Application Portfolio Estimation – How to Get the Most Value Out of My Investment
by Frank Vogelezang
Most organizations have an expanding software application portfolio. To keep this up to date, organizations must prioritize software investment. Better estimates enable limited resources to be used in the most efficient way.
Improved estimation maturity helps organizations get the most value out of investments in their portfolio. The basic prioritization mechanism is the WSJF (weighted shortest job first) from the SAFe framework. WSJF means that the investment where [Value]/[Cost] gives the highest result should be done first. This principle can be used at any level, from the portfolio level down to individual sprints. At the epic or portfolio level formal techniques add the most value, where at the sprint level also informal techniques can be used.
With WSJF organizations have a balanced mechanism to determine the value of investment proposals. This is expressed as Cost of Delay. The Cost of Delay is built up from the user-business value, time criticality, risk reduction value and the opportunity enablement value. Most organizations understand the user-business value, but the other three are essential as well to keep the portfolio technologically healthy and ready for future developments.
To come up with a cost estimate at least two, preferably three, different approaches need to be combined. The estimation approaches we see as the most valuable are:
- Sprint or team estimates (people times duration)
- Analogy comparison with previous investments
- Probabilistic estimate, based on historical data
All estimates should lead to a range or a three-point estimate. For organizations with a low estimation maturity that’s the first Learning point. It is a change in thinking going from a fixed number estimate (with almost 100% probability of not hitting that number) to a range that reflects the (un)certainty of the estimate.
When the change in thinking becomes part of the decision making process, the use of probabilistic techniques is quickly accepted, since each probabilistic estimate is accompanied by a certainty percentage. So decision makers can choose to claim critical resources with a high certainty percentage (usually 70% or up) and use a lower percentage for normal investment proposals.
In the presentation I will show real examples (anonymised) of how this process works and what the effects are with increasing estimation maturity.
Read Frank’s presentation