Sunday, August 14

When the supermarket knows more about you than your own mother: this is how loyalty algorithms work


If there is something that makes one’s day happy, it is going to the supermarket and being able to take out that customer card that you didn’t even remember you had when the checkout staff asks “Membership card?” The simple fact of saying “yes, I have” fills you up to no end. Points? Discounts? We don’t even know what they are for. But what a joy to finally be able to use them! A mysterious offer on the shampoo you wanted, discounts on the pizzas you buy every weekend and even a promotion to win that champagne you could never buy.

It seems that the supermarket knows us better than we know ourselves. And not only supermarkets, but also clothing stores or department stores. There is an explanation.

For decades, one of the main problems that many companies have faced is customer loyalty. Now, in the new digital paradigm in the cloud, elements such as loyalty cards have become a key element for brands to maintain contact and trust with their customers. So much so that 81% of consumers Spaniards recognize that they have any of these cards, according to a study by American Express Spain.

The most requested, in the food sector, of course. No less than 65% of the total. Then, those at gas stations, where 4 out of 10 customers admit to having them, while almost 3 out of 10 consumers own one from the fashion, sports or perfumery sectors.

Why do people decide to join these loyalty clubs? Discounts or obtaining offers in the following purchases that are made. Specific, 39% of users assure that they transfer their data personal when you choose to enter a sweepstakes or receive discounts; while 35% do so to receive personalized promotions on their purchases. Only 2 out of 10 affirm that they do not give their personal data.

And on the other side are, of course, the companies. The main benefit of the retailer when running a loyalty program is not rewarding customers for their loyalty, but rather the collection of information in an attempt to foster a better relationship with customers and guarantee a greater market share. The element that will make the difference will be the algorithm that transforms all the information collected from the client and their interaction with the brand into action.

Also Read  Ransomware continues to plague organizations

And to know what these algorithms are about and what bases are taken into account to create them, at Xataka we have contacted experts in the creation of the most advanced loyalty cards in Spain.

How algorithms work

Comerzzia is a modular platform for comprehensive store management that allows retailers to optimize their customers’ shopping experience. “Companies have realized that in order to increase the average ticket they need to know the customer, and for this they have to go further with personalized commercial actions so that they buy things that they do not usually buy but that intersect with what they do normally buy”, explains José Luis Cordero, director of operations at Comerzzia.

All of this entails the design of that promotion strategy: offer coupons, points, events, trips and even discounts with third parties. “It is not about specific benefits, the client has to see them and notice them every day. It is not worth it with a coupon of 5 euros at the beginning and that’s itLamb adds.

84% of users of this type of card consider that the gifts received in a year cost less than 300 euros, and place the average value at 212 euros. But the use of loyalty programs varies according to age groups, provinces and lifestyles.

Lidl Plus App

Cordero explains that the algorithm can range from very simple to very complex: “Once I have all the documentation for that client (bank card, name, gender, address, etc.), with those tickets I am able to apply the segmentation algorithms for start treating them differently. The algorithm does not tell me that this segment is an athlete or a family, but it tells me that 3% spends on organic products, or spends it on fabric softener for children, for baby food or to lose weight.

Segmenting customers: infonomics

You can also find out how sensitive customers are to price increases, and the comparison between products between, for example, Hacendado or Mercadona’s own brands. To do this, most loyalty companies usually apply three phases: get to know the customer, analyze him (his sales and behavior) and act (endorse him with the promotional part so that he spends more on his visit to the store)

Also Read  High-capacity hard drives: the best models from 8 TB that you can buy

“We use our own algorithm: basically I give this artificial intelligence a ticket sample, a pattern or typology and the program gives me a percentage belonging to each employer”, says Cadero. Major patterns typically refer to an RFN (Recency, Frequency, Monetary) analysis. In this first segmentation, a differentiating first impression is already obtained. “We also use other algorithms aimed at knowing the behavior of the customer in more detail: we analyze from who buys gluten-free, who buys products gourmet and who spends almost nothing”.

“Then the fence is specified: Healthy, athletic… Depending on the company, the customer can already be segmented. One who buys at Massimo Dutti, for example, is already segmented by the simple fact of entering there. But we can achieve that all people who buy “eco” in Carrefour receive offers of soy milk”, adds Cordero.

How to have your customer and loyalty cards on your mobile, the best applications

Customer information is the key component in CRM, but many organizations do not manage it as another company asset and the monetary value of that information is ignored. The complex and sometimes non-existent legislation regarding consumer privacy and the very perception that society may have can create obstacles when it comes to monetizing customer information. The monetization of information is the practice to which they have coined the term Infonomics. “For us “Juan Pérez” is a number, it is what matters to us. It is an anonymized process, we do not care if it is Juan or Carlos. We want to see how it behaves,” Cordero concludes.

Contrary to what you might think, income level is not a determining factor when using a loyalty program. In fact, 91% of respondents with incomes above €50,000 use these programs whenever they can, compared to 75% with incomes below €40,000. The latter, moreover, are the ones who most often forget to use their loyalty cards (8%).

Big data to improve sales

Loyalty will be, according to the retail expertskey in the coming years. Supermarkets and Hypermarkets such as El Corte Inglés are absorbing part of the strategies of other sectors, specifically those streaming platformsto strengthen the consumption of customers who once opted for them.

Also Read  A good photo is not the same as a well-done photo: the eternal debate between technique and quality

The Día, Lidl, Eroski or Carrefour food chains use technology big data and analytics to break down consumption data and offer each customer a series of differential discount coupons. In the case of Día, the implementation of the pilot test improved coupon redemption by 40%, customer redemption by another 40% and 30% incremental sales in just three months.

Club Carrefour Card

“We opted for non-collaborative filter algorithms because it relates the dimension of the customer to the article. It is about algorithms machine learning looking for patterns in information that generates the interaction between the two verticals including the dimension of context. All part of the Day data system. There are three main entities: customers, articles and tickets. It is a giant matrix that the algorithm tries to fill in with information and then predict the missing gaps,” explained Día’s data scientist, Saúl Lugo, in this report.

One of the challenges facing the implementation of these systems is the connect the physical world with the world in the cloud and that the interaction is effective. It is striking, then, that the German distribution giant, Lidl, does not use a physical loyalty card, but rather everything is managed through the application.

In addition, there is the social situation that a loyalty card does not necessarily represent a person but a family, for example. The Juan Roig chain has never had a traditional loyalty program. Its business philosophy is based on always giving the best price, so it does not consider this leg a priority, given that its customer retention data, although it has decreased after the pandemic, remains at acceptable values.

Fidelity, what is that? The battle to collect customer information has begun. The objective is the same as when our grandparents bought: to improve market share. Now with mobile apps.

Images: Eroski, Lidl, Carrefour

Leave a Reply

Your email address will not be published.