Bench Marking

Forces:

Having an OnsiteCustomer to "sit with the team, ... answer questions, resolve disputes, and set small-scale priorities" is expensive. For very large scale projects (like designing automobiles) with long lead times (like 2 to 8 years), having an OnsiteCustomer may require many customers, and may require replacing them regularly (as they may start to think like the other members of the team, instead of like "real customers"). See also: TheMachineThatChangedTheWorld.

Your competitors (or even companies in other industries) may have already solved similar problems or satisfied similar customers.

Therefore:

Benchmark easily measured features of other companies' products and processes. Use these benchmarks as the Design Requirements for your product or process. See also: BigDesignUpFront, TheDeadline.

But:

Benchmarks do not capture everything that is important to the customer.
Benchmarks do not set priorities for different features.
Over-reliance on BenchMarking can lead to designing the wrong product, and delivering it late. For example:

"It feels like a car designed by gifted engineers who researched diligently what benchmark numbers it needed to generate," said Markus. "It generates those numbers, but the engineers aren't car guys." -- Car and Driver, September 1999, describing the Oldsmobile Alero GL.

Discussion:

I'm not sure if this is a ProtoPattern or an AntiPattern. It certainly has more than 3 applications, though. -- JasperPaulsen