Thanks to Predictive Analytics, data is now more than just a log of the past. We are now focusing on the future, analyzing past data to uncover important patterns. At this point in time, we are now capable, with the help of analytics, to combine the power of prediction with our current grasp of present events.
One common example occurs in the world of telecommunications, an industry that covers millions of subscribers. A certain telecom company makes use of a predictive model that handles data on how users are accessing their network—if they are using certain apps, making a call or accessing particular websites.
With this information, the company can then use prediction and event processing to push appropriate offers at the best possible time. The model they’re using will know if a customer tends to make international calls and might be better suited with an upgraded plan. Instead of sending an offer when the customer is likely to be busy, the telecom company knows to send an offer right after he or she ends an international call, making it much more timely and relevant.
Pushing an offer at the right place and the right time is a very powerful action. This is very evident in cases with retail establishments. For example, a customer who was browsing a retailer’s site earlier in the day or is logged into an account that has his or her mobile number stored walks past a branch of the said retail store. The customer then receives an immediate offer for a product they were looking at earlier, telling him or her that the store has it in stock and that a 10 percent discount is waiting if they decide to purchase now. The ability to interact in real-time with customers is a powerful weapon in retail.
Companies used to require long stretches of time just to overhaul generators because they would rarely know when maintenance is needed most. Now, with the help of data from sensors monitoring fuel use, temperature and cylinder pressure, it is easier to spot when parts are nearing their point of failure. With real-time analytics, the predictive model is able to notice a significant drop in pressure and, by comparing it with past records, is also able to predict that the current rate of use will lead to equipment failure within a set period of time.