You start off with a need to do statistics, which may seem like a weird statement but then again you’re reading this book.... Perhaps your need arises from:

Often, your requirements help you understand the type of data you require, the statistics you need, and the like. All too often, businesspeople or readers sabotage their own research because they do not adequately understand what they wanted to achieve in the first place, causing them to gather the wrong data or do the wrong analyses. Therefore, try to have a really good idea of your needs prior to going any further. Identifying the right problem is a major up-front challenge.

Note that I am not saying your intended outcomes from a statistics project must be exactly defined. It is fine to say that you just want to explore your company’s data for useful patterns, or the like. However, knowing what you do and do not want to achieve will invariably help guide your efforts.

In addition, always think deeply about the context of the data situation. By context, I mean the environment of the situation (such as the culture of the people involved), the physical environment, and the like. Consider how the context affects the data involved, and whether this context gives you extra information about what or who to study.

For instance, say you are studying financial investment portfolios. Your study and expectations should be heavily affected by the broader financial or economic environment at the time. If there is a major financial and economic crash in the middle of the time period of your study, your entire thinking about the problem and possibly your data choices may change. In the case of Accu-Phi, you have already made some choices about what might affect sales – are there other aspects of the sales process and environment issues you have omitted?