Businesses need to judge how best to spend their money so they need to know costs and timelines. Only when software is finished and deployed does it start to save costs or generate revenue so, in general terms, the longer it takes to create the poorer the investment.
Forecasting the cost of software, though essential, is not straightforward. Building software requires lots of unpredictable problem solving and a large part of the agile body of knowledge describes tools and techniques to help with this.
Ian Carroll walks through some of these issues and pragmatic ways to tackle them.