Check the data first
An event may have fired twice, some traffic may not have been tagged, or tracking settings may have changed after a website update. Before drawing conclusions about visitor behaviour, make sure you are comparing equivalent actions and comparable periods.
Separate the signal from the explanation
A drop in enquiries reveals a change, but not its cause. Traffic sources, seasonality, the offer and product availability all affect results. Develop and test several explanations before changing the design.
Connect metrics to the objective
Different websites depend on different actions: requesting a quote, viewing contact details or placing an order. Choose measures that help assess the customer journey. An attractive report is only useful when it makes the next experiment clear.