Building a software product requires thousands of design decisions to be made and making those decisions quickly will significantly reduce the time to market. Getting things shipped quickly can significantly reduce commercial risks through gaining early feedback. But making these design decisions effectively requires data on all kinds of things from commercial issues like user behaviour and early sales figures as well as technical things like run time performance, storage usage and quality metrics. Some of this data maybe easy to collect or intuitively felt while other aspects require determined engineering effort to harvest.
Finding ways to collect and use this kind of data can be transformative by turning assumptions into assertions sooner and the best teams strive to improve this constantly. Jon Fulton talks here about data collection, F1 teams and risk management at a recent visit to Agile Yorkshire.