Jeff Sutherland said, that in 2006 Palm was trying to build a web operating system. They were running three-week sprints, but doing their testing later. The number one issue on a lean list is that nothing is partially complete. Their management said that they did not have enough testers to do the quality assurance inside of the sprint. So, I suggested an experiment like Edwards Deming would do. We would run a few sprints with the tester inside and also do the same QA in the current process. A few weeks later they had the data and found that one hour of testing inside the sprint took 24 hours outside of the sprint. So I said that that shows why they did not have enough testers. They were shocked. They asked, ‘Does that mean it takes two years to ship something we should ship in a month?’ I said, ‘If you can multiply it does!