We have a lot of security protocols that the IETF does not deploy on its own sites.
There is thus a push from many quarters to have the IETF eat its own dog food which is of course an excellent idea.
But could I just point out that the point is not to eat the dog food so as to convince ourselves just how tasty and yummy it is. The value of a dog food program comes when the manufacturer realizes that there is a reason the dogs don't like it and changes the recipe.
Without feedback, the process is like doing usability evaluations in a lab and taking no notice of the outcomes.
I am seeing a lot of groups demanding the IETF eat the dog food but rather fewer asking why they dog food isn't being eaten.