Originally published in the Cutter IT Journal.
In the August 2004 issue of the Cutter IT Journal, I and a number of my colleagues explored the difficulties of measuring and demonstrating ROI for IT investments. While the articles explore the challenges of ROI from the perspective of both the process taken and the tools used, there is another angle to IT justification that bears consideration. If we look at the types of IT investments we make and the rationale that underlies each, ROI provides a clear basis for analyzing some benefits – but it may not be the best form of decision making for all of them.
Infrastructure investments represent a particular obstacle in trying to use ROI. Paul Strassman raises the issue in part with his discussion of justifying the operating costs of infrastructure – and reasonably argues that any infrastructure investments should result in cost reductions, improvements in operating effectiveness or strategic gains. While very reasonable, the difficulty is introduced when you pose the question “Compared to what?”
Leaving aside strategic gains for a moment, both cost reductions and improvement in operating effectiveness assume wringing efficiencies out of an existing base. If we maintain the same level of functionality and capability, this becomes a straightforward exercise. It’s when we make changes in our infrastructure that the complications arise. Being able to justify IT infrastructure investments requires answering a seemingly simple but highly loaded question – “Just why do we invest in infrastructure?”
If you’re happy with the operations of your systems, and you don’t need any more functionality, further investment isn’t required unless something breaks – the very reason that more law firms than many would realize still run WordPerfect 5.1 on Windows 98. Moore’s Law of doubling performance every 18 months has resulted in most of our companies having a far greater level of infrastructure capability than our legal friends.
The biggest hurdle of justifying infrastructure is that it does not deliver value on its own. What it does create is opportunity – unutilized capacity on which strategic investments that do create ROI can be built. Upgrading servers and operating systems, implementing new network bandwidth or changing up database versions is a lot like twinning a highway – for a while after construction, the utilization of that capacity is the same as it was before. Sooner or later, though, new uses for the capacity are found and traffic densities start to move closer and closer to where they were before the investment was made. In those first few months, though, a lot was spent to create a capacity that wasn’t used – it’s only when a decision is made to use infrastructure that it actually provides a ‘return’.
There is, of course, the argument that infrastructure needs to be replaced or upgraded before it fails. While this is true, and the costs have been demonstrated in instances where failures have occurred before an investment was made, this argument still suffers from the problem that we are trying to put a value on something that will no longer happen. If we replace a server before it fails, what would the cost and impact have been if it did fail? When would that have occurred? How much would the ‘savings’ be if it failed earlier or later?
While the barriers are significant in justifying investment in infrastructure, the alternative is much worse. Forcing a justification of ROI for investment in infrastructure is one of the underlying reasons no-one believes IT investment numbers, simply because infrastructure typically does not create a direct return. This means that projects either don’t proceed until they are too late, with the cost of recovery from a failure almost always being greater than the cost of investment, or the investments get ‘carried’ in strategic projects that do provide a sufficient return to still make them happen. Once one project has justified the infrastructure, however, every other project that comes along gets a free ride – the capability is there for the using, no rationalization of investment required.
There is no question that investments in strategic technologies can produce a significant benefit, with double-digit rates of return. In almost all instances, however, these are business investments – approaches that result in new business models, processes or means of interacting with our customers. ROI is absolutely the right measure, and the more the merrier.
Our infrastructure investments need to be looked at differently, however. They are not exclusively about keeping the lights on and the plant running. Nor are they about realizing strategic goals. They live in a middle ground between these two points – essential and valuable, yes, but not for their own sake. Their value lies in not what they do, but what they make possible. It is this possibility that needs to be measured, not a direct return that they cannot guarantee.
Measuring possibility is a much harder thing than committing to a tangible outcome. It means willingness to truly ‘invest’ in the future – because it is feasible to visualize how potential can be realized, and the strategic investments that will utilize this new capacity are reasonably understood. It requires different rules of investment, though. Approaches that address this can be as simple as apportioning the infrastructure investment over all strategic initiatives that will utilize it. Or we may choose the harder road of trying to truly measure the possibility the investment represents. But ROI isn’t necessarily always going to be what gets us to the right choice.