Tribal knowledge is loosely defined as unwritten knowhow that is required to do your job. In a perfect world, all processes would be fully documented and any new hire could read a manual to learn everything about his job. Even the location of the nearest washroom would be documented (after all, that is the first item of tribal knowledge each of us learns the first day on a new job).

While the location of the washroom is easily discovered and probably doesn't need documentation, some tribal knowledge is impossible to discover without going through much of the original design or testing of a product. You either have to be told by someone who knows, or it may be lost forever if a particular employee leaves the company. Lost tribal knowledge that must be recovered results in wasted time and resources.

Years ago at a giant aerospace company, an engineer designed a new hydraulic pump that was more efficient and lighter weight than existing ones. He had five of them running in the engineering lab and their performance validated his design. The company commissioned an initial run of twenty five pumps built for testing and marketing purposes. When assembled by the production group, not one of them worked! The engineer and his top technician started disassembling the pumps, and his technician said, "No wonder they don't work—they forgot to mill off this little extra piece on the frame". Was this step documented—no. Who knew about this issue—just the one technician.

Even this minor upset because of tribal knowledge caused wasted time and money—and greatly embarrassed all involved.

A lot of tribal knowledge is caused by what I call lazy engineering. I often hear, "Our spec doesn't have to have everything documented because everyone knows that missing piece of information." One recent example of missing spec data I've seen is the voltage thresholds of on-off electrical signals on aircraft. While many engineers think in terms of on (zero volts) or off (maximum volts) for both the sending circuit and the receiving circuit, that approximation is not achievable in the real world, and doesn't define how the needed design margin will be realized. Sometimes it is a good idea to have someone outside of your group or even an outside expert review your specs, looking for missing items that are known to the group, but might not be known by new employees or those outside of your group.

Besides making life harder for new hires, tribal knowledge forces your experienced employees to tutor each new hire. This diverts some of your valuable resources to a repetitive task that is easily avoided.

-Don Burtis