2021 Puppet State of DevOps Analysis: Where Is the Middle and How to Escape It

Ten years of DevOps

In the technology space, DevOps has been the topic-du-jour since 2011. For the last ten years, companies like Chef, Puppet and RedHat have taken the time to monitor, on an annual basis, the technology industry's most fashionable trends, most hyped tools and most effective practices. Each year DevOps has topped that list to such an extent that Puppet authors an annual address on the matter. 

The Puppet State of DevOps report has grown in notoriety to become one of the industry's most rigorous, comprehensive and influential publications when it comes to shaping current and future trends in software/infrastructure development culture and best practices. Having collected such a deep pool of data from which to base its conclusions, Puppet's 2021 State of DevOps Report sheds a light on several trends which are as surprising as they are illuminating, that is to say: 

Though that statement may sound harsh, Puppet's report is not without suggestions on how to diagnose, mend and/or enhance your company's DevOps operating model and the underlying maturity thereof. Overall, the data is rigorously collected, analyzed and presented, and just as it has in the past, the 2021 Puppet State of DevOps report promises high-quality insight that any company can use to adopt or further mature their DevOps practices.

Breaking down the middle

The theme of this year's State of DevOps report is "Breaking down the middle," and for good reason. According to Puppet's 10 years of data collection, while 79 percent of surveyed organizations said that they were in the "mid-evolution" phase of DevOps adoption back in 2011, in 2021, 78 percent of organizations claim the same. Adding to the worrying knowledge that over the course of 10 years there has been a reduction of just 1 percent of companies claiming to be in the mid-evolution stage of DevOps adoption, there are also some implications to the fact that in 2021 only 18 percent of these companies claim that they have adopted a highly evolved DevOps practice (up a measly 8% from 10 years ago). 

What this data implies is that while it is not terribly difficult for low-evolution DevOps organizations to clear the runway and begin their ascent into becoming highly evolved DevOps organizations, there are preverbal "gale force winds" that make achieving optimal altitude elusive for most organizations. Puppet's data indicates that the challenges most often cited as preventing organizations from breaking through the mid-evolution stage of DevOps adoption are as follows:

• 33% - Shortage of skills.
• 29% - Legacy architecture.
• 21% - Organizational resistance to change.
• 19% - Limited or lack of automation.

These blockers are fundamentally different from those of highly evolved DevOps organizations because highly evolved DevOps organizations need only worry about the shortage of skills and legacy architecture issues, both being industry-wide challenges in their own right. What this data reveals is that the most critical blockers for organizations that wish to become highly evolved DevOps practitioners lie less in the technological hurdles of DevOps adoption, and instead lie in the vastly more challenging task of augmenting organizational thought and culture. 

DevOps as a culture

Historically many IT leaders have written off the challenges to becoming a highly evolved DevOps organization as insurmountable because of the lack of understanding of the following concept: 

This tendency has resulted in the nearly universal focus of most organizations to channel their efforts into becoming highly evolved DevOps practitioners almost exclusively on the technological aspects of DevOps. While this tendency to focus on the technology can get an organization to the lower-middle stage of DevOps evolution relatively quickly, the mentality that it breeds is almost always antithetical to breaking past the middle and becoming a highly evolved DevOps organization. Instead, Puppet posits that thinking about such challenges as simply "culture problems" is neither useful nor productive. 

The data suggests that most organizations who consider themselves to be in the mid-evolution stage of DevOps maturity are likely suffering from some combination of resistance to change, insufficient feedback loops and/or a culture that discourages risk as barriers to more advanced DevOps practices. 

Signs that these problems exist in an organization can be as simple as hearing the "it is what it is" mentality when questioning the rationale of, or trying to improve upon, existing processes and procedures. Examples of where these problems may manifest are in the existence of Change Advisory Boards (CABs), Architecture Review Boards (ARBs) or security reviews that act as gates to product delivery or deployments.

Adopting practices that work

The report posits a couple of key actions that technology leaders can deploy to combat these "cultural problems," beginning with identifying flow patterns within one's organization and then optimizing them. The data suggests that when organizational leaders give individual product teams more autonomy by investing their DevOps efforts in the development of valuable and usable processes that promote self-service capabilities, the often artificial barriers that historically act as blockers to product delivery/deployment can either be augmented to promote faster flow patterns, or even systematically erased all together. Putting in place initiatives that promote knowledge sharing helps create explicit mandates and expectations between teams, which can help provide clarity of purpose to teams internally, as well as enable teams to begin sharing best practices and tooling.

The data clearly shows that highly evolved DevOps organizations almost wholesale eliminate production deployment gates by: 

• 85% - Sharing and documenting best practices.
• 85% - Explicitly understanding their IT infrastructure landscape.
• 82% - Utilize change management procedures that inform appropriate stakeholders.

Organizations that employ platform teams that leverage the "Platform as a Product" mindset to provide software delivery solutions such as infrastructure, data and CI/CD tooling for multiple feature teams through self-service APIs are nearly 8x more likely to be in the highly evolved DevOps evolution stage or 5x more likely to be in the mid-evolution stage than organizations that do not (3 percent). These data correspond extremely well when one considers the recent uptick of platform accelerator programs that consulting organizations (such as WWT) have begun conducting in all sectors that leverage technology, and the success that they've seen with them.

The takeaway

Overall, the most visceral insight that Puppet's 2021 State of DevOps report offers is not that many organizations have no interest in becoming highly evolved DevOps practitioners, it is that the adoption of tools and technologies are often viewed as the best way by which to reach that most elusive of goals. The data suggests, in a multitude of ways, that the barrier of entry to becoming an organization that leverages DevOps is quite low; however, breaking past the middle is almost always a matter of augmenting and improving an organization's culture rather than adopting the newest automation technology or leveraging the cloud. 

The breaking down of communication and process barriers from the leadership and the individual contributor levels are inherently what makes highly evolved DevOps organizations so effective, as well as the liberal usage of platform and stream-aligned teams in the development of products and services. While all organizations who are actively trying to "do the DevOps" are on a journey of growth and greater efficiency, Puppet's 2021 State of DevOps report uncovers that it is difficult for most companies to truly embrace these ideas without fundamentally changing the DNA of how they work and what they value. 

If your organization is focused more on the technical aspects of DevOps than it is around the people and process aspects, it is statistically more likely to be somewhere in the mid-evolution stage of DevOps adoption. Given the challenges of changing an organization's culture, it can seem that growing beyond the mid-evolution stage can be nearly insurmountable, but Puppet offers a variety of very clear, highly effective, data-driven solutions that just might enable your teams to optimize the flow and feedback patterns that ultimately help them grow into highly evolved DevOps practitioners, reaping all the benefits that come therein.