Jeff replied to one question on Quora regarding Story Point & Hour estimation: Agile Development: Agile Development: how do I estimate story points, and how do I decouple them from hours?
Basically he mentioned 3 research reports, which reported similar results:
- Microsoft (Scrum + Engineering Practices: Experiences of Three Microsoft Teams): agile estimation is astoundingly more accurate than traditional project estimation.
- Latest Standish Group analysis (shown in the figure): agile projects have three times the success rate of traditional projects.
- Forrester Group research (Common Project Management Metrics Doom IT Departments to Failure): The metrics organizations commonly use to determine whether an IT project is a success or a failure—whether the project is completed on time, on budget, and delivered the initial requirements—do more harm than good for IT departments, according to a new report from Forrester Research (eh, you have to pay $499 for this report….).
He also mentioned personal experience which I have the same feelings: “The venture capitalists I work with say they have never seen a correct GANTT chart in a board meeting.” Especially below sentence:
When they dig down into the problem they say none of their management teams knew the velocity of their teams before they implemented Scrum.
Finally he mentioned that the common used practice “Planning Poker” is embedded with the “Wide Band Delphi” technique, which is recommended by Rand Corporation, whose research showed that “humans are not good at estimating hours” and “practical experience repeatedly confirms the research”. (Jeff might be mentioning this one “The Prediction of Social and Technological Events“, but I’m not sure).