Developing an appropriate research design — taking into account requirements for temporal and spatial sampling, replication, controls, and counterfactuals — before initiating data collection is critical, for several reasons. These include avoiding wasting resources on unnecessary monitoring and accurately budgeting for required monitoring. Further, research design should inform monitoring decisions (e.g. frequency, sample size and other methodological decisions). Have a qualified scientist and/or statistician develop or review the research design to ensure that it provides the necessary rigor to match the desired level of certainty.