Some people have suggested that when doing an estimate one should make a lower and upper range on the expected time to delivery. The few project tools I have seen, seem to demand one fixed date. Are there any tools that support this concept of a estimation range?
+5
A:
Joel touts Evidence-Based Scheduling in their FogBugz 6.0 software.
Nathan
2008-12-05 17:17:26
A:
I've used Merlin2 which is a project management product for the Mac. When you are starting a new project it asks you the start date and end date - which look fixed, but when you look at the project plan inspector you see that there is actually an "Earliest Date" and "Latest Date" for both the Start and End dates which can be edited. By default it adds the start date into "Earliest Start Date" and the end date to "Latest End Date" - and you can tweak as necessary.
Tim K.
2008-12-05 17:22:55
Normally "Early Start" and "Late Start" (and finishing) dates are used with Critical Paths. That is there is one critical fairly fixed path through your project to work, but there is generally a lot of slack around the other tasks... which determines the early/late start dates.
Mark Nold
2008-12-09 01:34:18
+3
A:
There's also the classic method of providing a best, worst and expected case estimate for each item and then computing a result
computed_result = (b + 4e + w)/6
You can use that to demonstrate how you derived your estimates.
HOWEVER, if you provide a range of time; all the client/sponsor/stakeholder is going to see is the lowest value. No mater what you say. So keep the range secret, and advertise the computed result.
CodeSlave
2008-12-05 18:02:49
The formula is good. The key thing is don't just use a minimum and maximum and average it - if it's normally distributed while you'll be right over all time, it sets you up to fail 50% of the time on your deliveries.
Jon Hopkins
2009-09-28 08:27:49
A:
"Some people have suggested that when doing an estimate one should make a lower and upper range on the expected time to delivery."
But what do your project stakeholders want? Will a range help them decide to fund your project?
Ranges don't really mean very much. Further, most people ignore the range and either see the low or the high number. Optimists have "happy-eyes", see the low number, and complain when you don't hit it, even if you're under the high number. Pessimists see the high number, say it's too big, and demand you replan the project to make the number smaller.
How -- precisely -- will a range help you? Who needs the range? What information will the range help them with? What decision do they have to make that requires a range?
I suggest that you plan each piece realistically.
Further, prioritize your project. After prioritizing, you'll see that there's some essential stuff, some important-but-not-essential stuff, and some optional stuff. This is your range. Cost to do the essential stuff is low. Cost to do important-but-not-essential is in the middle. Cost to do the optional stuff is high.
When someone asks you to "replan", you trim optional stuff.
It isn't a simplistic range. It's a realistic view of what you'll get done and what value it has.
S.Lott
2008-12-08 11:44:33