The Right Track
Podcast

For people building data cultures.

We’ll hear from leaders in data, engineering, and product, as they share frustrating and inspiring stories on building the best products for their customers.

Hosted by Stefania Olafsdottir.

Building Relationships in Data with Emilie Schario of Amplify Partners, Netlify and GitLab

Emilie Schario was Gitlab’s first data analyst (huge) and eventually became Chief of Staff to Sid Sijbrandij, Gitlab’s CEO. She then joined Netlify as Director of Data and Business Intelligence and built an incredibly inspiring data culture there, which we heard about from Laurie Voss @ Netlify in a previous episode. She’s now a Data Strategist in Residence at Amplify Partners, helping lucky companies build out their data strategy, data team, and data culture.

Emilie was member #50 in dbt’s Slack community back in 2016, and she’s an admin in the Locally Optimistic Slack community. These are two valuable data communities I recommend. Her contribution there are a testament to her dedication to the data community, and so are the tons of valuable insights she’s shared throughs talks and articles. 

To put it mildly, I’m very excited to have Emilie on The Right Track. We covered a lot of ground, with a strong theme around data cultures:

• Is “data science” a helpful title? We discussed one of my favorite talk, Down with “Data Science” talk at dbt’s Coalesce 2021, where she argued for why “data science” isn’t a helpful reference to a role for data professionals, neither for clarifying what they do in their job, or as a title they can build a career path on.

• Data teams—from “number fetchers” to strategic partners: I loved and related to her thesis that the biggest shift in the industry is how data teams are evolving away from being “IT” and “number fetchers”, to being a strategic part of the organization. But accomplishing that takes work. Emilie shared actionable advice on how to get your data team to that strategic partner level. Including how she introduced so-called “insights time” at Netlify (also covered in episode 6 with Laurie Voss), where every data team member would carve out time for proactive data research.

• “Relationships first, numbers second”: A repeated theme in our conversation was the importance of relationship building and identifying partners across the organization. This is true both when building a data team and when building your individual data role. Organizational data trust and data quality also depends on relationship building; on bringing data consumers and data producers closer together. For example for the process of managing analytics releases, that is, producing product analytics events with every single product update; that really requires a tight collaboration between product, data and engineering. In her words: “Relationships first, numbers second”.

Listen in for Emilie’s thoughts on strategy and tactics for making your data team a strategic part of your organization, and enjoy the ride as she sheds a light on the challenges and joys of being a data professional. As she put it: “Data is hard, but sooo fuuun”.

Defining North Star Metrics with Che Sharma CEO & Founder of Eppo

Che Sharma is Founder and CEO of Eppo, a platform for teams to run better experiments faster. He was one of the first data scientist on the Airbnb data team, where he worked on anything that needed to get done to scale their data culture—including production machine learning, analytics infrastructure, actual analytics, and data tools. A special shout out to the infamous Airbnb Knowledge Repo, which Che built and open sourced. He was also the 2nd data scientist at Webflow. So he’s truly seen the creation and rise of successful data teams. He’s now the Founder and CEO of Eppo, a very exciting platform for teams to run better experiments faster—even when they aren’t the Airbnb data team.

As Head of Data Science at QuizUp, the Airbnb data blog was an inspiration for me personally; an important mirror to validate that we, a small team from Iceland, were actually doing pretty good stuff. So it was a great pleasure to have Che on The Right Track.

We discussed who should own your org’s northstar metric, what defines a good northstar metric, and how good metrics empower teams to be proactive, to think outside the box, and make quality decisions. 

We also talked about how vastly underappreciated the data lifecycle is and the amount of investment it takes. And how cross-functional data culture will positively impact your upstream and downstream data quality and infrastructure. Because the more people are interested in leveraging data, the more people care about data quality in the entire data lifecycle—anywhere from reliable event instrumentation to well maintained documentation about which revenue dashboards you should trust.

What’s the first thing you should do when starting a data team? What’s the first thing you should do as the first data leader of an organization?

Listen in for Che’s advice on starting and scaling a data team, including a reminder that the point of a data team is not to build a data warehouse; the point of a data team is to improve decision quality at your organization.

Getting to Know Your Users with Boris Jabes, CEO & Co-Founder of Census

Boris Jabes is CEO and Co-Founder of Census, a platform to operationalize data through a process often referred to as reverse ETL. Boris previously co-founded Meldium, a ground breaking account and password management solution for teams, acquired by LogMeIn in 2014. Before that Boris was a senior Product Manager at Microsoft where he worked on Visual Studio – a widely adopted IDE.

Fun fact: This is a crossover episode! Stef joined Boris in Aug ‘21 on his podcast The Sequel Show for a great chat.

We discussed how SaaS has shifted mindsets for product releases. From when you shipped software products and couldn’t really change it for years, to now when software products get shipped early and often, released experimentally and iterated on rapidly. This shift has changed the feedback loop and made product analytics a fundamental part of business strategy. 

We also talked about the amalgamation of product data and business data, the challenges of stale data and whether we should build expiration dates into dashboards, how there are hidden data pipelines all over in every company, and how the the hub and spoke model can be applied to ownership of different data sets in an organization.

Listen in for Boris’s insightful thoughts on data ownership and making every team in your organization empowered with data.

