It's a long time since I've actually had to do any of this - so I hope I'm not out of date. Hopefully other posters can correct anything I might have gotten wrong and provide you with their own perspectives.
If you go with backing up the data only, in the event of a server crash that corrupts your system (assuming the hardware is still running properly) you may have to reload the operating system and all software from CD, find all your keys and enter them, find your (uptodate!) documentation of your set up and reconfigure the server. This includes, for instance, recreating all user accounts and work groups.
This could take a a couple of days (during which your server will be completely down) depending on your set up, documentation and consultant. And you will probably have more little things to fix afterward because your documentation won't be totally up to date and somethings you just won't be able to recreate (accounts will have to get new passwords for instance).
You'll almost certainly need to pay your consultant to do all this - possibly at his/her call out rate.
And it's not just a server crash. If you run updates that have unintended consequences - depending on your current set up - you may have no way to roll back to an earlier configuration without that full copy (which might again mean a full install from fresh or suffering with whatever issues it brings up until a fix is found).
But if your server crashes and the hardware itself needs replacing the advantage of a full server backup often disappears. As, most likely, your replacement hardware will not be identical and won't be able to take a straight copy from your original setup.
To work out which way to go you should consider (i.e. guess at):
How likely a server crash that corrupts your system (or nasty upgrade) is and how much your consultant would charge to do each type of install (i.e. a full fresh install or copy from a tape drive).
What the likelihood is that you will have all the documentation and software media you need up-to-date and available and factor that into your calculations of how much recovery might cost.
What your organization's attitude to risk is - i.e. do you prefer to minimize risk or do you prefer to minimize expenses spent on planning for disasters? For instance some organizations find that disasters knock them off their feet and they can't get back up easily - fairly smooth sailing is essential to their culture of success. Others find that finding funding to cover the cost of disasters is easier than getting funding for contingency planning (and others just turn away from contingency planning and hope ;-) ).
You might also consider whether a full, fresh install would be a blessing in disguise from time to time. Sometimes there are all sorts of improvements that don't get done because of the expense of reconfiguring. If you could bank the difference between the $1400 and the cost of the 1TB hard drive system and use it in the event of a crash could you then put in place improvements you otherwise couldn't afford and would your organization be better off?
There are other pros and cons - tapes give you a better option to store data for longer (annual backups kept for several years could be handy if you ever find a discrepancy in, say, accounting between year end figures) and are easier to take off-site. On the other hand restoring from tape can be a nightmare, especially if you no longer have access to the machine the tape was recorded on. And tapes are more easily lost or poorly handled which could put your data at more risk.
Finally TechSoup has a nice basic article on backups from ONE/Northwest:
http://www.techsoup.org/learningcenter/software/page6089.cfm?cg=searchterms&sg=backup I suspect you've already covered most of what's in it but skimming it in case there's something you haven't considered might be useful.
Helen