The "eat your own dogfood" approach has never sat well with me. I think of Microsoft forcing it's employees to use Bing and Windows - rather than improving the products it seemed to lower morale and ignored the reasons why people were choosing Google or Apple.

If the idea is to improve the product by actually using it I think a better approach would be to "eat everyone's dogfood including your own," because so many good ideas come in bits and pieces from competition and new ideas can come from surveying the entire market to see what's missing. It's not just about experiencing the good/bad of your own product, but seeing where you fit in the market, who would choose your product, and why.