Software Estimation – Getting it right

Software estimation is hard – and it’s not likely to change in complexity overnight.  I’ve seen a number of estimation process and tools used across a large scale of projects (from 10K to 2Billion in pounds!). 


I though it might be interesting to write about these and give you an idea of my point of view.  Throughout the literally hundreds of estimations I’ve been involved in (for customers) I’ve got it right most of the time but not always…


Accurate estimation is driven by three factors:


  1. Experience of the person doing the estimate
  2. Historical evidence of the teams performance
  3. The Law of errors


Flame disclaimer: I’m sure an expert on estimation would strongly disagree with a number of my views.  I’m just passing on how I have seen estimation done though-out my career.  It’s hard to have a single rule for this stuff – project size, team drive and experience does count for an awful lot!


So now for a little on each point...



Don’t forget to register for our free technology event ‘Microsoft Connect’ Christchurch 7th April, Wellington 12th April and Auckland 14th April.

Comments (2)

  1. Darrell says:

    Take a look at a few notes I took from a Microosft software estimation webcast (presented by SEs from Construx):

  2. SMcBreen says:


    Great notes! I’ll refer to them in my next post. In the last big org I worked for we were SLIM users and a few of our own tools (Target for example).

    So when I get into the details of building an estimate on historical information I will cover these points off and talk about the pros and cons of Bottom-up and Top-down methods.

    I agree with most of the content in your notes as well 😉



Skip to main content