--
AndyBerner - 10 Apr 2009
Initial Estimate With Risk
Comment on
Project Selection and Review Process:
When a proposal is being evaluated (sometime during step 2 in Arthur Ryman's description at the link above), the proposal should have access to an estimate to assess the cost/schedule/resources needed along with the variance of those metrics based on the assumptions in the proposal. Schedule variance can be expressed as "probability of completion within time." (See also
Software Estimation Use Cases for suggestion of how risk estimates are provided) For example, for a particular proposal, instead of saying "we estimate this will take 18 months," the time to completion would be expressed as the following set of probabilities, with a wide variance due to the risk involved with some highly innovative content:
10% probability of completion within 8 months
20% probability of completion within 12 months
30% probability of completion within 13 months
40% probability of completion within 16 months
50% probability of completion within 18 months
60% probability of completion within 19 months
70% probability of completion within 20 months
80% probability of completion within 24 months
90% probability of completion within 36 months
Contrast this to another project, also estimated to "take 18 months," that has this distribution, with less risk maybe because there is less innovative content, and there are many precedents to the development techniques needed:
10% probability of completion within 12 months
20% probability of completion within 14 months
30% probability of completion within 16 months
40% probability of completion within 17 months
50% probability of completion within 18 months
60% probability of completion within 19 months
70% probability of completion within 20 months
80% probability of completion within 22 months
90% probability of completion within 24 months
By making the risk explicit, the portfolio manager can weigh the benefits expected (also with variance!) against the risk of completion within an expected time. Presumably the innovative project would have higher payoff. (Ah! But how do you capture short term vs. long term value of innovation!)