The DevOps Enterprise Summit once again exceeded my expectations. Swarming with witty smart people (who were also fun to hang with), an impressive hallway track emerged. Much learning transpired – below are my highlights in the order they occurred.
- A dedicated Dojo space for Learning
- Challenges: 30 day experiences to introduce and level up expertise on DevOps, Lean and Agile.
- Flash workshops: 1-3 day events to create usable features or solve a problem.
- Open Labs: 90 min sessions twice a week for Q&A and inspiration.
I walked out of the Target talk seeing how dedicated space could be a measure of leadership buy-in for implementing DevOps and continuous improvement.
- The Impromptu Open Space turned out better than I ever imagined.
We set up 2 whiteboards – one for attendees to post topics they wanted to do a deeper dive on and the second to communicate times for when the highest voted topics would occur.
Gene Kim gave me 2 minutes on the main stage to explain how open space worked. I flubbed the first sentence saying “Diaper Deeve”, instead of “Deeper Dive” – a good reminder to laugh mistakes off. It was great to watch 25+ people gather around a long table to Diaper Deeve. Yay for self-organized spirited conversations!
- Breath Customer Oxygen
Jody Mulkey emphasized business metrics over descriptive operations metrics during his Ticketmaster talk. Knowing predictive revenue/actual revenue can activate the business lens for tech teams. Jody described viable career path scenarios for how Ops people and how they can provide more business value – “Own your system end to end”.
- People are willing to sit on the floor to hear about metrics
With the standing room only space fully consumed, people sat on the floor between the room columns to hear Troy Magennis and Julia Wester’s Metrics & Modeling presentation. They showed folks how to (and how not to) use metrics for improvement and how to visualize the data.
The metrics open space led by Eric Passmore kicked off lively (sometimes heated) conversations on problematic metrics, which Nicole Forsgren’s tweet captured succinctly: “If you do comparisons across a single metric, it will inevitably be gamed.”
Passion for metrics was a major theme throughout DOES, which mirrored my observations while working with the metrics team on the DOES forum report.
- Transform with Stealth DevOps
One of my new all-time favorite storytellers, Charles Nelles talked about implementing DevOps from the bottom up and from Ops out. Charles opened with; “this is from the perspective of an old grouchy Ops guy.” This appealed to me because DOES is intended to be a conference for horses and not unicorns.
When asked how his team dealt with management push back, Charles said their stealth DevOps approach and focus on data got them a few wins that showed undeniable improvement. Data was the currency for lowering the “You Ops guys suck” threshold. I adored his dry humor when referencing The Phoenix Project, “We didn’t have a fairy godfather waving a magic wand to help us.”
- Uncertainty is turns out is a shape, not a number
Preparing a talk equals study, study, and study for me. I found myself re-reading The Flaw of Averages, and was so inspired by it that I changed my talk title to include “Uncertainty”. Uncertainty is the essence of variability that flow metrics capture so accurately and that this illustration from the book exemplifies so well.
My talk explained how to use a kanban flow approach to measure the probability of delivering work on time. The Holy Grail is controlling work-in-process queues instead of timelines because queues are a leading indicator of flow time.
- Allow the j-curve to happen
I plan on watching both Michael and Paula’s talks again when the 2015 videos get posted.
BTW – DOES14 videos are still available if you haven’t gotten a chance to view them.
- The Hallway Track is sometimes the best place to learn
John Willis, Mark Imbriaco, Alan Kraft and Brandon Holcomb were huddled around a round table in what looked liked intriguing conversation, so I joined them and leaned-in. Topics ranged from evil SLA’s to pulling the andon cord.
I got a huge kick from hearing that Alan and Mark have the same repulsive response as I do, to the term “Best Practice”. Much of the knowledge work we do is complicated and/or complex where there really can’t be a best practice until cause and effect known.
- Learn by Observing great speakers
Courtney Kissler embodies the mastery of speaking that I strive for. Her elegant yet friendly stage presence, pitch and speed are inspiring. Add Nordstrom’s evolving journey for content and we can understand why Gene Kim brought her back to the main stage again.
From Value stream mapping to Coaching kata to studying cycle time trends, Courtney’s talk with Jason Josephy delivered significant learning.
- Empirical Evidence Trumps Speculation
Elizabeth Hendrickson’s talk on feedback was music to my ears – for a number of reasons. Yes, she covered feedback in full glory, (anything in our process that prevents change prevents feedback), but she also emphasized two more topics I hold near and dear – decisions and visibility.
I’ve been known to say that making work visible is the first best step to starting a necessary conversation. So, it great to hear Elizabeth validate the need to base decisions on data vs. opinions and confirm that, “Visibility is more important than I ever imagined – Just by making work in our pipelines visible, we had better conversations.”