I believe the most effective method for resolving an experiment failure is more testing. This gives you the opportunity to try out different parameters and see their interactions. It will become clear down the line which ones have the highest level of success. Also, root cause analysis is very important. Finding out exactly where the failure happened will help create a more robust solution. In our company, we use a lot of software tools. Often, the programs written will have bugs that we have to find. Seeing the results of the program and working backwards helps us find where the origin of the problem lies. This in turn, will allow us to change the code and make a better one for future use.