Avoid ‘Big Bang’ Delivery

30 years into the software revolution, we've learned a thing or two about what works and what doesn't. 'Big Bang' Delivery falls squarely into that second category. 

This may be entirely obvious to you, having learned the hard way or simply accruing the learnings of others through the adoption of Lean, Agile, Continuous Delivery, CICD or some combination of those. It's worth reminding ourselves why ‘Big Bang’ Delivery doesn’t work.

The failings of 'Big Bang' Delivery can be summed up in four ways. 

• No feedback, no course correction. 

• Expectations rise, rapidly. 

• It's easier to measure small things than big things. Estimations and forecasts go bananas. 

• Concentrated risk - all your eggs are in one big, fragile basket. 

Related to these, there's increased deployment risk, and lack of believability of progress for key stakeholders. 

It's not clear to me what the relative weighting of these factors is, or even whether the relative contribution of these factors are consistent across projects. I am confident that they each contribute to the likelihood of failure. If I had to put money on one factor over another, it would be that feedback is vital. And shielding a project from early and regular feedback is a sure-fire way to deprive it of the information necessary to help it evolve, adapt and grow. 

A final point about 'Big Bang' delivery. 'Big Bang' delivery is an indication that things are going badly. Teams that know better, deliver continuously.

Teams that know better but have run into difficulty (technical, coordination, political) drift towards Big Bang.