IT Value Measurement

I'm in a professional group that meets the first Friday of every month. This morning's discussion was led by Joe Lengyel, a senior manager at the STL Federal Reserve. He recently completed his PhD studying ways that IT value is measured and, more importantly, why it generally isn't.

You would think that companies that are spending hundreds of millions of dollars annually on IT would be at least a little interested in whether they are getting their money's worth. Yet, there is pretty low adoption of formal metrics for measuring IT value.

Many good points were raised. My take on all this is that, while it always makes sense to perform some introspective analysis, the real question of value should be put to the business. After all, they are the consumer and have the greatest interest in spending IT dollars wisely. In my experience, the best way to do this is to bill IT services back to the business. And, not in some amorphous blended allocation but in real dollars spent on work pertaining to specific business needs. This can be problematic for infrastructure since it's pretty much impossible to say exactly how much each business area is benefiting from having a decent network, but for applications development it's usually pretty clear.

Of course, this only works if the business has direct control over what's delivered (taxation without representation is tyranny!) We meet with our business users every week not merely to give status, but to get their direction on where we should be focusing our efforts. Technically, we chunk it out in 3-week iterations, but the mid-iteration feedback is useful, too. Macro-level budgeting is, of course, necessary to set staffing levels and have some idea of what the big deliverables are for a year. But, it's this regular contact directing our priorities week to week that allows us to give the maximum benefit to the business given the resources dedicated in the budgeting process.