Type: Process
Status: final
Version: 2007-07-27
Gist: show the participants of the experiment that they are not very good at making estimates and not very good at knowing what the requirements for a solution are
Source: Niels Malotaux, Gilb Seminar on Decision Making, London, UK, 2007
S1: Explain the 'project': "multiply 2 numbers of 4 digits each, on paper". Every participant will conduct the project on his or her own.
S2: Ask for everybody's estimate, how long the project will take to be finished.
S3: Everybody shall measure the time needed, from reception of the requirements to delivery of the result.
S4: Give the requirements: "The numbers are: 4-5-6-7 and 9-8-7-6".
S5: Let them re-estimate.
S6: Interrupt them now and then by talking, because this happens to every project.
S7: Yot down the times needed.
S8: Ask if they have testet the result. Make them test their results and take the time again.
S9: Add the times and compare them to the original and second estimates.
S10: Show them a solution to very different requirements. Make something up, like changing the order of the digits.
Note: There are a couple of interesting questions to ask here.
Like "What does 'finished' mean?",
"What was the reason for the adjustment of the original estimate?",
"What was the impact of the interrupts?",
"Is it rational to forget testing, and why does it happen all the time?",
"Why were we assuming that we really understood the requirements?"
No comments:
Post a Comment