From John Hempton (here: http://brontecapital.blogspot.com/2010/09/a123-or-how-to-lose-money-on-youtube.html):
"To make real money in tech your company must do three things. Two out these three and the results will be (at best) inadequate.
You need to have an idea that effectively changes the world in some way (even small ideas are OK as there are surprising profits if you can pull the next two tricks).
You need to execute – that is you need to bring the idea to reality.
And you need to keep the competition out.
Of these normally number 3 is the thing that trips up tech companies – they work really hard to get the idea implemented and then someone with less expense – and with the benefit of watching your failures, trials and tribulations – copies the idea (usually slightly better or less clunky) and the margins go to zip. Microsoft is such a fantastic company not because they have the technology right – but because people build on them proprietary software (developers, developers, developers) and that makes people reluctant to change even if the new product is superior.
But a tech company can easily be tripped up on the execution phase as well. (Anyone remember Friendster? Remember when their site took 2 minutes to load a page because they couldn't get the IT implemented properly? And look how valuable the position they lost is…)."