There was a piece on National Public Radio this morning that got me thinking about IT infrastructure funding. I think this is especially relevant as companies consider Web 2.0 technologies and ask the inevitable question – “What’s the value proposition?” It’s a great question, unless that question gets morphed (as it so often does) into “What’s the return on investment?”
The NPR article was about the levees along the Mississippi River. Apparently, after the huge floods in 1993, a few Midwestern states got together with the Army Corps of Engineers to draw up a plan to build higher levee walls. Unfortunately, the plan, which could have helped limit the amount of flooding the region is now seeing, ranked the project low on the Army Corps of Engineers’ priority list, because a cost/benefit analysis showed an economic benefit to cost ratio of less than 10%. This was due to the combination of high cost (over 1,000 miles of river bank) and the fact that the land was already protected by existing levees, so the new plan would have only added protection when the water got really high (which, of course, has just happened, and in retrospect, the levee work would have been well justified!)
As someone on the news piece pointed out, this is a classic error of trying to use cost/benefit analysis for projects that are infrastructural in nature. This has been exactly my experience with IT infrastructure. When IT infrastructure (or any type of infrastructure) works well, it is invisible, so nobody wants to fund improvements to it. This is exacerbated by the fact that the economic benefits for a given IT infrastructure investment, while potentially significant, are often indirect and/or intangible. There is extensive research on this phenomena, and a literature review shows that traditional financial evaluation techniques, such as NPV, undervalue IT infrastructure.
There are a variety of preferable techniques such as Real Options that include the effect of project inter-dependencies and recognize that an IT infrastructure project may have a negative NPV when evaluated on stand-alone basis, but the project nonetheless can provide an option to launch future value-added capabilities. Unless the option value of this flexibility is taken into consideration, it is impossible to accurately represent the strategic business value, or justify, strategic investments in IT. A related technique is the Value Net approach to estimate project benefits based on interactions between stakeholders. A Value Net is a map that links a firm to various player segments: customers, competitors, suppliers and partners.
For all of the research on this topic (see, for example, the Journal of Information Technology Management’s “Justifying Investments in IT” by Sasha Dekleva at DePaul University for an excellent treatment of this topic) I find IT management surprisingly lackadaisical on their approach to IT infrastructure funding, with the result that infrastructure is misunderstood, undervalued, and therefore underfunded.
All this, of course, course, depends upon your definition of IT infrastructure. I still find that Professor Peter Weill’s definition works most effectively:
“The base foundation of budgeted-for IT capability (both technical and human), shared throughout the firm as reliable services, and centrally coordinated.”
This gets at the key distinctions of “foundational”, “budgeted-for”, “technical and human”, “shared throughout the firm”, “services”, and “centrally coordinated” (note, not necessarily centrally-managed. This definition, together with appropriate justification techniques such as Real Options and Value Net approaches, together with a strong partnership between the CIO and CFO, can lead to a more robust understanding of the role of IT infrastructure, and a more valuable IT capability for the firm. By the way, these techniques help not only in that they might get an investment approved (and/or appropriately prioritized), but that they ultimately improve value realization by clarifying the cause and effect chains between IT infrastructure investments and the value they enable.