The Downside of Shared Services

As companies continue the drive towards fewer data centers and a base of centralized applications running on virtualized servers, IT is re-entering an old way of doing business: shared services.

While this may be good for the cost-conscience CEO, the bottom line for CIOs is they are will probably—if not already—face the reality of charging business units for the IT they consume. On the surface, this may seem a fairly straight forward matter. Simply divide total IT spend by head count or some other common metric and send out the monthly bills.

But, as with most things IT related, the reality is usually more complicated. Sure, you can, and many companies do, use very simple metrics to bill for IT. But what happens when the head of marketing questions it’s bill after a lunch with the head of accounting, which has consumed far more resources—network bandwidth, server time, developer’s hours, etc.—but has fewer employees. Those complaints will go straight back to you, the CIO.

What about things like phones service and email? Do you bill based on just having a phone or email account or you bill on the number of calls or emails sent? What about the utilities that manage the servers these applications run on? The power they consume?

“The model (companies) are moving towards in the first instance is a cost-recovery model,” said Rob Stroud, VP of Brand Strategy for CA. “All they’re looking to do is to agree to metrics to justify the budget IT spends. In a shared-services environment we actually can’t go the old way by … splitting up the costs because no body agrees with it.”

What to Measure

Of course, IT can measure and monitor many things today: CPU cycles, bandwidth consumption, IP addresses accessed, etc. but the trick is to find measurements that actually means something to business units that will be billed based on those units. If you get too granular, then tracking those metrics and managing the billing based on them gets unwieldy and complicated. This may end up costing more to administer than some of the bills are worth and may lead to endless rounds of meetings to decipher, decode and rectify the bills with line-of-business managers.

Take databases and storage: two IT services generally consumed on a broad basis, accessed by multiple applications from different divisions and generally very hard to partition for billing purposes. What about enterprise content management apps? Collaboration apps? ERP? CRM?

Again, these apps are consumed by a broad cross-section of your enterprise to a lesser or greater degree and are generally very expensive to buy and maintain. These systems are also prime targets for rationalization and consolidation as many companies do their best to get app-creep back under control. So, what do you do?

To keep yourself from being burned at the stake, suggests Steve Pickett, a past-president of the Society of Information Management (SIM) and currently the CIO of a multi-billion dollar company, start with a business definition of the services IT provides. Use terms like telephone v. VoIP, storage v. SAN or NAS, computing time v. CPU cycles, etc.

“It’s got to be something the user can touch and feel,” said Pickett. “He can see his telephone but he can’t see his calls, for example.”

In Pickett’s experience some of most successful chargeback operations work from a services catalog. IT publishes a price list of the services it provides and managers can pick and choose from the list as they see fit. Of course, in this environment the services must be two things: of high quality and fairly priced.

“It needs to be set up in such a way that the consumer or the end user can look on the open market and find he’s getting value for what he’s paying,” said Pickett. “He shouldn’t be able to buy the same service on the outside for a lot less.”