In a previous blog on this general topic, I mentioned the importance for the CIO to understand Return on Investment (ROI) and its purpose in the area of financial management. Historically the concept of ROI and its various measures including payback period, internal rate of return (IRR) and net present value (NPV) were designed to allow an organization to rank alternative capital investment opportunities with different cash flows as part of their capital budgeting process. Those with the greatest ROI relative to risk were selected and funded.
As I learned in graduate school, one of the most common mistakes people make in applying ROI methodologies is to focus on the measurements themselves and not the underlying assumptions surrounding the cash flow forecasts that produce the ROI. Here is an example from my own experience.
When I was the CIO for the Department of Human Services (DHS) in Oregon, we initiated a state wide data center consolidation project. Seventeen agencies participated and DHS was the largest. The capital investment was estimated to be approximately $24.0 million to build a new state of the art data center. In order to justify the investment the Department of Administrative Services (DAS) was required to show a two year payback in cost savings, based solely on projected personnel savings (i.e., you don’t need 17 data center managers in a consolidated facility). A two year payback is roughly a 50% IRR.
When the project was completed, DAS and the state legislature considered it a failure because the estimated personnel savings were not achieved. To fill the hole, the state agencies who participated were required to cut their budgets by the amount of the shortfall, which displeased the agency heads and agency leadership in general. It was seen as another IT fiasco. What went wrong?
First and foremost, setting an ROI goal for this project at a two year payback or 50% IRR created the wrong incentive for management. In order to achieve this goal, DAS artificially reduced the number of staff required in the new consolidated data center and accelerated the staff reductions to increase the projected savings and achieve a two year payback. This was a mistake and lead to the perception of a failed project, reinforcing my earlier point that the validity of the cash flow projections is paramount. So where did Oregon go wrong?
Essentially, the cash flow projections didn’t reflect reality and were backed into in order to achieve an arbitrary ROI goal. Here are some examples of items that were overlooked:
1. In the out years, the new consolidated data center had the capacity to provide computing services to all state agencies. This benefit was not estimated or included in the cash flow projections.
2. The project consolidated 17 separate agency data networks generating significant savings not included in the cash flow forecasts.
3. The consolidation of the data centers freed up valuable real estate in agency headquarters that was not valued or included in the cash flow forecast.
4. The construction of the data center was completed on schedule and at a cost $2.0 million below the budgeted $24.0 million.
5. Positions were eliminated (approximately 30+ at DHS) and the people who were in these position were able to fill other jobs that were vacant, leave state government for the private sector or retire.
6. The need to invest in 17 separate facilities going forward was avoided but again this benefit was not estimated or included in the cash flows. The State wouldn’t accept cost avoidance as a benefit but maybe this has changed after Katrina and the Gulf of Mexico oil well blowout.
Bottom line, an IT project that produced substantial benefits to the State over the long run was deemed a failure due to highly suspect ROI analysis and guess who took the heat, the CIO’s. In reality and taking into consideration all the benefits produced, if this project achieved a 4 year payback and only a 25% IRR, that’s a lot better than I’m doing in the market.