“The most successful retail companies are utilizing data science and predictive analytics (PA) to improve efficiency, improve marketing campaigns, and gain significant customer insight for a competitive advantage” says Christine Kern, contributing for Innovative Retail Technology. But what about the “not so successful” retailers? How can they share in the advantages that Big Data and PA offer? Retailers can – by using predictive analytics.
What is Predictive Analytics?
Predictive analytics is a set of business intelligence technologies that uncovers relationships and patterns within large volumes of data that can be used to predict behaviour and events, according to Eckerson (2007) 1. Or, as Eckerson states it more bluntly “Predictive Analytics is like an “intelligent” robot that rummages through all your data until it finds something interesting to show you.”
Also, forecasting is about predicting the future, and predictive analytics adds questions regarding what would have happened in the past, given different conditions. Therefore, PA attempts to quickly and inexpensively approximate relationships between variables while still using deductive mathematical methods to draw conclusions 2.
Gregg Brunnick, Director of Product Management & Technical Services, Business Systems Division, Epson America explains the usefulness of PA: “If you know how many cheeseburgers John sold during last Tuesday’s lunch hour, for instance, you can improve the efficiency of your food ordering, preparation, labor, and marketing operations.”
The value of Predictive Analytics for retailers
Deon Abott of Smarter HQ writing in Inside Big Data, suggests that data science and predictive modeling have become the holy grail for the retail industry. For this reason retailers built reports summarizing customer behavior using metrics such as conversion rate, average order value, recency of purchase and total amount spent in recent transactions.
These measurements provided general insight into the behavioral tendencies of customers. However, says Deon “In order for retailers to create a meaningful dialogue with customers that honors the shopper’s preferred level and mode of engagement, it takes more than summarized reports, which is why customer intelligence and predictive analytics provide the opportunity to significantly change the retail marketing industry.”
Generic uses of Predictive Analytics are according SAS the following:
- Detecting fraud. Combining multiple analytics methods can improve pattern detection and prevent criminal behavior. As cyber-security becomes a growing concern, high-performance behavioral analytics examines all actions on a network in real time to spot abnormalities that may indicate fraud.
- Optimizing marketing campaigns. Predictive analytics are used to determine customer responses or purchases, as well as promote cross-sell opportunities. Predictive models help businesses attract, retain and grow their most profitable customers.
- Improving operations. Many companies use predictive models to forecast inventory and manage resources. Predictive analytics enables organizations to function more efficiently.
- Reducing risk. Credit scores are used to assess a buyer’s likelihood of default for purchases and are a well-known example of predictive analytics. A credit score is a number generated by a predictive model that incorporates all data relevant to a person’s creditworthiness.
Erick Siegel of Big Think suggests that predictive analytics allows for a keen assessment of the probability that any one person will buy, sell, click, lie, die, etc. PA doesn’t just predict the future; it can influence it as well.
The challenges of using Predictive Analytics
The big challenge for retailers is to use PA correctly. Not using PA appropriately can cause loss of brand equity and market share with astonishing speed. The key is in understanding the customer’s “digital body language”, suggests Earley (2014) 3. Retailers need to understand customer data – the attributes, needs, characteristics, life stage, behaviour, demographics, and psycho-graphics. The information coming from the data may be used to help customers behave in a way that satisfies their needs 3.
Unfortunately, the use of PA by some retailers has been reported as controversial. Not only are most companies not informing their customers of when and what data they are collecting, but they are not letting them know about their analysis policies, according to Corrigan et al (2014) 4.
According to Arliss Coates from EConsultancy retailers should note the following when using PA:
- Is automation driving out your innovation and originality?
- Do you have people that know how to interpret the results of PA?
- Scenario planning – humans cannot prepare the machines to anticipate every possible nuance or scenario.
- An over-reliance on data to substantiate decision-making may hampers innovation.
- The “garbage in, garbage out” principle – bad data will render bad results.
The explosion of data is here to stay. At this moment it seems that the availability and use of big data and predictive analytics will grow exponentially. In spite of some controversy and challenges, PA couldn’t have come at a better time for retailers. Predictive analytics may help retailers to integrate their channels more smoothly and thereby keeping in pace with their competitors.
Read also: Big Data for Small Retailers – Is it Doable?
Have a look at this practical demonstration of PA from IBM “”Predictive Analytics for Retail – Introduction”:
1 Eckerson, W.W. 2007. Predictive Analytics. Extending the Value of Your Data Warehousing Investment, TDWI Best Practices Report, Q1.
2 Waller, M.A. and Fawcett, S.E. 2013. Data science, predictive analytics, and big data: a revolution that will transform supply chain design and management, Journal of Business Logistics, 34(2):77-84.
3 Earley, S. 2014. Big Data and Predictive Analytics: What’s New? IT Professional, 16(1):13-15.
4 Corrigan, H.B., Craciun, G. and Powell, A.M. 2014. How does target know so much about its customers? Utilizing customer analytics to make marketing decisions, Marketing Education Review, 24(2):159-166.