I run a company that does just this (BitPusher, LLC, http://www.bitpusher.com/).
There are a number of different kinds of individuals and companies that do this, and there are challenges with all of them.
Individuals who do this are generally cheapest, and sometimes you can get past the trust issue by using a friend or a friend of a friend. On the flip side, they have a tendency to leave for day jobs when times get tight (or be hard to reach when busy if they're moonlighting).
Companies give you a few advantages -- they tend to be more stable (if one person leaves, you still have backup if they're well organized) and they can have a wider variety of expertise (so the odds of needing to bring in other parties is lower). But there are two challenges here. The first is that if they're not sufficiently well organized, you'll sometimes get someone who doesn't know your environment. The second -- if they charge by the hour -- is that no matter how worthy of your trust they are, it's still in their best interests to bill you for as many hours as possible. This means that most of the successful ones either find ways to get you to buy extra work, or don't do the proactive maintenance they should (so they get more revenue from the things that break).
We've taken the approach of managing IT infrastructure for a fixed monthly fee. This is tricky, but after a couple of years of using this model it's coming together. It's a difficult sale, because you have to get customers to agree in advance to pay for hidden IT labor costs (both ongoing minor maintenance and the periodic major incident) -- small companies are used to paying for IT reactively rather than budgeting for it. Also, you need to become actually good at the things that IT organizations always talk about being good at -- sharing information, doing proactive maintenance, documentation, following processes, etc. And the work is generally front-loaded, so most customers aren't profitable until the second year. And this model ends up having a little more overhead (despite being more efficient per customer) so it's hard for the whole thing to be profitable until you've grown enough to have some economies of scale.
That said, about two years after converting from an hourly and project model to a fixed-monthly-fee model, we find it working quite well (and we're finally profitable again).
Slashdot Mirror
Individuals who do this are generally cheapest, and sometimes you can get past the trust issue by using a friend or a friend of a friend. On the flip side, they have a tendency to leave for day jobs when times get tight (or be hard to reach when busy if they're moonlighting).
Companies give you a few advantages -- they tend to be more stable (if one person leaves, you still have backup if they're well organized) and they can have a wider variety of expertise (so the odds of needing to bring in other parties is lower). But there are two challenges here. The first is that if they're not sufficiently well organized, you'll sometimes get someone who doesn't know your environment. The second -- if they charge by the hour -- is that no matter how worthy of your trust they are, it's still in their best interests to bill you for as many hours as possible. This means that most of the successful ones either find ways to get you to buy extra work, or don't do the proactive maintenance they should (so they get more revenue from the things that break).
We've taken the approach of managing IT infrastructure for a fixed monthly fee. This is tricky, but after a couple of years of using this model it's coming together. It's a difficult sale, because you have to get customers to agree in advance to pay for hidden IT labor costs (both ongoing minor maintenance and the periodic major incident) -- small companies are used to paying for IT reactively rather than budgeting for it. Also, you need to become actually good at the things that IT organizations always talk about being good at -- sharing information, doing proactive maintenance, documentation, following processes, etc. And the work is generally front-loaded, so most customers aren't profitable until the second year. And this model ends up having a little more overhead (despite being more efficient per customer) so it's hard for the whole thing to be profitable until you've grown enough to have some economies of scale. That said, about two years after converting from an hourly and project model to a fixed-monthly-fee model, we find it working quite well (and we're finally profitable again).