Project managers need to be comfortable with different estimation techniques. Foundational project management courses will teach you about analogous, bottom-up, parametric and three-point estimating. Take a course covering agile delivery and you’ll learn about relative sizing techniques such as estimating poker or t-shirt sizing.
But have you heard of RBS?
I’m not referring to a Risk Breakdown Structure, a Resource Breakdown Structure (seriously, couldn’t they have found a different name for one or the other to avoid creating confusion about which RBS is being referred to?) or any other type of breakdown structure, but rather Randomized Branch Sampling.
This technique was originally proposed by Raymond Jessen in 1955 for the agriculture industry as a method of estimating how much fruit would be found on a given tree (hence the significance of the word “branch”). This approach has been adapted by Dimitar Bakardzhiev for use on software projects following an agile delivery approach and could be similarly adapted for a traditional, deterministic project where a WBS has been created.
I would encourage my blog followers to read Dimitar’s article but here is an overview of the process once you have decomposed your project to an appropriate level of details (e.g. epics & stories or control accounts & work packages).
- Make a note of how many epics or control account have been identified. Let’s call this “M”.
- Randomly pick one of the epics or control accounts.
- Make a note of how many stories or work packages are associated with that epic or control account. Let’s call this “N”.
- Randomly select one of those stories or work packages.
- Estimate the size or effort of that story or work package. Let’s call this “X”.
- Calculate the overall size or effort of the project or release using the formula X/(1/M * 1/N)
- Do steps 2-6 between 7 to a dozen times. This will provide you with a probability distribution for your project’s total size or effort as well as an average or mean which you could use for your estimate.
This approach does require that your project is large enough to have at least a dozen epics or control accounts and does assume that the decomposition is clear and complete.
While you might be comfortable with tried and true estimation methods it’s a good idea to use more than one estimating method. Certain techniques are appropriate at specific points in time over the life of your project based on stakeholder needs and the level of uncertainty remaining. There’s also the benefit of being able to sanity check estimates when multiple approaches get used.
Learning new techniques such as RBS can help us become more effective project managers but only if we understand when and how to use them as well as their limits.
“A Fool with a Tool is still a Fool”