Democratizing Data Science: Community Spotlight with Avo’s Stefania Olafsdottir Ashley Dotterweich
Welcome to the Heavybit Community Spotlight Series, where we highlight the people that make up the Heavybit community, and the interesting projects and ideas that they’re exploring right now. This week we’re turning our spotlight to Stefania Olafsdottir, CEO and co-founder of Avo. We chatted with Stef about why she’s passionate about empowering teams to use data more effectively, how the role of the data science teams is evolving, and what she’s doing to adapt to life under COVID.
Tell us a bit about your background. What led you to start Avo?
I am a mathematician and a philosopher who went into software development and data science and statistics research in the genetics industry. I was one of the few people in Iceland with experience working with huge data sets when an Icelandic startup called QuizUp blew up and reached a million users in one week, which was the fastest growing app in the App Store.
I joined there as the first data person and learned that to build a data culture and power data-driven product development, it’s a twofold issue: it’s cultural, and it’s technical. The culture part is helping everyone ask and answer the right questions to move our product in the right direction. The technical side is building the infrastructure that we need so everyone has the answers to their questions at the tip of their fingers. At QuizUp, I learned a lot about these things and we invented a lot of wheels to solve the problem internally, like collaboration software and data validation tools.
Fast forward a few years later and we realized we had to build these tools again at another company that I founded with my friends. We discovered that all of the leading data-driven organizations end up building these things internally. It was so frustrating that we decided to build these tools as a product to solve it for more people.
It sounds like you came into the tech industry because you were the right person for a particular job at the time — did you think that moving into tech would be a temporary career move?
The genetics company I was working for — who is now testing everyone in Iceland for COVID — sponsored me to do a masters and I was invited to do a PhD in bioinformatics — a mix of statistics, algorithms and computer programming. I was on that path, but then I got an offer from QuizUp. I figured those PhDs would always be there, but how often would I have a chance to build up a data science team from scratch?
Then I got very excited about the cultural aspect of helping people be data-driven and making data-driven decisions. I never decided “I am now in tech.” I continued to do some classes on the side, but I dove very deep into the tech work culture.
Was that when you decided you wanted to change the culture around data in tech? When was the “aha” moment when you realized that there was an opportunity to make an impact on how companies use their data?
I didn’t realize how early QuizUp had been in building internal data tools. I think it’s related to the fact that the CTO and I were a couple (we’re married today). The competitive advantage we accidentally created is that we would come home and complain about our days, and we were coming at this problem of data quality and engineering efficiency from two completely different angles. We realized there was an opportunity to change the way developers were working and increase the data quality immensely if we built a tool that helps these teams collaborate better. So we did that to solve each other’s pain points. I think in most companies the data team doesn’t work so closely with developers.
When QuizUp was acquired, I didn’t want to work for another tech company, I was done. I didn’t want to be head of data science anywhere else. We had to build too many tools and enact too many cultural shifts. I took a sabbatical, and that’s when a few friends and I from QuizUp decided to found another company (not Avo).
I told them, “I’m not going to be responsible for the data science part of this product.” And then about four or five months into the product, I realized we had shipped af feature based on incorrect data. I get shivers just thinking about that moment. I knew that as long as I was in the digital product space, working with data would be so painful because there was no way for us to do that reliably.
We kept building that product, but I kept talking to people in the space — I talked to Twitch, I talked to Spotify, I talked to Airbnb. And all of them had very recently built data tools like QuizUp had built. I realized that this was a huge challenge everywhere, and that I had to do something about it. That was really the drive.
What changes are you seeing in the data science field right now?
There are two major shifts happening right now that have been driving change in the data science space. There’s been a culture shift with an expectation of easy access to information. Every person in a company should be able to answer their questions quickly, without having to rely on a data expert. Until recently, we would have isolated business intelligence teams who would deliver static reports to stakeholders a couple of days or weeks after it was requested.
This is not the case anymore. Every person in the company is expected to be able to answer questions super fast, so that they can act quickly and build better experiences for their customers.
There’s a myriad of tools available to solve this problem, and the role of the data team has gone from being an isolated BI team that generates reports, to being a team that’s integrated into the entire company and supports a self-serve analytics culture. Making sure that stakeholders have easy access to analytics infrastructure, helping everyone ask the right questions, get the right answers, and make decisions based on data they can trust.
The other shift is a technical change. We now need to understand the entire user journey. We used to just be taking snapshots of some operational database — maybe every week we would count how many users we had. But now every user interaction is logged, and we create these holistic user journeys. But the prerequisites to be able to use that data requires us to build infrastructure and processes that make this immense volume of data reliable, relevant and transparent.
It sounds like we’re at the beginning of a revolution in the world of data science. Why is it so important for teams to invest in data science?
The shift towards data culture is behind why we’re building the product we’re building at Avo. The situation in the world is that product teams have never had to understand their users better or faster. There’s a digital gold rush going on — everything is being digitized, and consumers have multiple options for digital products. So we’re able to jump from one product to another very easily. And if those products don’t provide us with the best user experience, we’ll leave. THat’s why product analytics and experience have become such a fundamental part of being a successful business today.
We’re seeing that especially now with COVID — the tools that will win the market are the ones that people will love the most. This has put an immense pressure on product teams to understand their customers.
Who are the people in the data science field that inspire you most? Who is doing something interesting with data analytics right now?
When I was a practitioner of data science at QuizUp, I very much looked up to Hilary Mason of FastForward Labs. I was navigating how to build a data culture, and I had such a difficult time choosing between spending my time on technical stuff — building infrastructure, writing bash scripts — or spending my time consulting within the company and answering questions. Hilary was one of the first people I saw give a lecture on the importance of mixing those two. She really put a dent in how I was thinking about things.
I really like the analytics team at BuzzFeed. Their publisher Dao Nyguen has a really inspiring analytics policy for their editorial team, and how they decide what content to prioritize and when to release it.
I’ve also always kept an eye on how Airbnb has approached data science; they were a big inspiration for us at QuizUp. They were doing similar things to us, but at a much bigger scale. It was always a good confirmation for us that we’re doing things that make sense — if it makes sense for Airbnb, then it makes sense for us.
The Avo team is a split between Iceland and San Francisco. How has being remote impacted the team?
We’ve been doing remote and asynchronous-first since last November, but what I’ve found as an unexpected side effect of COVID is that it’s been a great forcing function to do it really well. Even though we were async-first, we had 3 offices and there were people working in these offices together. But now no one has access to these offices, and everyone knows what it’s like to not have the right documentation or to be able to talk to other people right in the moment.
The rabbit hole that I’m digging myself into is working with the Iceland team until the afternoon, and then working with the San Francisco team until late. I think that’s the danger of being remote and working across two time zones.
What have you been doing in your down time?
Growing our houseplants, and baking sourdough!
Because I’m spending so much time at home, I’ve become obsessed with making my home a really nice place. I recently did the MasterClass on interior design, and the instructor said “your home and your spaces around you are so much about making you feel happy in them, so of course you should invest in them.” I’ve often not prioritized making my surroundings beautiful, but it really nourishes your soul every day.
Are there any habits that you’ve picked up during COVID that you’d like to continue doing in the future?
I’ve tried to keep a routine a little bit better. It doesn’t happen organically because you’re not going to the office or the gym. I start my lunch break with some exercise, then grab a quick lunch and get back to work. In the afternoon I keep my schedule clear for sales calls.