Please login or join to subscribe to this thread
Totally depends on the approach taken to deliver the project (e.g. adaptive vs predictive) and the degree of complexity and uncertainty about the scope and how to achieve that scope.
With a low complexity, simple project following a predictive approach, it should be possible to get rough estimates for the different stages of delivery as a percentage of an overall ROM estimate.
Relative estimation (i.e. story points) using Delphi or some other method of the team's choosing. The team should create and own the plan so it's ultimately up to them what inputs and estimation methods they use but most dev teams I come across use relative estimation techniques.
Kiron made a valid point.
If you don't get some kind of high level estimates at the outset, you have no assurance there is even a valid business case.
Sometimes it is similar enough to another project that we can do a parametric or analogous estimate of the whole project at the outset. That's a good sanity check. If it is unique enough, we will compile estimates from the various teams involved. They may have multiple ways of doing that, so I do provide a template to make sure all the inputs consider key aspects like project phases included in the scope.
Providing a template also helps "normalize" the inputs so they can be compiled. I don't have a standard template though. I will often create them specific to the project based on the nature of the work, and input from business management and program leaders on what information is needed to make steering decisions prior to investing too much money.
On most projects, I have used a two-level approach: ROM, then LOE.
The rough order of magnitude is usually meant to determine whether or not to proceed with the project or initiative. The number is less important than the magnitude: is it 1 day, 10 days, 100 days, ...? Based on a relatively small set of information, the ROM usually has a large variance (e.g. ± 50%).
Once the decision is made to proceed, and additional information is collected, you can then come up with a level of effort. The LOE is used for planning (e.g. budget and schedule) and should change as additional information is collected. The LOE variance is usually smaller (e.g. ± 10%).
Please login or join to reply