Call it the DevOps disconnect. Excitement about DevOps keeps rising, with one report even suggesting that the number of companies that have adopted DevOps or plan to do so has surpassed 90%. And the DevOps tools market is expected to grow to over $10 billion by 2023.
At the same time, Gartner predicts that through 2022, 75% of DevOps initiatives will fail to meet expectations due to issues around organizational learning and change.
By unifying development and operations end to end, DevOps promises to increase companies’ speed and agility in software delivery as they grow their digital initiatives, without sacrificing quality.
Yet while many organizations are finding success with DevOps pilot projects, they’ve struggled to scale those results further in the enterprise.
So, everything is awesome with DevOps, except when it’s not. What’s getting in the way? Here are the five most common issues that imperil DevOps initiatives and how developer teams can avoid them.
1. Failure to partner with the business.
One of the biggest barriers to scaling DevOps efforts, Gartner says, “is the mindset of infrastructure which, after years of standardization to drive efficiency, sees an intractable conflict between its mandates for reliability versus speed.”
What that describes is an organizational tension inherent in DevOps: The business is always pushing to release software more quickly, while DevOps necessitates taking a breath and going slower initially to put in the processes and tools to move faster over time. It’s the classic “urgent and important” vs “not urgent and important” quandary described by Steven Covey in “The 7 Habits of Highly Effective People”.
Abraham Lincoln said, “Give me six hours to chop down a tree and I will spend the first four sharpening the ax.” IT and the business need to take this wisdom to heart and align so they’re investing in short-term strategies and resources that will pay off in the long term.
2. Underestimating the culture shift.
This is huge. Because change is hard and feels risky, an organization’s ability to overcome cultural barriers to embracing change is essential to DevOps success. This is especially true in those that for so long have built and run the software “factory” a certain way. People are used to and trust that process.
So, work on culture change. Talk about it, rally around it. Bring new people, both leaders and individual contributors into the team who can help.
3. Failure to measure.
Doing DevOps for DevOps’ sake is silly. Organizations need to set specific parameters of what success looks like. Is software being deployed faster or more frequently? Has quality increased?
This is an area where companies often fall short. When success is ill-defined, it’s hard to tell if progress is being made. And it’s easy to abandon projects when things get hard.
Pro tip: Pick the right teams and right applications to score quick wins that can demonstrate to others what success looks like.
4. Not managing DevOps like a product.
DevOps initiatives should have a detailed pipeline just as a product would. There should be a roadmap of projects and technology rollouts, shared with the business, rather than a big bang approach. One of our largest customers has a published roadmap of releases for their DevOps capabilities. Their internal customers know what to expect in future releases, how to leverage them and build them into their own resource plans.
Have a DevOps roadmap that defines “releases” of DevOps capability. For example: “This quarter, we will change the team construct to include the right players. We’ll pull database administrators into the process. Next quarter, we’ll automate builds/deployment of applications and have the DBA team automate database deployments. The quarter after that, we’ll bring the two pipelines together.”
5. Overlooking hidden icebergs.
People tend to underestimate dependencies for success in DevOps projects. Sometimes, the reason is culture. Other times, it’s failing to recognize that a critical part of the application delivery isn't part of the process.
Too often, organizations fail to automate these processes, wonder why they can't get the ROI or success they expected in DevOps, and then realize you must address the most important asset your company has… your data.
Enterprise organizations are looking to DevOps to get the software speed, stability, and availability needed to increase profitability and customer satisfaction. But they must be aware of the pitfalls or risk ending up flailing in the DevOps disconnect. Pilot project wins are nice, but true DevOps success means being able to extend those results to other critical application pipelines throughout the enterprise.
Derek Hutson is CEO of Datical, an Austin, TX-based database release automation software company.