Hi George,

You bring up a great question!

I think that the "relative" property of Story Points is exactly what you need and you don't want to change that. Story points (similar to Function Points & Use Case points) have been introduced so that projects can be measured and quantified much sooner in the development process, based on the requirements and desired features, and do it in a way that can be compared across projects, etc.
The idea is that you can compare the relative complexity of requirements between projects - which is probably all you need at the high-level.

If you want to go to a deeper level of granularity you would need to employ a formula for converting Points (story, function, use case, etc.) into effort rather than changing the relative nature of points. Unfortunately, that is not easy since given the same exact requirement (sorry - I mean story) the actual effort could vary depending on a number of parameters such as:

• type of technology employed,
• type of programming language,
• technical team skills set,
• familiarity with the subject matter (on the BA side),
• etc.

What is it exactly that you are trying to accomplish when comparing projects across departments?

If you're using story points, you will still need some level of governance and/or standards to ensure that story points are counted consistently across teams/departments.

Just some thoughts!

- Adrian

George -definitely take a look at Mike Cohn's book 'Agile planning and estimating.'  This scenario is explicity addressed.

What I am about to do this next fortnight is use story points across two different projects and get a compare. I am not following Cohn's advice exactly but am doing the following; Creating two sets of SP estimates independent of each other - obe is from an existing 'in train' project and one is new. The new one will only have  high level storis - hich means that th scale won't be comparable. 

(One will have a bunch of 2, 5 and 13 point stories, the other if using the same baseline will be full of 100point stories.)

We'll then be picking a sample from one of the new stories and and breaking it down so we can do a detailed compare of the SP at a detailed level.  Then we'll calibrating the new to the old.

This sounds complex, but that may just be my explanation.

The real issue is that team velocity will vary based on a number of factors, some of which Adrian mentioned.  So Story points alone don't give an end date.  You really have to invest a sprint two (or 3) before you start to get  picture of the duration.