A few days before launch, the client asked us to change how some of the data was coming in.
I knew immediately it was the wrong call. The request wasn't unreasonable on its face - they had reasons - but changing a data contract days before a high-stakes launch wasn't a technical risk, it was a usability risk. The product would ship broken in ways that mattered, and we wouldn't have time to find out how.
I went to my producer, Damian. He wasn't happy that we were going to push back on the client. He made the case anyway, and he had my back the whole way through. We agreed to make the API change post-launch as a fast-follow. The product shipped the way it was supposed to.
That moment has stayed with me. The instinct to say yes to a client, especially close to launch, is almost overwhelming. But protecting the work - knowing when to hold the line - is part of the job.
The product
The Digital Trading Room was the Dodgers' version of Moneyball. The experience had two parts: a tablet running a responsive web application, and a set of very large screens in the team's back office. The tablet was the control surface. Coaches and analysts could pull up a player, throw his stats on the big screen, and evaluate whether he was being undervalued - whether he was worth picking up.
The goal was to give a baseball team a real-time analytical edge. The execution required something that didn't obviously exist yet: a data layer that could surface Bloomberg Sports' player data through APIs a web application could actually consume.
That was my first workstream.
The build
I was embedded at Bloomberg Sports for the early phase of the project. Their data was the engine; my job was to build the APIs our web applications would use to pull it. I wrote them in Ruby on Rails.
I should mention: I didn't know Ruby on Rails before this project. I learned it specifically for this engagement - picked it up, got productive, and shipped production APIs in it. This is a pattern in my career I've stopped apologizing for.
While I was there I also wrote unit tests in RSpec - not just for the APIs I was building, but for significant portions of a system that had no test coverage at all. That work wasn't glamorous and it wasn't in my original scope. It needed doing.
The second workstream was the web application itself: completing features, fixing bugs, getting the tablet experience to the state it needed to be for launch.
Three CLIOs
The project won three CLIOs: two bronze, for Digital and Digital Technique, and a silver from CLIO Sports - all for the Digital Trading Room itself.
Awards are nice. What I think about more is the launch-day call, and Damian backing me up when I made it. The product shipped right. That's the thing that matters.