Thursday, 4 August 2011

On Finding Savings Through IT Efficiencies

Today the Government of Canada did a shit-stupid thing by announcing "saving $100 million to $200 million annually" by creating "a more efficient and secure information technology system, but also a cheaper one".

They stressed the "initiative will have no net cost"

Oh really?????

How do you consolidate "a hodgepodge of email systems, data centres and networks" that use "more than 100 different email formats, over 300 data centres and more than 3,000 'overlapping and unco-ordinated' electronic networks" running on "aging IT systems" with no net costs?

Short answer, you don't. It is impossible. As in reverse the Earth's orbit around the sun level of impossible.

Even assuming that the government has current systems that can handle the additional load (which is about as likely as me winning the lottery this weekend, and I haven't bought a ticket), there are still project maps, conversion tests, programming, testing cycles, etc. that still have to be performed. These don't come cheap. This usually requires outside consultants to help with the additional workload and to keep the project focused and moving. These people do not work for free.

In reality, what will have to happen, is that all new equipment and software will have to be purchased, tweaked, modified, tested, in addition to the above. As data is migrated to these systems, they will run in parallel with the existing systems until all the tests and training are completed, and then the old systems are shutdown and everyone cuts over to the new systems. This is how it is done. This is how it is done properly. And yes, it costs a lot of money. In the case of the feds, probably several hundreds of millions expensive as they want to do a few hundred of these conversions. Can you say KA-CHING? This will be an expensive process, and will take several (not two or three) years.

Don't get me wrong. This process probably needs to be done. As Sheila Fraser's audit found, much of this process is past due, and if not done soon, we risk system failures of over-worked tired older systems. But to claim there will be no net costs is a bald-faced lie. The only way to hide the costs is to annualize them out over a 25 or 50 year period, probably beyond the life of the systems migrated to. Trust me, that will require some accounting trickery that will border on the illegal. The net savings of efficiencies made will never add up to the costs invested. Information Technology is a cost center, a money sucking pit if you view it uncharitably (and oftentimes I even see it that way). I will guarantee you that after this conversion/upgrade process is complete, federal IT costs will be the same, if not more, then they are now. But if done properly, we will have a stable, secure, faster performing system. But that won't translate into any direct savings. It never does.

Don't believe me? Look at it this way. If you are in 3 old office buildings that are in desperate need of repair, and they are too small to meet your needs, what do you do? You build one new bigger office building and move everyone into it. Is there a cost savings to doing that? No. The cost of construction will just have to be absorbed. But it still needs to be done, just to keep your organisation going. IT system upgrades are the same damned thing.

UPDATE: the Auditor General..said that it would cost $2 billion for just five major entities

3 comments:

M@ said...

The gun registry was another brilliant software initiative that was going to cost nothing and deliver everything. The political right still uses that as a rallying cry to angry up the blood in its troops ("billion-dollar boondoggle!").

I agree with you about how this will play out, but it'll never be interesting enough to be a headline in the Sun, so I'm betting this will be the last we ever hear of this.

Catelli said...

What concerns me is that the cost-cutting angle will cause the Feds to screw this up and make a bad situation worse, or the costs will cause them to cut other essential services or both.

I do not discount their ability to spend tons of money, screw the system up worse and cut services elsewhere to cover the cost of their screwup.

Christopher Parsons said...

You mean you can't just outsource all this crap to Google for free?

:P