Big Data at the service of retail

How retailers use big data to improve personalization in three key aspects for the buyer – assortment, offer and delivery, told in Umbrella IT

In the late 1990s, entrepreneurs from all walks of life came to realize that data is a valuable resource that, if used properly, can become a powerful tool of influence. The problem was that the volume of data increased exponentially, and the methods of processing and analyzing information that existed at that time were not effective enough.

In the 2000s, technology took a quantum leap. Scalable solutions have appeared on the market that can process unstructured information, cope with high workloads, build logical connections and translate chaotic data into an interpretable format that can be understood by a person.

Today, big data is included in one of the nine areas of the Digital Economy of the Russian Federation program, occupying the top lines in the ratings and expense items of companies. The largest investments in big data technologies are made by companies from the trading, financial and telecommunications sectors.

In 10-20 years, big data will become the main means of capitalization and will play a role in society comparable in importance to the power industry, analysts say.

Today’s shoppers are no longer a faceless mass of statistics, but well-defined individuals with unique characteristics and needs. They are selective and will switch to a competitor’s brand without regret if their offer seems more attractive. That is why retailers use big data, which allows them to interact with customers in a targeted and accurate way, focusing on the principle of “a unique consumer – a unique service.”

1. Personalized assortment and efficient use of space

In most cases, the final decision “to buy or not to buy” takes place already in the store near the shelf with goods. According to Nielsen statistics, the buyer spends only 15 seconds searching for the right product on the shelf. This means that it is very important for a business to supply the optimal assortment to a particular store and present it correctly. In order for the assortment to meet demand, and the display to promote sales, it is necessary to study different categories of big data:

• local demographics,
• solvency,
• buying perception,
• loyalty program purchases and much more.

For example, assessing the frequency of purchases of a certain category of goods and measuring the “switchability” of a buyer from one product to another will help to immediately understand which item sells better, which is redundant, and, therefore, more rationally redistribute cash resources and plan store space.

A separate direction in the development of solutions based on big data is the efficient use of space. It is data, and not intuition, that merchandisers now rely on when laying out goods.

In X5 Retail Group hypermarkets, product layouts are generated automatically, taking into account the properties of retail equipment, customer preferences, data on the history of sales of certain categories of goods, and other factors.

At the same time, the correctness of the layout and the number of goods on the shelf are monitored in real time: video analytics and computer vision technologies analyze the video stream coming from the cameras and highlight events according to the specified parameters. For example, store employees will receive a signal that jars of canned peas are in the wrong place or that condensed milk has run out on the shelves.

2. Personalized offer

Personalization for consumers is a priority: according to research by Edelman and Accenture, 80% of buyers are more likely to buy a product if a retailer makes a personalized offer or gives a discount; moreover, 48% of respondents do not hesitate to go to competitors if product recommendations are not accurate and do not meet needs.

To meet customer expectations, retailers are actively implementing IT solutions and analytics tools that collect, structure and analyze customer data to help understand the consumer and bring interaction to a personal level. One of the popular formats among buyers – the section of product recommendations “you may be interested” and “buy with this product” – is also formed based on the analysis of past purchases and preferences.

Amazon generates these recommendations using collaborative filtering algorithms (a recommendation method that uses the known preferences of a group of users to predict the unknown preferences of another user). According to company representatives, 30% of all sales are due to the Amazon recommender system.

3. Personalized delivery

It is important for a modern buyer to receive the desired product quickly, regardless of whether it is the delivery of an order from an online store or the arrival of the desired products on the supermarket shelves. But speed alone is not enough: today everything is delivered quickly. The individual approach is also valuable.

Most large retailers and carriers have vehicles equipped with many sensors and RFID tags (used to identify and track goods), from which huge amounts of information are received: data on the current location, size and weight of the cargo, traffic congestion, weather conditions, and even driver behavior.

The analysis of this data not only helps to create the most economical and fastest track of the route in real time, but also ensures the transparency of the delivery process for buyers, who have the opportunity to track the progress of their order.

Delivery personalization is a key factor for the buyer at the “last mile” stage. A retailer that combines customer and logistics data at the strategic decision-making stage will be able to promptly offer the client to pick up the goods from the point of issue, where it will be the fastest and cheapest to deliver it. The offer to receive the goods on the same day or the next, along with a discount on delivery, will encourage the client to go even to the other end of the city.

Amazon, as usual, went ahead of the competition by patenting predictive logistics technology powered by predictive analytics. The bottom line is that the retailer collects data:

• about the user’s past purchases,
• about the products added to the cart,
• about products added to the wishlist,
• about cursor movements.

Machine learning algorithms analyze this information and predict which product the customer is most likely to buy. The item is then shipped via cheaper standard shipping to the shipping hub closest to the user.

The modern buyer is ready to pay for an individual approach and a unique experience twice – with money and information. Providing the proper level of service, taking into account the personal preferences of customers, is possible only with the help of big data. While industry leaders are creating entire structural units to work with projects in the field of big data, small and medium-sized businesses are betting on boxed solutions. But the common goal is to build an accurate consumer profile, understand consumer pains and determine the triggers that affect the purchase decision, highlight the purchase lists and create a comprehensive personalized service that will encourage buying more and more.