Biggest Challenges in DevOps
Thoughts ahead of DevOpsEnterprise #DOES15
Technology has enabled people to consume products in new ways. Mobile devices, online social, and cloud computing allow customers to easily find what they want and consume products on the go. These consumption patterns lead to entirely new business models and are disrupting existing companies.
Often these challenges are addressed with a new technology architectures to build winning platforms, or by hiring folks "from the valley". These approaches are too tops-down and take a considerable amount of time and money to validate. It is a big bet to make on any one platform or any one person.
New CIO, CTOs, and new tech platforms typically run an 18 to 24 month cycle from beginning to validation. At the end of that period Technology leaders will often leave for new jobs and the organization will be left with a mandate driven technology platform. The business is often no better off from the technology adventure.
Technology is enabling new consumption patterns which in turn is disrupting businesses. Organizations are experiencing a lot of change. If they hope to compete organizations must adapt and respond to these changes. We tend to focus on vaguely defined outcomes instead of developing the organizational capacity to truly compete.
The focus should be placed on the rate of change, the lead time for those changes, and success rate of those changes. Advancing these measures enable organizations to adapt and respond thereby closing the gap with and outpacing competitors.
Key diagnostic questions to ponder for your team.
1. How long does it take to push a simple one line change in software or configuration to customers?
2. How long would it take for a service, website, or client app to double capacity?
3. When you add more technologists does your rate of successful change increase proportionally?
4. What percentages of changes are treated as experiments against a control (normal)?
The biggest challenge is the wrong focus big bet strategies. We really want a dynamic system of discovery to build great products. Teams need to learn and earn their way into the best strategies and they need to continuously improve those strategies. It's easy to address by running an experiment. Take a single system and make it great. Use rate of change, lead time, and success rate as your guide. Aim to double rate of change, half lead time, and double success rate.
On completion you will have a little super car capable of doing amazing things, and you will want to spread that goodness throughout the entire org. That is where the adventure begins, and the path leads to both innovation and predictable low-cost delivery.
Making software leaders better