Agree, it’s the only way for us to test our assumptions.
I wonder how well the experimentation approach will work in this case though, as:
1. Effects might be subtle and outside of the expected impact area.
2. Timescale might be really important (effects over a time much longer than we anticipate. Or even a timescale that makes it impossible to run many experiments in reasonable time)
I wonder how well the experimentation approach will work in this case though, as:
1. Effects might be subtle and outside of the expected impact area.
2. Timescale might be really important (effects over a time much longer than we anticipate. Or even a timescale that makes it impossible to run many experiments in reasonable time)