Here's what he wrote about yesterday's "contrast: thinking about hardware" post:
The major issue has been cost - adding function with no revenue even with the promise of future revenue makes no sense if the function has significant cost, which can in the hardware manufacturers parlance be pennies.. therefore the only functions that the manufacturers will gladly upgrade with software are those things for which the hardware is already there. In the age of microcontrollers that's a lot more today than in the past. Also service upgrades such as call home services may be targets because the added hardware would be borne by the customer."If you own the hardware and software you can do this," he commented later on. "But once you break apart hardware and software - the whole motivation for delivering unrealized functions went away. I would think medical equipment would be the ideal candidates for this, where all the cost - like in an MRI machine - is in the magnets and the installation, and the controller and tech doing the analysis is only 5% of the cost. If can monitor licenses too. . .that would enable folks to be more free about not just delivering hardware but software that's not activated until you've got the right license. This gives you the freedom to upgrade software and hardware with a single software function, because it minimizes the risk of not getting paid."
The difference between hardware and software is gross margin, even Intel who have had the best gross margin are under pressure but software still gets in excess of 80%. Development cost are written off as incurred but manufacturing cost are in inventory and must be recovered in the price; software recovers development cost thru license rev, but since the cost are written of in prior periods the rev goes to the bottom line in terms of profit.
So once you get the initial capital to fund the development and write off the cost, the rest is mostly profit minus sg&a... so if you want to be rich start a software company.
I think your idea will work today for products that are based on microcontrollers so sort your targets by that metric.
1 comment:
Thus, in building systems, then, I believe the traditional rule would apply: hard now, soft later.
ahem.
Post a Comment