The other day I was looking at the analytics discussion board that I moderate on the Information Management site. I had posted a topic entitled “the value of advanced analytics.” I noticed that the number of views on this topic was at least 3 times as many as on other topics that had been posted on the forum. The second post that generated a lot of traffic was a question about a practical guide to predictive analytics.

Clearly, companies are curious and excited about advanced analytics. Advanced analytics utilizes sophisticated techniques to understand patterns and predict outcomes. It includes complex techniques such as statistical modeling, machine learning, linear programming, mathematics, and even natural language processing (on the unstructured side). While many kinds of “advanced analytics” have been around for the last 20+ years (I utilized it extensively in the 80s) and the term may simply be a way to invigorate the business analytics market, the point is that companies are finally starting to realize the value this kind of analysis can provide.

Companies want to better understand the value this technology brings and how to get started. And, while the number of users interested in advanced analytics continues to increase, the reality is that there will likely be a skills shortage in this area. Why? Because advanced analytics isn’t the same beast as what I refer to as, “slicing and dicing” data to produce reports that might include information such as sales per region, revenue per customer, etc.

So what skills are needed for the business user to face the advanced analytics challenge? It’s a tough question. There is a certain thought process that goes into advanced analytics. Here are five (there are no doubt, more) skills I would say at a minimum, you should have:

1. It’s about the data. So, thoroughly understand your data. A business user needs to understand all aspects of his or her data. This includes answers to questions such as, “What is a customer?” “What does it mean if a data field is blank?” “Is there seasonality in my time series data?” It also means understanding what kind of derived variables (e.g. a ratio) you might be interested in and how you want to calculate them.
2. Garbage in, Garbage out. Appreciate data quality issues. A business user analyzing data cannot simply assume that the data (from whatever source) is absolutely fine. It might be the case, but you still need to check. Part of this ties to understanding your data, but it also means first looking at the data and asking if it make sense. And, what do you do with data that doesn’t make sense?
3. Know what questions to ask. I remember a time in graduate school when, excited by having my data and trying to analyze it, a wise professor told me not to simply throw statistical models at the data because you can. First, know what questions you are trying to answer from the data. Ask yourself if you have the right data to answer the questions. Look at the data to see what it is telling you. Then start to consider the models. Knowing what questions to ask will require business acumen.
4. Don’t skip the training step. Know how to use tools and what the tools can do for you. Again, it is simple to throw data at a model, especially if the software system suggests a certain model. However, it is important to understand what the models are good for. When does it make sense to use a decision tree? What about survival analysis? Certain tools will take your data and suggest a model. My concern is that if you don’t know what the model means, it makes it more difficult to defend your output. That is why vendors suggest training.
5. Be able to defend your output. At the end of the day, you’re the one who needs to present your analysis to your company. Make sure you know enough to defend it. Turn the analysis upside down, ask questions of it, and make sure you can articulate the output

I could go on and on but I’ll stop here. Advanced analytics tools are simply that—tools. And they will be only as good as the person utilizing them. This will require understanding the tools as well as how to think and strategize around the analysis. So my message? Utilized properly these tools can be great. Utilized incorrectly—well—it’s analogous to a do-it-yourself electrician who burns down the house.