Helen Sun wanted to move fast.
With 20 years of tech experience behind her, the CTO at Stats Perform knew that if you want to stand out from the competition, your team had to be brilliant — and be brilliant often.
Shortly after her arrival as CTO in mid-2018, the data and AI-powered sports statistics company decided that to hit new levels of achievement — without hitting new levels of burnout — they would need to adopt a new methodology.
They landed on extreme programming (XP), which falls under the popular Agile umbrella. While some of the tenants are pretty common sense — feedback, communication and respect, for example — the structure also touts some ambitious practices, like one-week cycles and an emphasis on pair programming and storytelling.
Compared to other engineering methodologies, like scrum or waterfall, extreme programming was created specifically to encourage tight turnarounds.
For example, following a test-first order of operations — another feature of XP — engineers create tests, then build out code which they know will pass the test. It’s a preventative approach: the team codes correctly the first time, rather than circling back to fix errors.
To accelerate a team’s technical operations, XP certainly stands out as a natural choice. For Sun, making the switch to the methodology seemed like a no-brainer.
“Engineers start the day together, take breaks together and leave work together — it’s an exhilarating environment,” Sun said.
Meet the team:
Sun is the chief technology officer and oversees AI, innovation, global engineering, product design, and information technology services.
Who she wants on her team: “Technology skills have a shelf life of a banana peel. I look for individuals who have a growth mindset, are thoughtful, have a systemic way of thinking, and are continuously learning.”
Peterson manages a team of software and data engineers who are building out sports betting predictions as well as the live-insights products.
Why he chose Stats Perform: “I was attracted to joining an extreme programming environment and saw an opportunity to have a large impact on the company.”
Nichele leads an engineering team, but she also leads a group of product owners to ensure that the product roadmap is being followed.
Why Stats Perform is a good fit for her: “Product owners own the backlog, prioritization of tasks and the extreme programming process in order for the team to drive toward milestones and product goals. Having a real sense of ownership and a direct impact on the product is one of the things that I appreciate most about my role.”
What emphasis does XP place on collaboration? Why is that helpful for the engineering team at this point in time?
Helen Sun: We are looking to grow and scale in order to deliver more solutions faster. In order to do that, we need to retool and reskill our current talent, as well as attract, retain and train new AI talent. Extreme programming enables us to onboard new engineers faster through paring and knowledge transfers. Engineers pair with AI scientists through XP on many projects to learn the nitty-gritty of AI, and vice-versa. Engineers start the day together, take breaks together and leave work together — it’s an exhilarating environment.
Collin Peterson: My teams are very collaborative, both amongst themselves and across teams. I encourage them to talk to each other to solve problems by jumping into a room and diagramming on a whiteboard to come to a solution that works for everyone. It’s critically important to foster an inclusive environment where each person is heard and the team is committed to finding the best possible option.
Another pillar of XP, “whole team,” emphasizes teamwork. We’ve all seen examples of teams that seem to be working on different wavelengths, so how does Stats Perform pull together and regroup?
Peterson: A couple of our teams recently realized that they were doing similar things in different ways. After some healthy debate about whose solution was the right one, we decided to shut ourselves in a room for the afternoon to talk through all of our different use cases. We talked through not only the immediate needs, but also about other known use cases and likely future needs. We ended up landing on a third option that we feel will benefit not only both teams but our entire company.
XP dictates that there is a one-week iteration period. That seems like an incredibly tight turnaround for incorporating feedback and fixes. How do you manage it?
Nicole Nichele: As a product owner, I concentrate on writing stories for the engineering team that are small and manageable. Each story should be adding value to the product and for the user.
To turn things around quickly, we take a minimum viable product (MVP) approach to both the design and development of products. This allows us to build out a product with just enough features to satisfy the first iteration of the product and to provide value to early customers. From this approach, we are able to then gather further feedback for future product development and further iteration.
Peterson: Each iteration starts with a planning meeting and ends with a retrospective. Each day starts with a stand-up and ends with a stand-down. Other than that, there are almost no additional meetings for our engineers. This means our engineers get to code for the majority of each day. We also follow trunk-based development with multiple production releases every day, which means our engineers are seeing their work come to life almost immediately.
Sun: As Collin mentioned, there’s no red tape and no disruption to engineers. Our engineers have the opportunity to work on cutting-edge technologies, cloud-native microservice-based architecture, and continuous integration and continuous delivery, which enables us to release multiple features to production in a day.
It sounds like the quick iteration cycles have been working well for your team. How many products have you been able to launch as a result of XP?
Sun: This year, we have launched four new products, delivered a new, major version of a computer vision tracking system that company pioneered 10 years ago, and released a new module of another AI product. The pace of innovation and the value of the work provided to our customers has made Stats Perform an incredibly exciting place to work.
Can you tell us a little bit about what you’re working on now? What’s next for Stats Perform in 2020?
Nichele: There are several different projects in progress. Notably, one project focuses on creating prediction data feeds using our rich historical data and the use of AI-driven models. The other project focuses on building visualizations that demonstrate the use of our predictions. Seeing both develop from start to finish is an exciting process that allows me, as a non-technical individual, to stretch and challenge myself, while also providing an interactive experience to the product.
