Last year this time, I was working for a corporation (I don't want to give names here, but hi guys! :) ). One of my pet peeves working there, was the "eat your own dogfood" decision, which had a CRM system forced upon us.
The explanation I had always gotten was that, if we showed that we didn't have enough confidence in it to use it internally, no customers would want to buy it.
It's a good marketing decision, but that's just it: it's a marketing decision, not something you'd impose on your developers (though, in this case I guess it's also something you'd impose on your developers).
I realized how wrong this is, as I read an article about World of Warcraft, where Wartenberg talked about how "eating your own dogfood" means “actually using your product before subjecting the consumers to what may be an ill-conceived and poorly designed product.”
Read that again: the first reason to use your own product internally, is to improve on it, to take it back to the design board. That is, to not shoot yourself in the leg, when you get it out in the market.
Think about it!
Then, please think about it, some more! (because I still have friends who suffer from that thing, every day).
Edit: I just remembered: the software had been bought, along with another company, probably for their marketshare (so, that lays the blame - if any - somewhere else). In fact, I doubt that, said company had been using this software internally, though they might have been.
Still, all this ended up as an environment most developers hated, something to be avoided, and part of the reasons I left (not because of this specific product, but because of the attitude that made decisions like this, the Modus Operandi).