[phelmig]

I have great experience with agreeing on the features to be delivered in a two-week sprint. You provide and estimate in hours and have a hourly rate for your work. This allows you a) to outperform your estimate and gives b) your customer some planning security. If you screw up, a the content of a two week sprint won't brake your neck. If you outperform you have more time for your project.

Additionally you get rid of the discussion if dev a is worth 1.2456 more than dev b.

Infrastructure is cost you bill additionally.