Guest post from Brad Hill, Market Manager, IBM Business Analytics
Follow Brad on Twitter: @bradhill14
Sir Howard Stringer, Chairman, CEO and President, Sony Corporation, once said, “All great companies have a rowboat mentality, they are rowing hard, but they are looking at the past.”
Reflecting on what has happened in the past is an important and a worthwhile exercise that businesses do all the time.
Many have business intelligence systems in place to consolidate their data and report on the current state of the business. Looking back at historical data helps identify business issues or areas which may need improvement, however because the process is user driven, it relies on experience or gut feel to identify the problems.
So how do we turn around and face the front to see what’s going to happen next? What if we don’t like the outcome? Is there anything that can be done to influence or change that outcome?
Enter predictive analytics.
Predictive analytics uses historical data and applied analytics to draw conclusions about current conditions or future events using techniques such as data mining and statistics to uncover relationships and patterns in the data. By finding these patterns, and understanding why and when an event is likely to occur, predictions can be made about future events, actions or behaviours.
For example, some law enforcement agencies, such as the Memphis Police Department, have analysed crime data to uncover patterns and figure out where to send patrols to prevent crime, allowing them to cut serious crime by 30 percent. Companies in the communications industry have been deploying predictive analytics to reduce customer turnover (churn), by identifying customers at risk of cancelling their service and using retention strategies to prevent them from leaving.
Predictive analytics can be applied to almost any organisation and can be used to solve one specific issue, or to address a number of problems across departments. Most commonly it is used in the area of customer analytics, to acquire new customers, grow the value of existing customers and to retain high value or good customers. And, many organizations are also using it inside of their operations to better plan supply and demand of parts, manage the day to day operations or processes and maximise the performance of an asset or employee.
If you had predictive capabilities, what would you predict?
No, don’t say lotto or stock prices (although some people are exploring the latter). Think about your current employer, what are its main business challenges?
For example, if there was a need to increase the average purchase made by a consumer in a retail store, wouldn’t it be beneficial if we knew what else that person was likely to buy? Or maybe if the crane on a truck broke or failed, could there have been a way to service the truck before it broke?
Identifying these key business problems is the first step on a predictive journey, but the most important part is putting it into action. For instance, knowing what to recommend to a shopper is of no value unless the store actually suggests it when the customer is on site. Or if we didn’t service that crane before it broke, or make arrangements for an additional one to replace it while it was going to be serviced, it would still have the same outcome of unscheduled down time.
Understanding the key drivers is often as valuable as the prediction itself. Knowing that a student is likely to drop out of his course and discontinue study is good first step, however knowing that the reason they are likely to do so because he/she is in their first year and is struggling with a key subject due to language barriers and the complexity of the content, is much more actionable. This allows for a more appropriate intervention that can prevent him/her from dropping out.
It is the key drivers that are usually under some degree of control, in this last example changing entry conditions or introducing a bridging course for foreign students, can shift the result in our favour and change the end result, of a decrease in student attrition.
So how do you shift this rowboat mentality?
Rather than companies employing a coxswain (the guy at the end of a boat that steers and instructs the rowers on what to do) consider adding analytics to help guide the boat. Using what has previously happened to understand why outcomes took place and how outcomes can be changed by letting the data tell the story.
For more information:
· Read the whitepaper, “Seven Reasons You Need Predictive Analytics Today”