Can We and Should We Measure Software Engineering Productivity?
You cannot improve what you cannot measure.
The constant challenge for engineering leaders is to improve their developer productivity to deliver business outcomes faster. The first step towards improving developer productivity is to identify the right metrics to measure, the ability to visualize them quickly to identify bottlenecks in the process, and unblock the engineers, explains Sudheer Bandaru, Founder & CEO, Insightly Analytics.
You cannot improve what you cannot measure. But, how do we measure engineering productivity where every outcome isn’t the same and cannot be measured by mere lines of code? I struggled with this for years, and here’s a brain dump of my evolving thoughts on this topic.
With the rising demand for tech talent, increasing salaries, and companies going remote – this has become a hot button topic at pretty much all the companies – and engineering leaders are grappling with this challenge of managing software engineering teams effectively. Simply increasing headcount without measuring or improving productivity is no guarantee of improved quality or number of releases.
The pandemic has caused a global employment crisis, which has hit the American tech market rather severely. The Bureau of Labor Statistics estimates that by 2026 there will be over 1.6 million vacant CS jobs. Given this shortage of engineers, even the most well-funded companies cannot attract top talent. Say they can hire top talent; it’s becoming more critical for them to justify the cost and effectively manage them to deliver quality outcomes for the business. If you are an engineering leader or a CEO with software teams, I trust you can resonate with what I am saying. If I am not wrong, you may have just approved the 50% increment your junior engineer who joined last month asked for as he got another offer at a different company with a 100% hike.
The Scientific Art of Measuring Engineering Productivity
Jokes apart, can we measure engineering productivity? Isn’t it art? Isn’t it a creative field where you cannot objectively measure their value? Yes, it’s an innovative field and is an art. But, hey, aren’t we measuring the sales function? You ask salespeople, and they say that selling is an art. But that’s not stopping the companies from objectively measuring sales teams and their outcomes. Luckily, it’s not that bad with software engineering.
It’s of no use when measuring the output – like lines of code, number of hours, or tracking the websites they visited. It also doesn’t differentiate between time spent on writing new code or fixing bad code. At its worst, measuring output rewards busy or inefficient work and penalizes smart work. You’d be surprised at how much busywork a developer does. A developer spends an average of about 13.5 hours a week on just technical debt. That’s over 700 hours a year spent on fixing past mistakes. Now that’s not something you can easily ignore!
See More: 4 Practical Ways to Tackle SaaS Sprawl Effectively
But what if you measure the ‘outcome’ and not the ‘output’? When you calculate the outcomes like the number of releases, time to market, the number of bugs, it helps the entire engineering org align towards the same goal, same metrics and augment gut decisions with data. It turns managers into leaders, and it gives autonomy to the engineers. As long as the engineers are connected to the purpose and mission of the company, this will empower them while still allowing the leaders to measure productivity objectively.
Rethinking Rework
Engineering can be measured, and it could then be a tool for success – for engineers, engineering leaders, and companies when done in the right way. Research by Stripe suggests that, on average, engineers spend four hours a week working on “bad code.” Large organizations lose around $93 million a year on unnecessary rework. Over a year, this amounts to $85 billion lost in opportunity cost. Only when this is measured can it be saved.
To summarize, in this world of data-driven culture, objectively measuring the right metrics brings the right culture to the teams, and it’s proven that elite teams aren’t formed magically but only because of their scientific approach to improvement by measuring, optimizing, and improving. Continuous improvement fostered by data analytics produces the best outcomes for companies. We have seen this with all functions and areas of life – no doubt analytics has gained popularity – because data can’t lie.
Are there other strategies to quantify engineering outcome and productivity? Tell us on LinkedIn, Twitter, or Facebook. We’d love to know what you think!
MORE ON PRODUCTIVITY
