Running Lean, Running Fat

by Venkat on February 10, 2014

The term lean has become way too overloaded with complex management ideologies and procedural scaffolding. This is one reason I have come to dislike it. Leanness is a variable in a process optimization model, not an ideology. Cutting the fat is not an unqualified good idea and can actually be dangerous. Lean as ideology often turns into anorexia.

But the core idea is a useful one. In any work process, leanness is simply the amount of work in various stages of completion between input and output. Stuff near the input is raw material inventory, stuff near the middle is parts inventory, and stuff near the output is completed inventory. The sum is Work-In-Progress (WIP), loosely speaking.

I suspect lean became an ideology because of the obesity of 1970s industrial processes in America. Now we’ve come to the other end of the pendulum swing. Anorexia is widespread.

How do you manage leanness without ideology?

I’ll assume a manufacturing context, but this stuff applies to any kind of work.

In operations research theory, WIP inventory is trapped money that is losing value. This is a useful way to think of some kinds of work: a stream of smaller output artifacts that are individually only a small fraction of the output. A single spoon is a tiny fraction of the output of a spoon manufacturing line over the course of a production run. Spoons that sit half-finished in the factory might rust and require “rework.”

But as the elements of the output get more coupled, leanness gets to be an increasingly dangerous idea. If there is only one output, such as a commissioned power plant, a completed book, or an opened bridge, this can be a particularly dangerous way to think. It may be worthwhile to “run obese”  — finish sub-parts out of sequence, maintain high WIP inventory, and buy up raw materials when the appropriate function of cost, degradation and storage costs is optimal.

Or to put it more simply, how lean you run is simply a function of three buckets of variables, corresponding respectively to the variability in external input boundary conditions, the sunk and variable costs of project operations (warehouses rented, salaries, material degradation rates), and degree of irreducible coupling in the output.

Lean ideologues tend to think that you can always manage the input boundary variability away so that somebody else is dealing with the problem of releasing inputs to your processes in a just-in-time (JIT) way; that the costs of project operations are always high; and that couplings in output can always be eliminated by clever work-breakdown structure (WBS).

None of these is always true. All three together are almost never true. You can get conditions that are so chaotic, and cost structures that are so weird, that it pays to just get all input whenever you can, process material whenever resources are available (because rework is unlikely and things have to get done sometime for sure), and assemble output as parts become available. This is running really obese. It’s the right strategy under some kinds of uncertainty.

One reason lean ideologues don’t get this is that they reduce everything to financial terms and apply ideas like inventory carrying costs to extreme levels.  In practice, only some kinds of process obesity actually contribute to project costs or affect actual returns. So long as you manage the leanness in this subset (in simple conditions, what is known as the critical path), that’s enough. The more complex and unique the project, the more modeling it financially becomes an art rather than application of a book-keeping formula.

But even when you cannot gauge conditions accurately, there is reason to run obese where possible: it takes far less thinking, and cognitive resources may be your most precious ones. This is one reason artists and other creative types often seem to have messy, obese processes of creation.

Share

Previous post:

Next post: