Predictive analytics is a hot topic. It has certainly created quite a buzz recently. In fact, a quick look at Google Trends reveals that there has been an over 300% spike in worldwide searches for predictive analytics since 2010. Moreover, Gartner’s Emerging Technologies Hype Cycle shows predictive analytics as the most proven emerging technology in 2012.
Today predictive analytics has emerged in multiple industry segments including, banking, healthcare, social media, retail, meteorology, military, and much more. Social media companies study a user’s online behavior to predict which advertisements will bring in higher conversion rates. Financial institutions run client records to predict the likelihood of a loan being paid back on time, and insurance companies use predictive analytics models to estimate upcoming claims.
We are certainly living at an interesting time. Charles Duhigg, illustrated this in a recent article for the New York Times . He wrote about how big box retailer, Target, uses predictive analytics technology to find the correlation between shopper behavior, purchase history, and the possibility that this shopper is pregnant. In fact, depending on the purchase history they have even broken it down by trimesters. The article brings up a very interesting situation where Target sent flyers to a possibly pregnant shopper promoting products for early pregnancy addressed to the shopper’s name. The problem was that she was a 16 year old girl, and her parents had no idea she was pregnant.
On the other hand, there are industries where predictive analytics has not been ironed out yet. A great example of that would be the stock market, one of the ultimate potential uses for predictive analytics. Imagine the ability to predict which stocks or futures will increase or decrease at any given time. However, while it may be possible to forecast several elementary factors that would affect a certain stock, it is extremely challenging to take into account important influences such as geo-political changes, legal and legislative decisions, world events, exchange rates, social trends, other stocks behaviors, and much more. Still, we have seen how many stock market decisions are made by computers analyzing data from news feeds, RSS, and Twitter. In April 2013, the Twitter account of Associated Press was hacked, and a tweet was sent out saying the White House had been hit by two explosions and that President Barack Obama was injured . This caused a 143 point fall in Dow Jones within minutes of the tweet, which immediately rebounded just 6 minutes later, when AP confirmed that its account has been hacked. This immediate reaction to the news on twitter was mostly as a result of an analytics system that scans for key words such as “Whitehouse”, and “explosions” and triggers preset actions with the stock. While this proved impressive, keep in mind that these systems are still reactionary, and not predictive.
So what does it take for predictive analytics system to be accurate, and efficient?
And which industries can benefit the most out of this technology?
The answer lays in Business Specific Predictive Analytics. The only way that predictive analytics could produce accurate results is if it is built specifically to your business. This means taking every single influencing factor of this business into account, and how they affect each other.
When it comes to Retail, whether it is fashion, automotive, jewelry, electronics, or liquor and wine, the supply chain process is fairly similar. It normally involves planning, purchasing, initial allocation, replenishment, assortment and merchandising decisions, price optimization, handling promotional events and other steps. The influencing factors therefore are also common: historic sales and consumer demand, trends, seasonal fluctuations, price elasticity of demand, lead times, and so on. Yes, each retailer will have a unique process, with its own nuances and complexities; yet the major part of the influencing factors is measurable and predictable. A good retail predictive analytics engine is built on a common analytic platform that is capable of accounting for all these factors, their inter-relations, as well as the business specific nuances.
Some of the most effective predictive analytics tools that traditional retailers already use are: Advanced Demand Forecasting; Predictive Price Optimization; and Inter-Store Inventory Balancing. These tools enable a retailer to have the right product, at the right place and the right time, proactively, before a customer walks in the door. Sounds like a fantasy? It’s not. Brick-and-mortar retailers that use predictive analytics are experiencing 25 to 40 percent in inventory cost reductions and a boost in turnover, in some cases, by up to 3.5 times. Furthermore, these retailers report an average increase in comparable store sales by 15 percent within just one calendar quarter.
In fact a recent Gartner report said that by 2016, 70 Percent of the most profitable companies will manage their business processes using real-time predictive analytics. The future is here, is your retail business going to be part of that statistic?
By: Yan Krupnik, Marketing Manager, Retalon