1. W. Edwards Deming's Red Bead demonstration underscored the fact that a production operation generates a random number of defective parts, for which the workers cannot be blamed.
2. A simulation that uses gun targets and histograms illustrates similarly the effect of variation on product dimensions, where the measurement is a continuous-scale number rather than attribute (good/bad, pass/fail) data.
3. Goldratt's and Cox's matchsticks-and-dice simulation illustrates the effect of variation on production control.
• The drum-buffer-rope (DBR) production control system is a pull production system that contains rather than corrects the variation and its underlying effects.
• Single-unit flow, rapid changeover, and elimination of batch-and-queue operations remove the variation, and thus reduce cycle time and inventory.
4. If a project has two or more critical paths, it is likely to finish late unless resources can be reallocated if one starts to run late. There is 50% chance that each critical path will be late.
• The same concept carries over into production management where absence of any item in the Bill of Materials will prevent production. Henry Ford said in 1922 (My Life and Work) that transportation must be absolutely reliable to avoid the need to carry protective inventory, and bought the Detroit, Toledo & Ironton railroad to ensure that it functioned reliabily.
5. An easy spreadsheet simulation illustrates the effect of variation on a service operation with a random arrival rate and a constant service rate. (A somewhat more complicated model assumes that both the customer arrival rate and service rate follow the Poisson distribution.)
• In practice, service managers will have cross-trained workers they can reassign when a backlog of customers develop. The same concept carries over into manufacturing where cross-trained workers can be assigned where needed.