A common business usability conflict gets the spotlight in this fascinating Business Week article about the $100 laptop. You'll really want to read the article, but for a quick summary, Nicholas Negroponte, founder of the MIT Media Lab, has designed a cheap laptop computer for children in developing nations. The laptop operating system, Sugar, was not usability tested before production and Jakob Nielsen criticized Negroponte's methodology.
If you've worked in a company where usability is new, the divide may seem familiar. On the one side is the usability expert, Nielsen, saying "it’s outright reckless in a case like this" to launch without user testing. Indeed, it's very risky and paternalistic for Negroponte's One Laptop per Child (OLPC) to develop and launch a product without studying the needs of the audience.
On the other side is Negroponte's "demo or die" engineering mentality. Anything that delays the launch schedule is a risk, and anyway, no prototype test is as useful as real world results.
The best way to bridge the divide is to employ quick, low-impact usability methods, such as paper prototyping. Prove usability's benefits by fixing bugs early and showing how input from users can improve the product.
Schedules are important, but usability testing need not lengthen the normal design process. Besides, it's hard to believe no time could be found for usability when the long-awaited laptop, called XO, has been two years in the prototype phase alone.