We all hope customers know exactly what they want when starting a project, but sometimes their needs change as what we do helps them achieve more clarity. Our older approaches tried to restrict the customer from changing while Lean-Agile approaches can allow changes even on fixed-price work to the degree the customer is willing to swap out an equal amount of now-lower-priority work in its place.
In working with large, complex projects that inevitably change, we have found that the Deming Cycle works well because it works on an empirical basis, meaning practitioners verify by observation or experience rather than by theory.
The Deming Cycle steps include:
Agile methods are empirical, or evidence-based, using check and adapt cycles, or what Scrum calls inspect and adapt cycles, to adjust team actions and plans based on principles to the emerging situation in your specific context. Every iteration has a reality check, providing more reliable predictions of milestone deliveries.
After using PDCA to make an improvement, standardize the new innovation that changed your existing practices so that you retain the gains. Then use PDCA again for the next gain to continually improve.
Let data inform your decision making. There are uncertainties when trying something new. Applying Agile the first time is new, and all eyes may be on you and your team.
Set expectations so people expect to use experimentation at the beginning. Use the Deming Cycle. The PDCA cycle allows you to go forward without knowing everything up front. Socialize your plan. Apply the Lean-Agile approach. Check your data. Then adjust to overcome challenges and to capitalize on opportunities.
The 'P' can also mean predict. Experiments are important during innovation to see what happens. Our testable prediction, or hypothesis, is that Lean-Agile will improve cycle time for learning experience development.
Focus on making improvements in how well you implement the principles. For example, does your focus on flow, rather than focusing solely on waste elimination, prove a better catalyst for continuous improvement for learning experience development?