Business value and reducing IT duplication

Duplication of features is where I consider one of the easiest areas of IT architecture (enterprise or otherwise) to reduce cost and increase the business value of investments in IT and related technology architecture. In general, I take the view that architecture pays for itself (and then some) when brought into, or given greater influence toward, an organization. This post focuses on architectural duplication, the idea being that an organization has software systems or technical infrastructure that are duplicates of one another. Any piece of information technology (IT) has costs; these costs start at the decision making around acquiring (or building), and continue into implementation, and then exist over the life-cycle of the IT item. Some say the cost of maintaining software is as high as 80% over the life of a piece of software, either way maintaining software (bought or built) cost money and having duplication of systems or features increases these costs unnecessarily. I believe that any architect worth their salt, and working in a medium size organization or larger, can easily pay for their salary simply by reducing duplication and optimizing implementation to reduce licencing costs.

Identifying and reducing
What is the best way to identify duplication? Start with creating an inventory of all software and systems (yes, surprisingly not every organization does this). Make sure this inventory crosses all silos of the organization. Do the follow-up, talk with all stakeholders responsible for software, infrastructure, and related licensing. Create a spreadsheet, diagram, database, or whatever works best for you to visualize all the software, systems, etc... make sure you have a good written description of each. Be sure you can discuss the diagram in a vocabulary that each stakeholder understands. If you can get a sense of cost associated with each, get that information... reach out to the help-desk to get information of which systems require the greatest support, look into the ongoing data entry (duplicate or otherwise) costs, find out how many simultaneous users each system has, gather licensing costs. With all this information you should be able to identify duplication across the silos and within the software features. There will be duplication that can be "easily" removed or licensing schemes that can be adjusted.

Regardless of the cost in doing all this work, it will be less than the savings collected from reduced duplication and over-subscribed licensing.

-- If you want assistance with this or similar IT architecture initiatives feel free to contact me;

Peter, the pragmatic architect.