Your data warehouse stores critical data telling you what is happening in your business and sometimes why it’s happening. But you can go beyond understanding why something went wrong. You can use past data to predict the future, correcting problems before they happen. In a recent survey that Oracle did of over 300 C level executives, 93% of them thought that their companies were losing an average of 14% of their total revenue because they couldn’t fully leverage the information they had already collected. One key way to do this (and you’ll hear more about this in a future survey) is to use predictive analytics. Let’s take a quick look at why and how.
Turkcell is a leading mobile phone provider in Turkey, with over 34 million subscribers. And like most mobile providers a majority of those subscribers use pre-paid accounts and pre-paid cards. Money launderers take advantage of this, and losses for this business are of the order of $5 for every $10,000. This may not seem like much, but with billions of transactions, this adds up to millions of dollars a year.
Like other companies, Turkcell examine huge quantities of data and build models that help it identify and ultimately predict and prevent fraudulent transactions. Unlike many other companies, Turkcell does this analysis in its data warehouse. With 100 TB of compressed data – representing over a petabyte uncompressed – it would take a long time to move that data out of the warehouse and keep it up to date as new data arrived. And the window to stop the next fraudulent transaction might have already closed.
Oracle Advanced Analytics enables you to perform sophisticated predictive analytics inside a data warehouse. You can mine your data directly while it is inside the Oracle Database using either SQL or R language APIs or the Oracle Data Miner SQL Developer “work flow” GUI extension, depending on your need and existing skills. You build models for past behavior and use that to predict future behavior, improving your accuracy with time. And best of all, there’s no need to move the data around which takes time you might not have and also leaves you exposed to security risks. As Turkcell said “...we can analyze large volumes of customer data and call-data records easier and faster than with any other tool”