Earlier this year, Puppet Labs released the 2015 State of DevOps Report, a follow-up to last year’s ground-breaking study on DevOps and its impact on organizations. I’ve had the pleasure of hearing Dr. Nicole Forsgren, the lead research investigator and co-author on the 2014 report, present the results of this year’s report at the 2015 DevOps Enterprise Summit and a recent DevOpsDC meetup. Those results are compelling.
To grab your attention, this year’s report found that high-performing IT organizations outperformed their peers on both speed and stability metrics by wide margins. How wide? This wide…
- 30x more frequent deploys.
- 200x shorter lead times.
- 60x fewer failures.
- 168x faster mean time to recover (MTTR).
In addition to the eye-popping performance metrics, I had five more key insights from the report.
1. You can have both speed and stability.
Conventional thinking states that to get more speed, you have to give up some stability and vice versa. The report shows this thinking is incorrect. High performers are getting both increased speed and increased stability. There is no tradeoff — speed and stability move together. For example, high performers are deploying 30 times more frequently than their non-high performing peers and have 168 times faster MTTR. More speed and more stability.
2. The constraint on speed has moved from an IT limitation to a business decision.
The 2015 report showed fairly similar speed-related metrics (e.g., deploy frequency, lead times) as the previous year’s study, but showed large increases in stability-related metrics (e.g., failure rate, MTTR). This result was unexpected — one would expect both metrics to go up together (see the first insight). The explanation for why this didn’t happen is that once you can deploy whenever you want (e.g., tens, hundreds, or thousands of times a day), you don’t get any more business value from deploying faster — you’re already moving as fast as your organization can make decisions. However, you do get business value from improving stability so that’s where the investment moved. For high performers, the primary limiter on speed is now no longer the IT capability of whether you can deploy changes, but a business decision on whether you want to deploy changes. How powerful would that be to your business?
3. There is a link between DevOps practices, IT performance, and organizational performance.
In IT, we’ve always craved clear and compelling evidence that what we do matters to our organization. Thanks to the research in the report, that evidence now exists. How well an organization follows a handful of continuous delivery and lean practices is a strong predictor of that organization’s IT performance and in turn its overall performance relative to productivity, profitability, and market share goals. In fact, last year’s report showed “the publicly traded companies that had high-performing IT teams had 50 percent higher market capitalization growth over three years than those with low-performing IT organizations.” What CEO or Board of Directors wouldn’t want that kind of performance?
4. DevOps scales developer productivity.
Time to shatter some more conventional thinking. If you want to scale overall software development productivity (i.e., get more done organizationally), you need to add more software developers, right? And you should just accept that individual software developer productivity will go down (i.e., an individual developer will get less done) because of diminishing returns, right? Nope — at least not for high-performing IT organizations. For those organizations using continuous delivery and lean practices, adding software developers to the team actually increases individual productivity — not decreases it, as measured by the number of deploys per day per developer. DevOps is a force multiplier for software development teams.
5. Employee burnout is avoidable.
IT is often a high stakes and high pressure environment. I’ve experienced that intensity many times over my career and I’m sure many of you have, too. My eyes were opened to the impact of that intensity after reading John Willis’ raw blog post on burnout and suicide. In a real and meaningful way, DevOps can make life better for people. So what is good for our business is also good for the humans in our business. Just as speed and stability move together, so do the health of our people and the performance of our organizations. If you care about the people you work with, please read the section in the report on burnout — and then act.
Thanks so much to the people who put their time, money, and care into creating the 2015 State of DevOps Report. It’s an incredibly powerful tool to support change for the better. I’m looking forward to what insights we glean from next year’s report.