How do you measure the success of a shift left initiative?

It depends on what are we left shifting. In any case we need define outcomes, metrics to indicate the outcome and a deterministic way to measure the metrics. The Trend of the metrics will show the success of the transformation.

Architecture impacts are hard, but there is some low hanging fruit like eliminating clones. That was one of the metrics we looked at. If you have duplicate capabilities, you can identify if you have good architecture design or not. Or if you have outages and incidents, is there a way to connect that back to a resiliency pattern embedded in your design? 

We have operational metrics and we are just starting to generate architecture outcome metrics so that we can start to correlate. I cannot yet show cause and effect to prove that operational health was better because there was good architecture, or vice versa. But that's where we’re heading. What we have seen is that people are coming to us for help, even with some of the softer things. People are asking about things like how to apply the complex governance and integration rules when putting a customer list together following an acquisition.