One question that I have always asked myself and have had stakeholders ask on many occasions is "when can we release this?". Through the magic of story points and velocity, my teams cobbled together approximate dates with a wide variance on confidence. When the opportunity arose to leverage different methodologies to help the team meet a fixed delivery date, I started to investigate methods to forecast delivery without the traditional point based velocity of scrum. I was looking for something more accurate to give me, my team, and leadership a way to forecast the future with a greater degree of certainty. I leveraged a statistical forecasting method that built a Kernel Density Estimate distribution function from historical data in JIRA. This data is now used to forecast our delivery dates as well as track our delivery progress on a daily basis. My talk will cover the background into the methodology, how to use it and when to use it, along with the ups and downs of getting people on board to this process during its maturation.
