I feel strongly about the damage done by interruption to maker's focus. My appreciation for this damage to system performance was greatly enhanced as I became a software developer.
In my work prior to software development I could see that interruptions were not ideal but they were not so damaging to my work and they do also have value. But work that requires extra focus, and software development sure did for me, the damage increasing greatly.
Over 20 years ago at the Madison Area Quality Improvement Network they had a simple visual management sign at everyones office and cube. You could dial between various settings - GREEN available (free to interrupt)... RED busy, don't interrupt... There were 3 to 5 settings. It worked very well for them.
Of course such a system is highly dependent on the overall management system. Just putting that up in most offices would likely fail. MAQIN was an organization that embodied modern management practices (Deming, lean, respect for people, organization as system, etc.). It worked very well for them.
When building a software development team I tried to protect developers from interruption while having very open communication between developers and program managers (product owners). As is often the case, the real roles didn't line up exactly. In our organization, the people that needed us to develop software (to varying levels) didn't want to take on the responsibilities that agile would suggest (setting priorities etc.). Nor did the that organization want to deal with the conflicting priorities of the various people in their organization had for software development. As part of making things work, I took on a bit of the priority setting roles.
One of the things I did was explain to those that had us do software development was why it was critical to allow software developers to have uninterrupted time to focus. I shared various articles and reports over time and would talk about this as I interacted with them. I encouraged people to talk to me first (at that time I had transitioned from developer/software-development-program-manager to nearly entirely a program manager role).
We also had weekly meetings (when it made sense) with the "product owners", me and the developers working on the software. We scheduled additional meetings when necessary.
I also encouraged people to use email when possible to just check on things, ask simple questions, ask to meet. Again I explained how just going up to talk to the developer could impose significantly higher costs to them being able to focus of their work than there were for interrupting my work or others that had work that could more easily be interrupted with lower consequences.
When there were very important deadlines I would increase these instructions not to interrupt the software developers. And I would talk to the developers about how things were going and if the developers wanted help pushing back a bit I would then do so - mainly by trying to work on the system (provide a set time to allow needed discussions, coach the "product owners" on the difficulty caused by interruptions etc.).
As you might expect this worked better with some people than others. It made a significant difference though. And in our organization it was actually a bit less effective because the software developers were all so nice. They could say how interruptions damaged what they could do systemically but were always super friendly whenever they were interrupted. I was by far the most willing to actually disappoint people face to face in the name of improving overall performance.
It is very hard to overestimate the systemic nature of all of this. By delivering great results we were able to reinforce that if we followed the modified Deming/agile methods we were using were important. Those needing work from us ranged from very skeptical to accepting of those ideas but our performance (both in meeting their desires for software and for interpersonal interactions) made them willing to go along with the methods we said were important.
Related: Deming and Software Development - Mistake Proofing Deployment of Software Code - Creating a Culture that Values Continual Improvement
1 comment:
I personally can say that this is very important, and also that I dod not appreciate the amount John did to make this happen in our environment until I left for another environment and found myself having to have this same battle repeatedly.
It's also worth pointing out that many programmers need to be prevented from being distractions to themselves. I find "Pomodoro Technique Illustrated" a very worthwhile book to help with that. It also introduces ideas like the PDCA loop, systematic measurement, and a systematic approach to thinking about one's self that start to set the foundations for thinking about the picture outside just one's own behavior.
Post a Comment