Check out the latest comments on this story (to say you the effort they are reproduced here) which sees an inhouse IT commentator put an alternative view on the calculations…

The 'savings calculator' is one of the most risible I have ever seen.
For a 600 person Exchange system, it will supposedly cost me £356K over
three years, vs. £50K for a hosted solution, resulting in a £306K
saving. Wow, can't wait for the managing partner to come dashing
through the door demanding to know why we haven't signed up yet.

Hold on though, before you long suffering IT Directors and IT
Managers groan in despair, I can save you! I've done a back of the fag
packet Excel based on their model, but using realistic assumptions

a) Nobody in legal IT writes stuff off over three years and then
immediately bins it. You are going to run it for at least two more
years, so a five year cost model is more reasonable and plenty of firms
would be sweating their ageing assets beyond even this.

b) I'm going to need hardware and licenses for domain controllers anyway, so it isn't a saving.

c) I'm going to need Windows licenses for my PCs anyway so it isn't a saving.

d) I'm going to have Outlook licenses anyway, as I need MS Office, so it isn't a saving.

e) I'm going to need a backup solution anyway, there might be a slight saving on the cost of backup tapes, but that is all.

f) If you can find anyone qualified to run a 600 person exchange
system for 25K per annum in London then send their CV over to me at
once! £40K is more realistic but still on the low side. In practice I
reckon you'd need about 0.3person/year to run a 600 user Exchange
system, not the 1.5 person/year they quote.

g) The calculator didn't even manage to work out their own monthly
per user cost and showed it as zero. I've (quite reasonably) added that
back in!

So I reckon the realisable saving of going hosted vs. running your own 600 person Exchange solution per year is:

They say (over five years and adding in their missing costs):

Own solution: £340K

Hosted Solution: £354K

Saving: (£14K)

I say (over five years and using real world assumptions):

Own solution: £130K

Hosted Solution: £292K

Saving: (162K)

I accept there are plenty of variables that one could add in and
ways to make either option more or less expensive, but using their own
assumptions it would be hard to show a significant saving. Bear in mind
their own model comes out at £14K more costly over five years to go
hosted, and I haven't even included any costs for actually making the
transition (which could be £30-50K without even trying too hard).

When a more realistic model is used based on the assumptions set
out above, the hosted solution comes out £162K more expensive over five
years! Hardly bargain of the century, is it? There might be good
reasons why you would want to go hosted (e.g. reliability,
availability, focus on core business, agility, etc.) and be prepared to
pay the premium, but cost savings are not the reason.