The December 2005 Business 2.0 has an article about Dell's newest manufacturing facility in Winston-Salem that reflects a few theory of constraints principles: Dude, You're Getting a Dell--Every Five Seconds (Christopher Null):
The PC maker's newest and biggest factory makes its other state-of-the-art plants look like slowpokes.
I don't know for sure that they have applied TOC thinking, but Dell are frequently discussed as a company that has taken to heart many of the principles of TOC. One could also read this article from the perspective of other continuous improvement philosophies. Dell have clearly spent a lot of time reviewing their other facilities in the construction of this new one.
The article is brief, but they highlight five areas where they have worked to make the new facility more responsive to customer demand. Here's my take on how they relate to TOC:
- Heavy Lifting: Remove repetitive strain complaints by deploying robots to do heavy lifting or awkward things, not necessarily those that appear to be more "productive." Reduce human worker complaints and you get more consistent processes.
- Supply on Demand: They have built assembly lines that can construct 40 different machines without re-tooling, removing the time involved with setups. Ideally, this lets them significantly reduce batch size, which reduces the time from order to delivery.
- Common Ground: Frequently-used parts are kept at each workstation and a runner restocks them as they get low. This keeps the workstations moving, which is important if those workstations are constraints in the process.
- Team Building: The assembly line workers are grouped together to help them learn from one another. Not necessarily a TOC idea, but Dell discovered that errors drop with these small teams.
- Speed-Read: Do a quick quality test at the final assembly location to check for the most common assembly problems. This gives the workers very fast feedback to correct errors before they get to final testing. This is a great example of getting as much as possible out of the constraint by catching and reducing its errors. And if final testing is the constraint, this helps by sending better-quality machines into final testing. The principle being that you don't want to waste important constraint time on bad machines.