One of the biggest mistakes that IT makes is confusing "critical" with "important". What I mean by this is that IT often thinks that just because the failure of a service or application has major ramifications for the business that the application is also important to the business.
There are lots of applications and services that fall into this sort of category and the way to understand whether something is "important" or just "critical" is to look at what happens when it goes away
When a critical service fails, or is removed, then it has a massive impact on the operation of other services, work arounds are needed and contingency plans are brought in. People really notice it when it isn't there.
Important services however are noticed when they are there they are the reason that the other services exist. If the Important service goes away then there is no point the other services continuing.
In other words Critical is about operations and Important is about business value.
To take this further I'll use the Tour de France which yesterday popped through the UK as an example of the difference.
About 1-2 hours before the race comes through the village there is the "caravan" these are the sponsors cars which dish out freebies to the crowd and are basically engaged in advertising. These are clearly support services, they aren't important to the race... well except that the sponsors give money and without the money the tour would be a much smaller event. So here we have a certain level of criticality for a support service, it clearly has an impact on the overall system but no-one would argue it is "important".
Next up comes the police, camera men and assorted random cars that must have some support job, but again its only support.
Then came the leading bunch of 5 riders, with a Brit in it, followed 5 minutes later by the pelaton. This is the important bit, without those riders then all of the other services, no matter how important, are pointless. There is no tour de France without the riders.
Following them are the teams with the spare bikes, liquids and the like, clearly these things are essential for any rider who wants to complete, let alone win, the race. If they went away then it would be practically impossible for a rider to compete, these are therefore critical support services. Their failure has a massive detrimental impact on the main business service (the riders).
In business the same applies, the phone system and electricity are "critical" to a modern business, but they are not "important", they are a relied upon commodity whose service level is important but which is not expected to create value. The same can be said of many things, HR, Procurement and often IT, and it is certainly true of IT systems. One of the big SaaS drivers is that these new solutions offer an answer to the "critical" question, while not requiring the costs of "important". Too often internal IT create critical systems whose costs far out weight the costs of truly important business services, indeed often these back-office critical elements suck the vast majority of the IT budget away leaving very little to deliver new innovation.
To really build a business case and deliver value from SOA its essential that IT takes an honest and pragmatic view on its current estate and differentiates the support service from the business services and the critical from the important.
So is that application you have been looking at one of the riders? The team support? Or just the people at the front giving away key-rings?