Go slow now to go fast later.
Again I shall recomend you read Jason Gorman's blog. Speed Comes Later. Start Slow. Focus On Getting It Right. I would go one further than him and say that the same thing applies to performance goals in the software. Make it go faster after you make it go right and good.
I have come to sunny Denver Colorado this week to catch up with my MDM team mates who work in the Denver Tech center. My flight was quite delayed. The first airplane displeased the captian and he sent it back for another. Who am I to argue? The second plane has some issues. We got to the runway four times and aborted three times due to a gauge that went dark whenever the engines were throttled up. I don't know if anyone actually fixed it the third time, we flew without it or it kept working. The other passengers were a mixture of annoyance at bewing 3 hours late and alarmed that two of the planes had problems.
I can't help wondering how good the "Sustained Engineering" effort for the airlines are. I know in software it often gets the short end of the stick.