To invent you have to experiment, and if you know in advance that it’s going to work, it’s not an experiment.
Great experiments are those in which the difference between the expected outcome and the actual outcome is significant..
Good experiment move needle. Example
Does user like small title in advertisement or bigger title
Do user want to see price of O365 or only info
Experimental failures are ok if you fail fast.
Operational failures are not ok. These are mistakes. If you have build hundreds of workflow and need to build another workflow and it fails, it is bad execution. If you spend time on building features, insight or aggregate columns that already exist it is bad execution.