But that's a tough concept to teach, or to relate to other people.
That's why I was somewhat excited when I picked up and read Geoffrey Moore's new book called "Dealing with Darwin." The book isn't really about picking what you outsource and what you don't. It is about picking what is core to your business, and what is contextual, and then allocating resources accordingly. But the lessons are directly applicable to the question of what to outsource.
This core / context matrix gets referenced a lot in business. But few people I've heard use this model have a crisp definition of what is or isn't core. Moore actually does a decent job of defining core and context. In a nutshell (to save you the 8 or 9 hours it takes to read the book), core work produces lasting differentiation that allows you to retain customers and hence charge more for what ever it is you do... Context is everything else.
Put differently, core is that which your customers believe they can and should buy only from you, context is all the stuff you do to manage to deliver it.
Moore defines a 2 X 2 matrix, as follows:
- Mission-Critical vs. Non-Mission-Critical on one axis
- Core versus Context on the other
The four quadrants, to summarize with some license applied, break down like this:
- Quadrant 1 -- This is where you innovate and invent. You create new differentiated offerings here.
- Quadrant 2 -- This is where you deploy your new differentiated offerings.
- Quadrant 3 -- This is where you operate. I think it's fair to assume that most IT work happens here.
- Quadrant 4 -- This is work you offload, or cease doing all together.
It's important for the sake of this exercise to separate services (stuff that the engineering or IT organization provides to the business) from a task inventory, for example, this might be a quick task inventory for "Exchange messaging services":
- Active Directory User provisioning.
- Exchange server patch management and server upgrades.
- Storage administration.
- Incident response and troubleshooting.
- Spam filter administration.
- MX record administration.
- User de-provisioning.
- Providing individual users with the ability to send, receive and store RFC 822 electronic messages from within the company, as well as from individuals at other organizations.
Unless your company builds mail servers and sells them, or offers hosted e-mail to your customers, I'd suggest that your messaging service is contextual. Your customers don't buy messaging from you. That said, it depends on your business whether this is mission-critical or not. Either way, this is a Quadrant 3 or Quadrant 4 service. It's contextual. In my book, that means it is a candidate for some kind of outsourcing. Maybe if it's mission critical, you manage the service, but seek offshore staff augmentation. Maybe if it's non-mission-critical, you outsource it entirely, and just buy it from some company that provides e-mail service to other enterprises.
If you go through everything in your service portfolio, and take the lessons Moore develops for markets and apply them to the much smaller fractal pattern of IT services or product engineering, you can quickly develop a landscape out of what once might have seemed to be a melange of unrelated tasks and services. And you can see, in an act of advanced common sense, what you should do about all these services you have to provide to your business.
- Those things that are core, and non-mission critical may need more investment.
- Those things that are context may need offloading, cessation, or augmentation.
I'm leading a facilitated discussion on this topic this week. So more on this later, after I try to walk some people through this line of reasoning.