Cesar Brea and I had breakfast a couple of weeks ago with our SAP friends Peg Culotta and Russ Lefevre. Their folks have been thinking about how investments get prioritized in government, and particularly about how politics and what economists call 'externalities' get factored into the process. Under a concept they've labeled 'Public ROI', they've been trying to understand how public-sector leaders could make a better case for technology investments.

With our research at Harvard focused heavily on the 'politics of cross-boundary initiatives,' SAP's ideas seemed particularly interesting.

Working these ideas over a few eggs, Cesar suggested a simple '3x5' card calculus to fuel the debate. The goal is a framework to better evaluate different opportunities against each other (snapshot view), and/or to better track and manage an initiative over time.

On Cesar's card, 'Public ROI' = R x S x P, where

R = Rational ROI -- more or less the traditional business case based on direct, measurable benefits and costs . In most settings, this is the analysis that justifies the budget you need. If an agency wants to do something badly enough to self-fund it, this analysis may stay internal. If new money is required, however, this is typically sent forward as a proposal for budget negotiation.

S = Social ROI -- more or less the indirect and difficult-to-measure 'public good' benefits and costs. These may be benefits and costs to citizens or to institutions in the value chain other than the institution proposing the project. They often include uncertain costs such as the reallocation of already budgeted staff (which usually dwarfs the new money budget request) and/or benefits captured not as cost reductions, but rather in the form of improved service or distributional equity (e.g., government's role in providing a safety net for have-nots). In Cesar's model, Social ROI would be expressed as a 'multiplier' factor. In a world of serious external effects, Social ROI might range from .5 - 2, with the precise number based on a descriptive listing of effects that are then subjectively quantified. While simple and rough, including Social ROI would be better than excluding it, as is often the de facto situation today.

P = Political ROI -- more or less the motivational feasibility of the project, or the benefits and costs for interested parties (champions, opponents, decision-makers). This could be expressed as another adjustment factor, with values ranging from 0.5-1.5. It would be based on a subjective evaluation of elements such as the confusion and conflict to be overcome during implementation, and the timing, sponsorship, and concentration/diffusion of benefits and costs among relevant parties. For example, concentrated benefits with diffused costs would increase motivational feasibility, while concentrated costs and diffused benefits would do the opposite. Timing is another element, with the Political ROI higher for a project that can show results by the next election and lower once lame duck territory has been reached.

* * *

We'd be interested in what you think about this Public ROI framework and its 'multipliers' -- do they work for you? Have you seen or used a better approach for prioritizing investments? If Public ROI, which is intended to improve analysis, itself needs improving, what does it need? Who has the best methodology out there for analyzing and prioritizing IT-related investments?

What, in any case, is YOUR methodology, and how well does it fit for 'cross-boundary' investments?