Gene Kim (@realgenekim) gave a great keynote at the Building on Success 2017 conference that described how TOC and related thinking can create truly breakthrough operations for technology.
He opened with his own story in learning about Theory of Constraints and applying those concepts in the technology development world. He's the author of The Phoenix Project (my review) with Kevin Behr and George Spafford. It is essentially a version of The Goal as applied in technology development and operations (DevOps).
What is one of the big problems in technology development? Maybe the core problem? Gene described is the "downward spiral" that starts with technical debt - the accumulation of junk that we promise to fix later. This junk (unclear code, shortcuts, fragile integrations) requires cleanup later or ongoing maintenance that makes later changes harder and harder. Further down the spiral, it leads to more and more difficulties in deploying new code, as it has to integrate across more and more fragile systems. The natural reaction to that is to deploy less frequently, until the worst case of being completely unable to deploy new capabilities.
With this as a framework, Gene described four "surprises" that he's had in his career.
Surprise #1: Business value (of DevOps) is higher than we thought. There is a core conflict for IT that goes along the lines of the goal of IT is to enable business value. In order to do that, we need to have stability of technology systems AND responsiveness to changing business needs. In order to have stability, don't make changes (or do it infrequently). In order to have responsiveness, make changes (and as often as needed). The conflict in the past has been resolved in the direction of making few changes and bigger batches of changes. See the downward spiral.
But what if the conflict gets resolved and removed a different way? What if you could get changes applied quickly AND those changes could easily be removed if they happened to cause problems? It turns out there is MASSIVE value is doing this, if it can be done well. Companies like Amazon, Google and Facebook are making thousands of deployed changes a day, instead of once a year or once a quarter. Gene's research has found that organizations which are doing this (even those that aren't the well-known technology companies), they have more stability and deliver more value than those which do not.
Surprise #2: Importance of measuring code and development lead time. If the changes are valuable to the business, then the various lead times can help point you in the direction of the next improvements. Do you need test automation? Do you need to break batches into smaller pieces? How do these things affect your change lead time. Gene has found typical bottlenecks to flow include the speed of environment creation, delays of code deployment, tests setup and run, overly tight architecture, product management ...
Surprise #3: DevOps is for the Unicorns ... and the Horses too. Gene has seen a number of "traditional" (non-technology) companies that have successfully used the DevOps concepts to transform their technical operations.
Surprise #4: Dynamic learning organizations. Gene is of the opinion that the concepts of Theory of Constraints, DevOps, Lean, and others will merge down the line into something that looks likek dynamic learning organizations. Interesting. This is certainly a directly these approaches can go - I certainly see a lot of connections across the various approaches.