Back-to-school season is here and, thanks to a weak first half of the year for retailers, it’s holding a little extra significance. A harsh winter cooled off retail sales early on in 2014 and caused the National Retail Federation to lower its full-year forecast. Still, the second half is supposed to regain some strength – thanks in part to an increase in back-to-school shopping. According to the NRF’s annual survey, total spending on back-to-school items is expected to reach $74.9 billion—up from $72.5 billion in 2013.
To be successful in this crucial and competitive period, retailers must make sure they are up to date with the latest analytics technology. Making customers happy – and, in turn, making them buy – is obviously key to a successful back-to-school season. Luckily, data about social media sentiment, call center feedback, buyer behavior, survey responses, sales, and more all hold clues as to what’s needed for a successful back-to-school season, from the right marketing campaigns to the right product mixes and everything in between.
Of course, there are different types (or levels) of retail analytics to this end. The most basic type is a simple volume assessment of your data, which involves counting the occurrences of issues and taking action as the volume and importance increases. The next progression 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.
But the most effective and advanced analytics – and those that have the most impact on back-to-school success – use all your data sources for predictive modeling.
By using retail analytics for predictive modeling, companies can actually see and understand how sentiment, emotion, and actions have changed; determine what is influencing that change; and make adjustments as needed (preferably in real time). This can take place 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 employed this back-to-school season.
1. Perfecting promotions.Customers are constantly giving feedback – in the actions they take, in the words they use to describe a store or company on social media and review sites, and in the words used in survey responses and call center interactions. By pairing data on customer behaviors with feedback, retailers can understand the root causes of buyer behavior.
For example, let’s say a retailer noticed that fewer customers were talking about their coupons and redeeming them less during this back-to-school season than in previous years. The basic level of retail analytics would alert the company to this difference. But by digging deeper into the data, the retailer could figure out exactly why – whether it was because of the checkout process, or because the coupons were expiring too quickly, and so on. Armed with this data, the company can then make an adjustment to make the coupons more effective for the remainder of the shopping season.
2. Preventing showrooming. Retailers know that customer loyalty is important and listen very carefully to loyal customers. But many retailers don’t understand why certain shoppers don’t buy their products and why showrooming occurs. Once again, this data can be found on Twitter, Facebook, and fan forums. When 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 make the necessary adjustments to solve the problem.
3. Adjusting the product mix.Another key part of retail analytics involves looking at historical customer feedback, whether from recent months and campaigns or from previous back-to-school seasons, to plan product mixes and pricing, anticipate changing customer demand, and more. To dive a bit deeper, this kind of predictive modeling means comparing historical trends with current trends and look for changes in rate, or the standard deviations. This is especially the case as the back-to-school shopping season starts earlier and earlier each year. As that trend takes place, retailers can analyze sales of specific products and adjust their orders as necessary. For example, let’s say Target notices that its sales of glue sticks were trending up early in the month. It can use that information to adjust its plan and determine which stores need re-orders.
4. Real-time response. Once again, predictive modeling can take place at a more individual level, allowing retailers to target specific customers in real time to improve sales. For example, retailers can use data to identify upsell opportunities or predict (and hopefully prevent) churn. And online retailers also use predictive modeling on a more individual level. They can use a combination of demographic data and activity to help the shopper to have a better omni-channel experience. For example, demographic data may show the retailer that a woman is a mother living in the suburbs, and her search activity may show that she is looking for kids’ shoes. Online retailers can use that information to predict that the mom also might need school uniforms and make those easier to find on the website. 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.
Predictive modeling can play a key role in retail analytics, allowing companies to detect changes, model trends, and act on that information to improve the customer experience and customer loyalty and, in turn, improve sales.
Sid Banerjee is the CEO and Co-Founder of Clarabridge. Over his career, Sid has amassed nearly 20 years of business intelligence leadership experience. Prior to Clarabridge, he co-founded Claraview, a leading BI strategy and technology consultancy firm. A founding employee at MicroStrategy, he held Vice President-level positions in both product marketing and worldwide services. Before joining MicroStrategy, Sid held management positions at Ernst & Young and Sprint International. Sid has a B.S. and M.S. in Electrical Engineering from the Massachusetts Institute of Technology.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.