Today we are talking to Bill Harding, the CEO at GitClear. And we discuss new ways of measuring developer output, what it takes to be an effective communicator, and how to hire and retain the best talent.
All of this, right here, right now, on the Modern CTO Podcast!
Bill has been obsessed by computer programming since 1992, when his parents splurged for a Packard Bell 486SX/33 with 4mb RAM and 310mb of hard drive space. He grew up in a flea-infested trailer with a leaky roof, on a 10 acre farm in a rural Washington town (population 4,300). These details are relevant to explain how, by the time Bill was 13 years old and enrolled in Computer Programming courses at his local community college, he had formed what would become a lifelong belief: that no other hobby would ever be half as interesting as programming computers. The next best entertainment options at that time and place were so incredibly boring by comparison, this truth was inescapable.
For the last 10 years, Bill has helped shepherd Bonanza.com from a programming side project to a company that is consistently recommended over Amazon and eBay by sellers participating in the largest annual poll of ecommerce sellers. Bonanza’s efforts have been buoyed by an amazing team that share his conviction for “listening to the customer” and “proving we care through our actions.” In the last five years, Bonanza has been named one of the 100 Best Companies to Work For. In 2017, Bill was an Ernst & Young finalist for Entrepreneur of the Year.
Aside from programming, Bill’s other lifelong passions are measurement and personal productivity. The thrill of measurement stems from it being integral to continuous learning & improvement. Trying to improve without reliable measurement is like trying to drive across the country without a map. To whatever extent it may be possible, it’s ridiculously inefficient. The only thing worse than trying to improve without measurement is trying to improve with unreliable measurement. This explains Bill’s occasional forays into masquerading as a writer, which include his popular essay scrutinizing the four metrics that agitate developers.
What we are:
GitClear is CliffsNotes™ for GitHub. We digest all your repository’s commits into a quantified data stream that lets managers and engineers get the gist of their code faster. For developers, we reduce tedious review work and leave more time for coding. For managers, we provide a window through which to observe the state of their developer team. For both, we provide a dashboard of code metrics that helps make decisions supported by data.
Why we are here:
GitClear sprung from the pain of scaling the development team for Bonanza.com. Once we hit 10 developers, it became immensely time-consuming for us to answer even simple questions about our development team:
Who is working on what? Is anyone stuck? Does our weekly “work from home” day impact productivity? By how much? How does Bob’s code output from his first three months compare to one of our top developers, like Alice? What should we talk about during Alice’s annual review, i.e., what was Alice working on 9 months ago and how did that go? For code review, GitHub could suffice, even if the time required to read a day’s worth of commits was clearly longer than necessary (“previous commit” button, anyone?).
For productivity-based questions, our options were far worse. All of the prior art seemed based on commit counting or code line counting — both effectively useless for gauging the impact made by a developer. Online discussions of development quantification were inexorably drawn toward hopelessness. But we were hopeful. If a Lead Developer can express the characteristics of their most valuable developers, then why couldn’t we teach a learning algorithm to do the same?
All of this brings us to present: a tool that can map how code changes over time, intelligently group commits of similar purpose, and present a team like ours with a productivity metric built to analyze code like a Lead Developer would.
Read more about: How our code analysis works