a difficult time. I hope that they take my suggestions and go back and ask the right questions. Hopefully over time through osmosis they would start thinking, I wonder what that person's thinking about, what are they trying to solve, why are they asking me.
Michelle: One of the things I wanted to highlight on this podcast is the soft skills. They are so hard to learn and people who do not feel like they have enough technical skills but are skilled in the soft skills can really shine. By being willing to learn and jumping into all the technical stuff as soon as possible, having that empathy really puts you a step ahead.
Annie: Yeah I think so. It’s nothing you can learn in college. I mean you might learn it through your friends, but there is no college course in empathy. I wish there was!
Michelle: If someone wanted to be you one day, what would the ideal career path be like?
Annie: I don't know, I kinda meandered through my career. I had a thought in college that I wanted to become an economist, so I studied math and econ together. I wanted to become an economist because I wanted to know how do you effectively change people’s decision-making process. Like increasing price, incentives, increasing demand, supply, etc. It is just happenstance that now there's a field called data science. Which, in the marketing field, is all about how do you entice people to do something. Click on the ad, look at the product, you're trying to help them make that transaction. So it’s really aligned to what I originally did in school but my path went through software engineering for many years; then product manager for many years and then the last ten to fifteen years is basically all data. So what I would advise is to be very curious about how to solve problems with data and then everything else will fall in line. I believe in just in time learning.
Michelle: It sounds like you had a varied experience and it all enhanced your career by learning different things.
Annie: Yes. I don't know that it was a planned thing, but it just so happened that I fell in with a professor right out of school where I did a lot of like financial programming. That taught me a lot. Then it was just making sure the next job was always a little bit like the last one but built on top of it until I landed at my current job.
Michelle: For the college student or career changer, what’s the entry-level job on your path? Is it a data engineer or something else where they can get started with a limited skill set?
Annie: I think probably not data engineering because with data engineering you need software engineering skills and have to understand data architecture, which comes with time and experience. I think the entry-level job would be data analysts or business analyst. I use them interchangeably, it depends on the company's size. As an analyst, you are given data that's already pre-formatted and you use that to find metrics. This way you will learn to write a report, learn some SQL, maybe get your hands on a business intelligence tool like Periscope, Tableau or Looker. After that, you could start learning more about the math, stats, and machine learning methods that come with any data science job. But definitely, data engineering jobs are now split out into their own specialty. Six or seven years ago everybody's doing everything at once but now it's all split up.
Michelle: It’s fascinating how things are changing because I thought I was data engineering, but it's not. I'm learning here today as well. Do you think if someone started as an analyst they'd be able to work their way up by learning on the job or do you think there's other courses or anything else you'd recommend learning aside from on the job?
Annie: Well if they want to be a well-rounded data scientist, they will need to learn something outside the job. On the job, you’ll be solving very specific problems and you might not get introduced to a lot of different methods or problem sets or other ways of solving things. I recommend to everybody who asks me this to try the free online classes first, so you get a feel for if you feel like it before you spend a lot of money. Learning things like linear regressions, random forests and support vector machines. Learn one of those things and how they apply to different problems. Learn how to apply different models and what's the pros and cons. All those are things that you don’t really encounter in a work situation. So you can’t learn everything at work and you have to supplement it.
Michelle: What's your next step?
Annie: I would like to take on more responsibility and build more products. I would like to manage larger teams and have more of a strategic say. Do what I’m doing, but on a bigger scale.
Michelle: How can you be reached on social media?
Follow @FromSourcePod for more episodes and to continue the conversation about what tech jobs are really like, from the good, the bad, to the boring.
Rate & review to support the show!