Last week was Fixit Week again at Color — one of our favorite times of the year. More than 30 designers, engineers, clinical lab specialists, and product managers hit pause on our normal week-to-week projects so we could initiate and pursue self-directed projects, with a goal of making Color better for everyone.
Some projects focused on improving our externally-facing client experience. Before Fixit Week began, two of our designers created a beautiful Book of Opportunity™ with the most commonly reported client confusion or pain points. Other projects worked to improve internal operations, from improving how we manage Family Testing Program applications to a redesigned rack-and-storage system for our lab specimen intake room. Finally, some engineers and designers chose to tackle internal team improvements — a containerized development environment, centralized web components and style guides, and a massive codebase upgrade to python3. A variety of prizes were awarded to the winners in each category.
Why is this necessary? Every fast-moving team naturally accrues various forms of “debt”: tech debt, product debt, design debt. We bias towards delivering new products and genetics to our clients quickly, so taking a little time periodically to work through one’s personal backlog can be immensely satisfying, not to mention beneficial for the team in the long run. So during Fixit Week, a few PM’s and designers bravely curated and organized more than 1,000 legacy JIRA tickets hiding in dormant projects, while I personally enjoyed finding and deleting 5,000 lines of legacy unused code, some of which dates back to the founding days of Color in 2014 when our CEO, Othman, was still coding a little!
What’s perhaps most unique about Color Fixit Week is the system we used to discover ideas and encourage focus on the most impactful projects. On the surface, it sounds simple: we solicited ideas from the whole company and asked everybody to vote on each other’s ideas. But last August, when we first tried this system, it was new and unfamiliar, so few ideas received many votes. This time, people self-organized to form PAC-like groups and ensured that their favorite projects received sufficient support; in the course of a few days we got more than 100 submissions and recorded nearly 600 votes! Here are a few of the most popular projects:
- Better synonymization/grouping of near-identical structural variants
- 80% faster page load time for our variant confirmation manifest tool
- Better mobile user experience: using the digits keyboard consistently for numeric fields
- Automated generation of chromatograms from external lab Sanger results, seamlessly incorporated into our clinical tools on AWS
- More thoughtful timing of test-related email notifications
- Upgrading our sizeable codebase from python 2.7 to 3.6, a multi-week undertaking that we pushed over the finish line during Fixit Week (and the subject of a future blog post)
Once the voting began, things really kicked into gear. Projects which received the most votes became hot commodities, with engineers hustling to claim and fix the underlying issue before anybody else could get to it. Whoever successfully completed a project earned a “bounty” equal to the number of votes that project received, with a Bounty Prize awarded to the individual who collected the most “bounty” throughout the course of the week. This resulted in a nice alignment of incentives and some friendly competition: fixers competed with each other to tackle the most popular ideas, and idea submitters rallied their colleagues for votes. The winner of the Bounty Prize tackled 7 projects worth a total of 85 votes/points.
Together we tackled more than 50 projects and issues over the course of the week, so we celebrated with demos and prizes on Friday — an entertaining and exciting way to wrap up our quarter.
Was Fixit Week successful? Absolutely. Could we improve it even more? Undoubtedly. Perhaps one of our next Fixit Week projects can be finding improvements to our last Fixit Week.
Want to join us for our next Fixit Week? Check out the open roles at Color.