These IT analysts also believe that the failures are not widely publicized. Of course failure is not documented as much as success because it makes people look bad. But why does outsourcing fail? TCO reduction is a myth. Sure you might be promised a 10% TCO cut off the start, but when you look at the details, some analysts believe you need to increase your management and logistical cost by 5%. (And what happens if your biggest disconnect is your middle management?)
And other analysts have written that if engagement is broken at either end, the outsourcing is a guaranteed fail. What does this mean? If the company doesn't have its IT house in order, that is, you don't have defined, mature processes, and/or the co-source is in the same boat, the overages are huge. And look at the risk. You might have just eliminated a large part of your business knowledge, and handed ownership and the ability to act over to someone else. Once it's been given away, it's almost impossible to get back. Not to mention some outsources count on the fact the initial employee/contractor count reduction (that yields the 10% TCO decrease) will often exceed the baseline by 20% in five years. By then the outsource is so entrenched, the company is truly at the outsource's mercy.
Core competency advocates say the best way to mitigate this is to invest in your people. IT folks should share the business knowledge with those domain experts. But this is a tough one. If you look at the credentials of many IT folks above the trenches, you'll find a melange of expertise that just might not understand the domain. So a lot of training and cross pollenation needs to be done. And this takes time.
The premise of Gartner's "IS Lite" is to outsource everything in IT, but retain five roles so your IT is a glue layer to your co-source(s):
- architecture development;
- business enhancement (which involves business process analysis), project management and business relationship management;
- technology advancement; and
- vendor management.
This is great if everything fits nicely into a defined package. But what if it doesn't? Most organically grown IT processes don't have well defined roles and responsibilities. Though that is a best practice (as promoted by ITIL), you can see folks getting territorial
when things don't fit the model, and people saying "Work the Role" instead of "Work the Problem". So high maintenance systems, where you need lots of expertise just to keep the thing running, break the mould. Couple this with a bench and co-source strategy, and you have IT that works Projects at the expense of efficient operations. You need to know how your business works first, and characterize the needs of the systems and processes, before trying to paint it all with one brush. (There will always be exceptions, and always quirks specific to a companies business. Without those quirks, what differentiates you from your competitors?)
And that is where the real solution is: efficiency. If TCO reduction is a myth, then process optimization is hard fact. It truly increases the value per dollar spent. The unfortunate catch .22 is that to do process improvement, you need to consider IT a core competency. It doesn't mean you can't outsource; just be careful what you do. Know your enemy and know yourself...