The Internet projects have a particular complexity, everything is changing constantly. In this kind of environment is very risky to base your decisions for new projects in previous ones since the possibility that the same formula works in this new very unique situation is almost null.
Due to the above mentioned situation is normal that managers and analysts have to make decisions based on very unique and particular scenarios, that is why I highly recommend going beyond the conventional Analytics activities focusing on comprehend the human mind and understand the analyzed system:
1- Every new problem must be taken as that…a new problem. Never try to compare it to some previous one.
2- Defining the solution or objective is a team work. This way you will eliminate your subjectivity and promote having plan clear and understandable for everyone.
3- There is one main object that prevent increasing your system (project) throughput. So, keep focused on optimize its efficiency.
4- One measure and action cycle at a time. If we do everything together we going to loose the possibility of understand the cause and effect relation among variables.
5- Keep focus. The information must be of use for making decisions and controlling, do not spend your time analyzing “interesting” information. Your time worth and the best way to invest it is making your project (company) generating profit.