Forecasting agile software project delivery dates without estimating

Even in the Agile world, management teams like to have an idea of the overall cost and timeframe for a project.  Justifiably so.  There is no point embarking on a project if the ROI is poor or if the actual delivery date is means the window of opportunity will have closed.

Throughout my career, I have intermittently experimented with Monte Carlo based estimation.  Each time I have discarded it in favour of the approach I blogged about a few years ago.  You can find that here.

The main reason for this is that, having run the Monte Carlo simulations, the generated dates were wildly pessimistic, and indeed the projects proved them to be so.  The likely reason for this is the accuracy of the available data that was input into the simulation.

Recently, a colleague pointed me to another resource for forecasting dates for Agile projects using Monte Carlo simulation.  It can be found at

I proceeded to run this against some some past projects and have found the forecast dates generated were remarkably accurate.

I will definitely be making use of the this in the future.

After downloading the spreadsheet, all you need is:

  • a clear understanding of project (or feature) scope AND
  • (preferably) a short history of the velocity of the team(s) conducting the work.

The real advantage with this spreadsheet, is that you do not need to have detailed technical understanding of the work that will be undertaken, just a good understanding of the scope.

This means the agile team has technical freedom, and there is no need to wait for technical resource to analyse the project and feed back the dates.

Additionally, you can update it will actual velocity and incomplete items and it will update automatically, making forecasting and delivery updates significantly easier, as well as telegraphing early when a project is 'slipping'.

Why don't you try it for yourself, and let me know how you get on in the comments.