Eating your own dogfood: Google vs. Amazon
Google engineer Steve Yegge let an internal Google memo slip publicly (he later called it "the great-granddaddy of all Reply-All screwups in tech history") where he argued that Google was royally screwing up by not building its social technology and other internal technologies as a platform. He compared Google to Amazon, which does.
If you're interested in understanding the importance of a company 'eating its own dogfood,' especially as it pertains to APIs and web services that even internal departments use with each other, give the piece a read.
Steve argues that what's made Amazon so successful is institutionalizing this into its culture early on.
You can also read his follow-up piece here, which he published after the slip was made public.
I really believe Steve is right about the importance of eating your own dog food, and as a company we're institutionalizing it into our culture. We're always working to 'eat our own dog food' internally, too. For example, we originally built Socialize into AppMakr as a custom feature; now that we have launched Socialize as a robust platform, we're ripping that custom code out and using our public Socialize API and SDK within AppMakr too. It gives us an opportunity to turn us into our best and most critical customer.
Whenever you can, build technology as a platform that can be consumed internally and externally. There's an elegant balance between fast iteration and building out a robust platform, which I can talk about in a later post if you'd like (just leave a request in the comments).