State of Devops 2022: The J-Curve

January 31, 2023

By Marty Henderson

State of DevOps

With 2022 behind us, DORA - that is, DevOps Research and Assessment - has released their 2022 State of DevOps Report, a thorough, 77-page PDF on DevOps as it impacts the success of software development. Sounds like an easy read, right? Well despite it being well written, there's a few important points that can be pulled out and discussed in context of where we see clients and organizations at today on their path to their goals and visions. One of the biggest points that we see often is something they refer to as the J-Curve.

Acknowledge the J-Curve

One thing we often read in headline and in blog posts are about how lean little startups can produce endless work and enjoy mountains of unchained success. How many of those startups keep that momentum though? Even startups that have gone big and stayed around, like Uber and AirBnB, have had slumps before resuming growth. However, most companies that are the pillars of success are not startups, but instead established companies with decades of experience behind them. Our clients range from start ups to these titans but they all share one common thing noted in this report - the J-Curve.

Specifically, as companies adapt to site reliability engineering - whether as a dedicated discipline or a function of another one - there's a lot of adjustments. Not only do people have to learn to use new tool and techniques, but they also must adapt mindsets. Despite our desire otherwise, we can't push out a mindset update like we can a new tool or hold a workshop on it like a technique. Instead, this mindset change takes time to shift to. DORA shows the J-Curve as

image of curve that dips then curves upward

Source: State of DevOps 2022 pg. 28

However, many people expect success to be linear. One you apply a new tool to a problem, it just gets better pretty much straight away, right?

Well, to take a real-life example, I bought a truck a few years back because I was about to buy a camper and would need to haul it places. So, bought the truck, drove it off the lot and to home. Seeing as it was a newer truck, it was actually a nicer drive than my old car. Got it home and parked it just fine. However, the next day, I needed to drive it to the office which had some minor struggles. I was used to the smaller space my car took and the fact I didn't need quite the same clearance room to change lanes and I didn't want to be "that guy with a truck" all over the interstate. So it took me about another minute to get to the office, not bad.

Then I had to park the thing. The parking lot at the office was a bit tight, but I was earlier than most, so I just parked a little further back where I wasn't surrounded, but it added another minute or two of walking to my desk. So, nice, smoother ride that didn't sometimes die randomly for the cost of about 3 minutes longer door to door. Seems like a good trade-off of reliability for speed.

However, I bought a camper a few month later. If you have never tried to pull 20' of camper behind a vehicle (without swaybars), it's a new experience. Instead of the normal 2ish hours home, it took nearly 3 of going slower, stopping and checking to make sure things were attached, and finding a gas station I could pull through with the camper on it. My normal efficiency of driving plummeted by 50% while bringing the camper back.

The ability to camp without a tent was my goal, however, and so with a lot more experience of driving the truck and the camper places I got faster and better at it. At this point, I am pretty confident of driving the camper wherever and enjoy camping again. Thus, my goal of enjoying and camping more often and being efficient have been met, but there was a solid 18 months of learning to drive the truck differently (especially in the snow up here in Wisconsin), pulling a trailer, and generally adapting to an entire new way to go places.

This is how the J-Curve works, not only did I need to learn a new tool (driving the truck) and a new technique (moving the camper), I also had to get over the fear of sliding the camper into other cars, thinking ahead on my turns, and planning places to take the camper. However, today in 2023, I can take the camper many places and I am no longer concerned about driving it, but rather did I remember to pack everything. As you look to adapt to new tools to ensure reliability and the ability to pull your organization to your vision, just realize there will be a time they slow down and learn. During this time, the whole organization will need to adapt to a new way of thinking. You can't just jump from a little car to a truck-and-trailer nor can you jump from a classic software design to a powerful, robust, and scalable one without several months of adapting to it.

Summary

The J-Curve is one factor of a DevOps journey - a journey you don't have to take alone! A good consulting partner, like Ocelot, can help make sure you follow the best of practices, keep your metaphorical trailer out of the ditch, and keep on your journey. Be sure to page through DORA's report above and see what part of the journey you're at and see if we can help you handle more than a J-Curve problem!

About the Author

Marty Henderson

Marty Henderson is a Staff Consultant at Ocelot Consulting. Certified Architect in Google Cloud and AWS, Marty leads and designs complex projects based on containers and cloud, with heavy emphasis on teaching and enablement of teams.

Acknowledgements

Header Photo Photo by John Lockwood on Unsplash

Peter, et al., "ACCELERATE 2022 - State of DevOps" https://services.google.com/fh/files/misc/2022_state_of_devops_report.pdf