I am pretty sure what Micah means by "great sysadmin" is "already skilled enough at managing services X,Y,Z so that you can do them with 0% chance that a 5-min thing turns into a 2-hour or 8-hour thing".
I could manage all of that stuff directly, but I shouldn't because I don't do it all day long which means that either I'll a) do it wrong or b) have to do several hours of research to make sure I don't do it wrong or c) routine multi-hour interruptions when things break or d) combine a+b+c because that's what will happen in reality.
It's because of this that we outsource the following things that we could do ourselves:
* Email hosting (Tucows)
* DNS servers (Tucows)
* VoIP (Vonage)
* 800# (Onebox)
* External system monitoring (Pingdom)
* General server sysadmin (a friend of mine)
* StreamSend (Email Marketing - not that good, but dealing with RBL and email deliverability is worse)
* Payflow Pro (credit card gateway - BAD customers service o/w OK)
It costs some REAL money, but saves much more in opportunity costs, headache, etc. Plus our services improve as these companies improve offerings.
We only directly manage things that can save us TONS of money or are strategically important. Here's the list of those things:
* SugarCRM
* RT (bug tracking. was outsourced but too $$ so we moved in house)
* Internal system monitoring
* Hosting (we run our own systems since we have tons of custom dependencies)