SAN FRANCISCO — The chief operating officer of Pivotal Labs, Edward Hieatt, has some advice for development teams:
By that, he means that you need to get good at estimating how much time it will take to complete development projects, and complete them in a predictably consistent timeframe. And you need to do them faster and faster.
“The development cycle needs to get tighter and tighter and tighter, especially at larger companies,” Hieatt said.
Related: Pivotal Labs is part of Pivotal, which announced a major new set of enterprise PaaS and big-data products today.
The background: Thanks to the advent of cloud-based software-as-a-service (SaaS), software can be updated and deployed to customers far more often than the old six-, twelve-, or eighteen-month refresh cycles. You can push a new product to customers every month, or every week — and many companies are.
What’s more, product owners — the managers who speak with customers and give direction to dev teams — have an annoying tendency to revise product descriptions and even overhaul their complete product vision.
In light of that, Hieatt said, speed is important — but it’s not the most important qualification. Product managers “would like the rate of output to be as constant as possible. They don’t want there to be any surprises,” he said.
“And they want all of this in the face of their right to change the product direction at any time.”
Developers may grumble about this, but the key to a productive relationship between business and software development is setting expectations and making it clear what is — and is not — possible.
In order to help dev teams meet that goal, he suggested creating a new metric to track software teams’ “volatility,” or the rate of change in how quickly they deliver on projects. He defines volatility in a precise, mathematical way, but in short, it’s the standard deviation of changes in speed, divided by the average speed, expressed as a percentage.
Pivotal’s Pivotal Tracker, a project-management tool, actually has this metric built in, so Hieatt was able to use it to analyze the consistency of various dev teams within Pivotal. What he found was that the median volatility was around 90-100 percent, meaning that the typical dev team varies its delivery time for projects by as much as 100 percent from week to week.
A volatility of 20-30 percent, by contrast, is incredibly consistent — and that’s just what business managers want to see.
“We’re in the business of serving the business side of the house. At the end of the day, it’s time to buckle down and try to produce against the backlog, whatever it may be that week,” Hieatt said.