Investors, watching from the outside: "oh reorgs are really dangerous, you need to run lots of 1:1s and make sure the results of the reorg aren't horrible."
This is somewhat useful (it is true that 1:1s are necessary to get a feel for the fallout) but not the most useful advice.
Conversation
The most useful advice I ever got was: "if something screws up post reorg, ask yourself what information channel could have existed that would have prevented the fallout, and go create that channel."
1
8
This is not a new idea!
In High Output Management, Andy Grove wrote about how you should audit your information sources every so often, and seek a few — redundant — sources that balance between a) timeliness, b) accuracy of information, c) effort.
Replying to
Grove sampled customer complaints at Intel each week for this reason. (High timeliness, medium accuracy, reasonably ok effort).
2
5
What I find interesting is that the general shape of the advice applies to metrics as well.
Colin Bryar once told me that a way Amazon used to pick new input metrics is to ask "what metric could we have been tracking that would have warned us of that screwup?"
14