Depending on where you look, you may see different posts stating that there are three, four, five or more types of analytics. So how many are there? There are three types of business analytics that are consistent among different sources: Descriptive, Predictive and Prescriptive.
Answers the question: “What has already happened?”
Descriptive analytics is the most basic form of analytics and usually the first type of analytics utilized by businesses. Descriptive analytics attempt to describe what has happened in the past. It does this by looking at historical data presented in a manner that is easily digestible such as reports, charts, dashboards, scorecards, etc. Most of the data that you see today are descriptive analytics.
- income statements
- balance sheets
- most information on your car’s dashboard: mpg, odometer, average speed
- school report cards
- fitness tracker applications
Depending on how this data is presented you can begin to identify trends, comparisons and changes from period to period. Descriptive analytics do not tell you anything about future performance or the likelihood an event will happen.
Answers the question: “What might happen?”
Predictive analytics can use most of the same data that make up descriptive analytics and attempt to predict probable futures based upon historical performance. You will notice that predictive analytics uses words like “might happen” or “probable” and not “will happen” and “certain”.
Everything is possible and Nothing is certain
The above quote demonstrates the point of predictive analytics – the fact that we cannot look into a crystal ball and determine the inevitable future, we need a way to determine on a scale from “Will Not” to “Will” how likely something will happen – this is done with probabilities. Since nobody knows what will happen, probabilities exist to protect you against your guesses about the future.
- business earnings forecasts
- other information on your car’s dashboard: remaining mileage based on fuel
- exit polls to predict election winners
Knowing that an event is likely or unlikely to happen – what can you do to take advantage or hedge against the event?
Answers the question: “What can I do about it?”
All organizations utilize at least one outcome of prescriptive analytics (wether they choose to or not) – that is, deciding to do nothing about a probable event. Inaction is just as much as a choice as performing a specific action. Prescriptive analytics aims to provide the action steps necessary to achieve results of predictive analytics and how each action will impacts everything else.
A simple way of viewing the phases of business analytics is to compare it to healthcare.
Before deciding to visit a doctor, there are typically symptoms experienced by an individual that can be readily described (e.g. headache, fatigue, pain, etc.) – in other words this is raw data. At this point a doctor can interpret symptoms and/or run additional tests to ultimately make a diagnosis, he is only telling you what disease or ailment that you have. This is Descriptive Analytics.
The next step a doctor can do is to analyze the situation further. From the data obtained from the additional tests or the severity of the symptoms a doctor can make a prognosis – that is a forecast of the most likely outcome of the diagnosed disease or ailment. This is Predictive Analytics.
Lastly, a doctor may be able to explain the possible treatments of the disease or ailment. The treatments may have varying success probabilities but may also have possible side effects. Choosing a treatment will have defined benefits such as treating symptoms – but can also introduce additional factors not previously seen before. This is Prescriptive Analytics.
The previous example is pretty rudimentary, but it demonstrates an important point. The reason why the patient was able to identify and take action on a perceived problem (symptoms) was by the seeking help of a medical professional who is trained in diagnosing and treating diseases. The patient in this case can be easily compared to any business professional. Analytics professionals can assist businesses identify and take actions on perceived problems, but the outcome relies heavily on the available data.