We continue to read headlines about “big data” and the power that it can have on almost any market, but it continues to be an amorphous entity without definition or bounds. What one person may deem as big data, another may simply refer to as a large data set. Definitions are important, and in a previous post I proposed two working definitions. I used the first working definition as the basis for my article last month that focused on questions to ask before or while implementing a big data solution. As a follow up, here are a few questions to ask when taking the first steps toward using your data.
There are many companies or organizations sitting on data that is not being used, is being underused, or is being misused. You can ameliorate the deleterious effects of these conditions, or at least understand how to by asking a few questions. It is wholly plausible that you may not have answers to the questions, or may not know how to use the answers—in which case, you would be well served to partner with an expert to help you get started. These questions are similar to those I recommend asking prior to a big data implementation but do have their own spin.
What do you hope to accomplish? This may seem like a simple starting point, but many people I have spoken with don’t have a business case or even a reason in mind to look at their data. Your goal may be general, “increase sales,” or specific, “increase sales by 50% in the southeast among 18-25 year-old males.” Often the starting point will be more general than it is specific and it will be refined.
What data do you currently have/ collect and what could you collect with little to no additional effort? Take a good hard look at what you have now—and realize that many of the programs you use everyday are actually collecting data in the background. An excellent analyst can help you find sources you may be overlooking, and can help transform some non-traditional data sources into usable data.
What outputs do you need? It is important to understand what you, or the decision makers in your organization need to see to gain value from the data. While it may be possible to find insights and answer your research questions, if it is not presented in a format that is easy for people to understand, it will not be used and will serve no function.
Above all, make sure that you have a plan. Trying to implement any analytical solution without one will result in failure. If you need help, it is important to bring in a consultant or analyst early. Spending money at the start of implementation may end up saving you money in the end.