Forecasting via Monte Carlo Simulation
M
Manuela Barth
We as an agile team want to make data-based plans. For this we want a forecast that shows us what amount of work we can commit with which confidence.
As product owner we want a forecast for our customer to give an data-based estimation when an item will be finished.
Both things can be done with monte-carlo simulations. At the moment we use the jira flow companion & jira to do this.
Log In
S
Shobha Srivas
This feature for forecasting is a critical need for us as well. So far, we are using tools like MS Project to help with forecasting but it is cumbersome
Skyler Reeves
This is the main thing we need as well. I’m considering just building my own ETL pipeline to produce these simulations in Google Sheets as a work around.
Jenny Lucas
Additional info from another client:
I've been playing around with Dashboards this morning - I can't find a way to product the core metric that we need - namely a statistical forecast / burnup, wherein a delivery date is forecast based on the number of tasks in a given scope multiplied by the avg. lead time that a team achieved over a given period of time.
So if I have 10 tasks, and an avg. lead time of 10 days, then my forecast stretches to 100 days from today.
The forecast should extend or contract based on both variability in the scope (i.e. if items get added or removed), and on the lead time achieved.
Ideally, you would also be able to choose between a lead time calc that is based on a delimited scope (so an epic or some sub-set of all the tasks in a given list), and the percentile lead time that I'm looking for (i.e. the lead time for 50th percentile will be lower than for 90th percentile, and I get to choose the level of confidence I want in my forecast).
For the time being, I'd take something more simple - just the ability to multiple lead time X the number of tasks in a scope, and get a likely # of days or hours that the scope will likely take to deliver.