This website covers topics on knowledge management, personal effectiveness, theory of constraints, amongst other topics. Opinions expressed here are strictly those of the owner, Jack Vinson, and those of the commenters.

Variability is good for you

FishburneSo often in business operations, the focus is on removing variability or eliminating waste or load leveling.  But how often do you hear about these projects and wonder whether it actually does anything for the bottom line.  I could write about why that is, but what has tickled me today is the idea that variability and "waste" can actually be helpful.

Really?  "Waste" can be helpful?  Isn't variability the focus of all those Six Sigma projects that GE and everyone under the sun are running?  Well...

Is excess capacity "waste" or good business?  Think of what happens when there is a glut of work, or if there is trouble that demands a quick response.  Without some excess capacity, it will be impossible to respond without Herculean efforts.

Variability is a bit harder to justify, it seems to make sense that you want to reduce variability everywhere.  But in most cases, it makes no difference.  There is still one step in your process that limits the overall throughput of the system.  If you reduce variability elsewhere, you don't increase the capacity of the system as a whole.  It's a waste of time and effort to reduce variability where it doesn't matter: and people get annoyed that all their efforts don't have a bottom line benefit.

I recently heard another nice connection between variability and innovation (on a TOC-related group on LinkedIn).  How will we know if can do better (faster, different) if there weren't good days and bad days to compare?  Isn't that what inspires the questions of "what was different here?"  If every day is the same, it's easy to get lulled into the belief that "all is well." 

Even worse is the idea of "load leveling."  Number one, it is nearly impossible to do: you'll more likely get people padding their time to appear level - and show a lower capacity.  If you manage to set every step to have the same capacity, what then?  What do you do when you need more capacity from the system?  You need to elevate the entire system, rather than having one place to focus (the constraint of the system).  And what happens when you have variability in a balanced system?  Any breakdowns, slowdowns or problems will ripple through the whole system, reducing the output.  And that loss can never be recovered.

Where should you reduce / remove variability?  You only need to do it in places that are damaging your ability to produce.  Does that business process frequently cause project trouble?  Is it always machine X that delays runs on your critical operating step?  Are deliveries from supplier Y so variable that you need to tie up funds in excess stock of their products?  You don't need to attack process, machines or suppliers: just those that are having a negative impact on your bottom line.

Feel free to add some more examples in the comments!

[Image: Drawing by Tom Fishburne in his article quality (out of) control thanks to Dave Snowden]

Buffers of risk

KM on the front line vs the back office