The Argument for Less Specialization with Erik Bernhardsson of Spotify and now Modal Labs

Erik was early at Spotify and built out their data team. He originally joined to work on music recommendations, but ended up prioritizing more important things. Eventually he did get to build the core recommendation system—making all our lives better. Along the way they built Luigi, one of the first major open source workflow schedulers, which we also used at QuizUp. He later joined Better.com as CTO and employee #8 and built a tech team of eventually 300 people—including the data team. Now he’s working on something new at modal.com.

Spotify’s org structure was an inspiration for how we structured our teams at QuizUp, including how we thought about the data team. So it’s a great pleasure to have Erik on The Right Track to talk about his experience building data cultures.

We covered a lot of ground. Including the mind-blowing experience when you first try cohort analysis and find something unexpected, how advanced machine learning should typically wait until you’ve fixed the data fundamentals—the data plumbing, and how and when you should prioritize that data plumbing you can’t ignore—including consistent event logging.

We also discussed data team recruiting and org structures, how hiring the right people to your early stage data team is the most important factor for success—which for Erik means something like a technical data journalist, meaning people who—sure, they’re good with numbers, they’re ok with stats, ok in software engineering—but above all, they’re driven by a pursuit of the truth and making an impact on the business.

Listen in for Erik’s thoughts on strategy when building and scaling a data team and data culture, and sit back and enjoy letting him convince you that simple things are usually far more impactful than complicated things.

Defining the Data Scientist with Josh Wills of Weavegrid (formerly at Slack and Google)

Josh has been around the block in his data career. He was the first Director of Data Engineering at Slack and built Slack’s logging infrastructure. He worked on Google’s foundation for experimentation around ads. He’s a self-proclaimed ex-statistician now working as a software engineer, but stays close to the data space with hot takes on Twitter and by angel investing in the tools he would have wanted back in the day. He’s now a software engineer at Weavegrid.

Josh coined a commonly quoted definition of “data scientist” as a “Person who is better at statistics than any software engineer and better at software engineering than any statistician.”

… a definition I heavily relate to. 

We discussed the role of a data scientist and whether that’s a helpful title for anyone. We covered his learnings at Slack, and how much of a difference it makes for a company’s data culture when data producers and data consumers are on the same team, working closely together – or furthermore when they are in fact the same person (which is rare). We also talked about what you can do and where to start, to get buy-in for good data culture, whether it’s at the genesis or when you’re knee deep in total chaos and want to make it better.

Listen in for Josh’s standup comedy and actionable insights from his life and trauma as a data person.

Data Mindset with Benn Stancil of Mode

Benn Stancil is Chief Analyst and Co-Founder of Mode, a product for data analysts and data scientists—and an influential voice with original thoughts in the data community.

He and his cofounders were on Yammer’s data team, where they built internal tools to create analyses and quickly share with their coworkers. When Yammer was acquired by Microsoft in 2012, the Microsoft teams started adopting their tools as well. From there they discovered that most other leading Silicon Valley companies had built the same things. So they started Mode. Benn wears many hats as Mode’s Chief Analyst, from the internal data teams at Mode to the wider community, knowledge sharing on anything-data and supporting analysts in building data strategies.

As two founders of data products, Stef and Benn went deep and philosophical. Listen in for discussions on the shift from vertical to horizontal data tools, why you won’t get self-serve analytics by buying a tool – just like you won’t get self-serve design by buying Figma, and advice for kickstarting a data team.

Domain Expertise with Laurie Voss of Netlify

Laurie Voss is a Senior Data Analyst (and data evangelist) at Netlify but a web developer at heart – ever since he started his first web development company 25 years ago; when he was so young that when he had business meetings, his mom would drive him to meetings because he didn’t have a driver’s license yet. Before Netlify, Laurie was co-founder, CDO, COO and CTO of npm – the most widely adopted package manager for Javascript development, acquired by GitHub in 2020.

In this show Laurie talks about the roles in the modern data team and whether PhDs are good first data team members, whether self-serve analytics will ever work, and how important domain expertise is for extracting insights from data.

Intangible Metrics with Elena Dyachkova of Peloton

Elena is a Senior Manager of Product Analytics at Peloton and has combined her passion and professional background in track and field with her background in sports research and analytics to build Peloton’s product analytics from the ground up. In the show Elena talks about how her team uses metrics and data to discover and build new user experiences, and how their most popular user features were powered by intuition with data.

Data Stewardship with Claire Armstrong of Fender

Claire is a Director of Product Management at Fender and product owner on Fender’s Data Squad. We talked about her hands-on strategy and tactics to better equip the Fender team to use reliable data—including her data stewardship workshops, where she educates their analytics stack, how their data works behind the scenes, what their analytics mean, and how to deliberately design good data, based on product goals.

Data Custodianship with Nick Threapleton of Culture Amp

Nick is a Senior Product Analyst at Culture Amp. He made the journey from industrial design through marketing into data. We discussed how to empower folks on your team to use data in their daily work. A little empathy goes a long way.

Data Literacy with Maura Church of Patreon

In this episode, Maura shares how to build strong data culture across disciplines and what it means when someone says "I don't trust the data." Data literacy is as much about people as it is about numbers.