What tools do you use to estimate Level Of Effort (LOE)?
What tools or methods are you using to estimate the level of effort on your projects? I've read a lot of books on the subject, but am curious what people are actually using "in the real world". Please also specify if you're talking about Commercial Off The Shelf (COTS) product development or consulting services, since I'm curious if the approach is different.
4 Answers
It really doesn't matter what is most common or widely used. The real question is what will work for you. This is HIGHLY correlated to who is on your team and what their experience has been.
EBS is nice, but it only works with a history of "guesses" and then fit against actual results.
You can't really dictate to people how to do this.
This is essentially one of the biggest, most difficult and unsolved problems (for the general case) in our field. Estimates are notoriously bad or unuseful.
As long as you are aware that it is a range with a probability, then you are better off than most people.
I have found that accurate estimates are easier when you are very familiar with the domain.
-
+1 for the range comment. I really like have the best case / worst case effort for a project. It helps give it a sanity check. – Jarie Bolander 15 years ago
From Joel on Software, Evidence Based Scheduling (EBS) - Definitely worth considering... Even if you don't adopt EBS, there's lots of goodness in this article.
EBS may seem like a lot of work (relative to just going with your gut or using some multiplier to give you a safety buffer), but it mitigates a lot of really tough things to account for (i.e. poor estimation skills, individual productivity factors, etc.).
-
EBS might be magical, however, how do those people come up with the estimate the Fogbugz then manipulates? All that does is take a guess, apply some probability score based on previous guesses and work, and comes up with a set of new numbers. I think the OP wants to know how to come up with the original number - and even more specifically, does not have a history of prior information. – Tim J 15 years ago
-
In my experience, most organizations never get beyond the practice of simply living with the poor original LOE estimates. Obviously, without prior information (i.e. experience with similar projects/tasks or the specific team members), solid estimation is pretty darn challenging (not impossible, just fraught with unknowns). I'll post some other ideas on that soon (in the hope of better addressing the OP's question). Torrie, please let us know what aspect(s) you were interested in. I just want to make sure we actually add value here. – Chris Hagner 15 years ago
-
I agree - most places don't get beyond that. It is perhaps one of the hardest things in our field of work. this has not changed for decades and it will always be that way IMO. Note also that this kind of discussion is typically not one that startups have. – Tim J 15 years ago
-
Thanks, Chris, FogBugz EBS is one of the things I've looked into, but haven't used myself yet. Are you actively using it for projects? Do you do COTS software or are you providing services to several clients with several different schedules? Mostly I'm interested in what people are actually using, or whether "a gut feeling & multiply by 3" is in fact the most common estimation method. – Torrie 15 years ago
For any project I do, the method is exactly the same. Simulate the project in OmniPlan or MS Project and track it via Key Process Indicators (KPIs). Driving my analysis is that I figure out Effort of each task not the Duration that each task would take. This is critical since the Duration of each task will vary depending on the Resources applied to it.
Once the project has been scoped out, then tracking it is done via the KPIs.
-
Thanks, Jarie. My main curiosity is, how do you "figure out Effort of each task"? – Torrie 15 years ago
-
Torrie- That can be a bit tricky. What I normally do is ask the people that will be doing the work. They usually have a good idea as to what needs to get done. If you don't have anyone, then I would suggest doing some research into how long other projects similar or ask someone you trust. Even if you are wrong, at least you have the tasks and a rough idea of effort. You can't change a plan until you have one. – Jarie Bolander 15 years ago
Defining units of work and their complexity is a bit challenging if it doesn't fit into a wider framework, their business value and inter-relation to other units.
Similar in scope to EBS (scanned it, good stuff), I use crystal clear's blitz planning which can be summarized as:
everyone writes down all the tasksArmed with this information, I can create feature lists, a product backlog list and ultimately a burn list. All very agile terms - but best described in this earned value and burn charts post. There's an excellent book on this subject of small team lightweight agile methodology - Crystal Clear, by Alistar Cockburn. Rather thank write about how great agile is, the book goes into detail about how to implement it.
they know of onto index cards, throws
them onto a long table, from which
point they sort, add estimates and
notes, look for the Walking Skeleton
and the First Delivery, strategize and
restrategize about roadblocks, costs,
time, resources, moving the cards
around as they go. Older folks will
recognize this as an annotated Pert
chart, constructed collaboratively
with cards.
