Class A Lesson 11: No Random Local Optima.

20 Lessons From Implementing Class A Manufacturing

Lesson 11: No Random Local Optima.

If you are a fan of Eliyahu Goldratt (The Goal, Theory Of Constraints, The Race, Haystack Syndrome, Necessary But Not Sufficient,…) this week’s message is a sermon to the choir.

How does it happen that a production order, required for filling customer backorders, can physically be accelerated through the supply chain, but isn’t? The discussions sound like this:

·     Sales: Folks, we’re on backorder for product X so we need to expedite the next batch.

·     Planner: I’m not taking a hit to my Reschedule metric.

·     Purchasing: I’m not paying Expediting fees on raw material.

·     Production: I’m not taking an overtime hit to my budget.

·     Quality: I’m not missing my Lab Turnaround metric.

·     Shipping: I’m not paying expedited Shipping fees.

·     Marketing: Are the rest of you guys crazy? We’re putting our best customers on backorder and that bad news will spread like wildfire.

What’s happening is a classic case of Random Local Optima and Global Sub-optima. Each person is rationally deciding how to handle an opportunity (or problem) based on optimizing their group’s (Local) performance, ignoring their negative impact on the overall business (Global Optima).

This situation becomes more difficult as organizational size increases and as compensation systems rely more on metrics and data. So what should we do?

Management must establish the culture of No Random Local Optima. Require people to assess things from the overall business’s perspective. When an analysis of an opportunity (or problem) reveals that the business will have a net Global benefit by pursuing a path that incurs some specific Local Sub-optima, management must set the example by pursuing that path, explaining how the Global business benefit is greater than the Local pain, and rewarding people accordingly.