An often-quoted paper by Sharpe and Keelin, back in 1998, reported how managers at GlaxoSmithKline redesigned their portfolio selection process. They were unhappy that it had become politicised, as strong-willed project leaders competed for resources for projects that only they fully understood. The process was seen as neither efficient nor objective. As one manager said “Figures can’t lie, but liars can figure”. The improved approach had three phases:

The first was to ask teams to make not one but four proposals: a base-line proposal, to continue the project as planned; a ‘buy-up’ proposal in which they could ask for larger resources for an enhanced project scope or speed; a ‘buy-down’ proposal for smaller scope; and a ‘minimal’ that would close the project but much as possible of what had been learned. This had the effect of moving teams away from ‘all or nothing’ advocacy toward a more business-centric approach.

In the second phase a common set of information was compiled about each project with the help of consultants and colleagues inside and outside the project. Valuations were produced using decision trees and resulting in an upper and a lower valuation for each project rather than a single-point valuation. These valuations were reviewed and debated by the selection panel until everyone was content.

In the third phase the portfolio was selected by an independent internal consultancy group who then presented it to the selection panel for review. The selection panel could now concentrate on the portfolio debate, without getting drawn back into valuation issues.

The process is reported to be very successful. The careful and open valuation process was accepted as fair and the new portfolio projected a 3-fold improvement in return on assets.

Read the full article:

How SmithKline Beecham makes better resource allocation decisions, Paul Sharpe and Tom Keelin, Harvard Business Review 1998

Recommended by and post by Rick Mitchell