Predictive Modeling Brings Holiday Cheer To Retailers

Black Friday — the Super Bowl of the retail industry — has come and gone, but holiday shopping is just getting started. November and December combined are two crucial and competitive months for retailers — and the good news is that the National Retail Federation expects sales during that period to increase 4.1% year over year, bringing total sales to $616.9 billion and dwarfing the 3.1% increase during the same time last year.

 

One way retailers can make the most out of the holiday shopping season is to ensure their company is up-to-date with the latest analytics technology. Making customers happy — and, in turn, making them buy — is imperative to a successful and cheerful holiday season. Luckily, data about social media sentiment, call center feedback, survey responses, buyer behavior, sales and more, hold clues as to what’s needed for a successful season — from the right marketing campaigns to the best product mixes and everything in between.

 

Of course, there are different levels of retail analytics to this end. The most basic type is a simple volume assessment of your data, which is the process of counting the number of occurrences of issues and taking action as the volume and importance increases. The next type is change analysis — that is, looking at the rate of change in the data, including spikes, and then determining the next best action based on dramatic increases. However, the most effective and advanced analytics — and those that have the most impact on holiday shopping season success — uses all data sources for predictive modeling.

 

Through the use of retail analytics for predictive modeling, companies can clearly see and understand how sentiment, emotion and actions have changed; determine what is influencing that change; and make the adjustments necessary in real time. This can occur at both a trend level and an individual level.

 

With that in mind, let’s take a look at four specific ways that this advanced level of retail analytics can be integrated in retailers’ marketing initiatives this holiday season.

 

Customers are consistently providing feedback, whether it be in the actions they take, the words they use to describe a store or company on social media and various review sites, and the words used in survey responses and call center communications. By narrowing down data on customer behaviors with feedback, retailers can dig into the root causes of buyer behavior. 

 

For example, if a retailer noticed that fewer customers were discussing their coupons and redeeming them less during this holiday season than in previous years, the basic level of retail analytics would alert the company to this change. But if retailers dig deeper into the data, they could figure out why — whether it was during the checkout process, or because the coupons were expired in a short time frame, and so on. With this data, the company can take actionable steps to adjust their coupons to be more effective for the remainder of the holiday season.

 

Customer loyalty is important, and retailers know to listen very carefully to committed customers. But there’s a challenge: Many retailers don’t understand why particular shoppers don’t buy their products and why showrooming occurs instead. The good news is, this data can be found on Facebook, Twitter and other fan forums. When data is monitored in real time, retailers can determine if there is a problem with the product, the staff at a location, the pricing or even the merchandising — and then take the necessary actions to solve the issue.  

 

Looking at historical customer feedback from past holiday seasons is also important so retailers to plan the best product mixes and pricing, anticipate changing customer demand, and more. More specifically, this type of predictive modeling encompasses comparing current trends with historical trends and looking at the changes in rate. Retailers can analyze each years’ trends in sales of specific products and adjust their orders as necessary. For example, let’s say Macy’s notices that its sale of Nespresso was trending up early on in the holiday season. It can use that information to adjust its plan and determine which stores need re-orders. 

 

As stated earlier, predictive modeling can take place at an individual level, allowing retailers to target specific customers in real time to improve sales. For example, retailers can use historic data to identify upsell opportunities or to predict churn. Online retailers use predictive modeling on a more individual level as well. They can use a combination of demographic data and activity to help the shopper have a better omnichannel experience. 

 

For example, demographic data may show the retailer that a man is located in the suburbs and his search activity may show that he is looking for a garden hose. Online retailers can use that information to predict that he also might need a lawn mower or garden tools and make those easier to discover on the web site. Similarly, real-time response and predictive modeling includes monitoring for abandoned online carts and comparing abandonment rates to previous rates — and then, of course, figuring out why the rate has changed.

 

All in all, predictive modeling will play a key role this holiday shopping season, allowing companies to detect changes, model trends and generally use data to improve the customer experience and, in turn, improve sales. 

---