I heard a distinction made recently that was highly unnerving. The distinction was between “legacy” (meaning mainframe) and “modern” (meaning their ERP package).
The package in question has been in place — the same release, the same product — since 1998.
As its fourteenth year in production continues, I’m forced to ask: at what point does something become legacy?
For most people I talk to, “legacy” meant “old”, a convoluted application written a long time ago that can’t be replaced easily and whose quirks and costs must just be lived with.
I tend to take the view that a good working definition of “legacy” is “in production”.
In other words, once something is delivered, the countdown to its removal has begun. Any other view lets you make the error of holding on to it too long, making it into something that is too complex, too risky and (possibly) too expensive to do something about.
By that definition, fourteen-year-old ERP packages fall into the “legacy” category.
Expensive enough when originally acquired and implemented, the thought of spending all that again just to make it up to the current release is enough to frighten off any planner or budget keeper. Yet isn’t this part of what we do with even older systems once written in house?
During the economic lifespan of an idea, there is a time to get the requirements right, a time to deploy the solution, a time to add onto it and refine it, and a time to replace it.
Hitting each of these marks properly is the hallmark of superior portfolio management: ensuring that each asset is enjoyed during its proper economic life, and replacing it while the costs and risks of doing so remain lower than they will be when forced into action.
But why replace early? Simply put, we are moving discretionary spending into the base budget by doing so.
Planning for annual change — just as we do for infrastructure — means that the economics of IT stabilize. Forcing the issue also forces out onto the table the consideration that these are corporate assets, not just the property of one department or another.
It is the job of IT leaders to manage corporate assets for best economic performance while anticipating the needs of the organization. Without plans to replace “legacy” in a timely manner, frankly, you’re just not doing enough of your job.