Build personalized customer connections, improve efficiency, and make intelligent, data-driven decisions using machine learning and image recognition.
It’s 7:15 in the morning and a customer – Amy – is on her way to work. She is on the hunt for coffee and breakfast and stops at a convenience store. She opens her store app in search of inspiration and offers.
The first offer Amy sees is for a pepperoni pizza, the second offer is for a Mountain Dew and the third is a sandwich – not coffee and not breakfast. Annoyed and dissatisfied, Amy pours herself a simple small coffee and leaves, forgetting she was hungry.
This customer might not be lost to this store forever, but she would surely have purchased more if the app supplied her with relevant offers meeting her needs at that exact moment.
Amy isn’t alone. According to a 2019 survey by Google/Ipsos U.S., “63% of smartphone users are more likely to purchase from companies whose mobile sites or apps offer them relevant recommendations on products they may be interested in.”
Machine Learning to the Rescue
7-Eleven Denmark, operated by Reitan Convenience, is a retail leader in technology innovation and adoption. In 2019, NACS awarded the International Convenience Retailer of the Year Award to Reitan Convenience for their forward-thinking approach to retail, communication, and technology. “Technology played an important role in helping customers navigate and choose products, with digital platforms providing nutritional information. 7-Eleven Denmark tied deals and loyalty initiatives into its app, which led to a significant increase in the number of downloads,” according to NACS.
7-Eleven was an early adopter, deploying a highly functional smartphone app to customers. App users can collect and earn rewards, play games, send gifts to friends and family and find stores and opening hours. Additionally, users receive offers for food service items – but not just random suggestions for every user.
Partnering with Liquid Barcodes, 7-Eleven developed an intricate machine learning algorithm that provides app users with highly-specialized offers based on time of day, day of week, weather, seasonality, prior purchases, and location. Additionally, by relying on machines to do the hard work of personalizing offers, the marketing staff doesn’t have to constantly sort offers and guess which promotions to give to an entire population of users.
Think back to Amy at the beginning of this article who unsurprisingly lost her appetite at seven in the morning when offered a pizza for breakfast. If she had been using such a highly personalized app, it would have told her she would like a croissant, a white chocolate mocha, and maybe a paleo shot for later.
Perhaps if Amy felt understood by her store, she would be back that evening on her commute home to see what the app might offer her for dinner. In fact, the app would also know that on weekends, Amy tends to choose more indulgent items as opposed to the healthier options she chooses during the week.
The Business Case for Machine Learning with Image Recognition AI
In late 2021, Liquid Barcodes and 7-Eleven Denmark conducted a study of its mobile app users. In just one week, they found that users were five times more likely to redeem one or more of the top three offers shown. Users also reported that the top three offers shown to them were 40% accurate to their preferences, which indicates that the app is perceived as very well personalized compared to a static group with non-personalized offers.
Miguel Regueira, Data Scientist for Liquid Barcodes, works extensively with customer data to ensure the output from the artificial intelligence programming is simply that – intelligent. For instance, the algorithm understands that it is summer and with weather data inputted into the system, it learns which customers will purchase a hot drink at noon even in the summer and which customers prefer a cold beverage. And it learns from purchasing behavior which customers prefer the energy drink offered in the front.
Once again, think back to Amy who would have been quite happy with her morning croissant and coffee. At the same time, same day, a gentleman – Bill – walks into the store that morning but Bill just finished his shift at the local hospital after working twelve long hours. He is starving and deeply in need of caffeine so he can get home with the energy to get his kids ready and off to school. Bill stops into this store and as he checks his app, he is suggested a small pepperoni pizza, candy bars, and an energy drink – but not just any drink, his favorite Monster Energy drink. Bill happily claims the offer and heads off to conquer what remains of his day.
What’s New? Machine Learning now with Image Recognition AI!
The machine-learning algorithm that provides app users with highly-specialized offers based on time of day, day of the week, weather, seasonality, prior purchases, and location continues to evolve. Miguel and his team at Liquid Barcodes have added image recognition machine learning with image recognition AI to the model to make the algorithms even more precise. 7-Eleven Denmark runs several promotions for bottled drinks at any given time and the program is even smarter now. It has learned that Bill chooses images that have a pink or white aluminum can (Monster Energy) more frequently than an image with a gray plastic bottle with red lettering (Coca-Cola). Amy, on the other hand, always chooses pastries in the morning at that time so the program chose a croissant promotional image to share.
Both customers were in the store at the same time, using the same app, but it is notable that each saw different coupons. “By mining the customer information and analyzing session data, we were able to build a model that not only tracks offer redemption, but our predictions have also resulted in 3.5 times greater accuracy in promotions personalization which leads to increased sales for the retailer and higher customer engagement with the app,” says Regueira.
When both customers return later in the day, they will receive different offers. For Amy, her first three coupons will be for a salad, followed by juice and finally, candy bars.
For Bill, he will see first a smoothie, then a wrap, and finally a Coke because the program learned he likes to buy those items together.
Promotions Personalization Delivers Increased Revenue
Customers prefer to interact with brands that know them, care about them, treat them with respect, and most of all, are relevant in their time of need. There is nothing more annoying than advertisements that completely miss the mark, suggesting a product at the wrong time of day or at the wrong place in time. Worse yet, being served ads that ‘overheard’ a conversation and presumed the customer wanted a product that was potentially offensive could lose their business.
Rewarding customers for shopping and interacting with an app in a way that makes them feel special is key to winning their future business. By taking the guesswork out of promotions, the customers are not only happier, but the marketing team can also focus on bigger and better things!
About Liquid Barcodes
Liquid Barcodes is a leading global loyalty and digital marketing technology company specializing in the convenience and foodservice industries. The proprietary, cloud-based technology platform allows retailers to create and manage their digital marketing campaigns with a process called the “customer connection cycle’ to engage, promote and reward customers’ activities in real-time across digital and media channels.
Liquid Barcodes loyalty platform is powering loyalty programs for industry leaders across several global markets, by offering unique machine learning, personalization, subscription, gamification, loyalty, and capabilities and has developed this app for use across 7-Eleven Denmark stores.
Learn more about Liquid Barcodes at LiquidBarcodes.com.
About the Author
Carolyn Schnare, Chief Content Officer, Liquid Barcodes
Carolyn Schnare has been involved with the convenience retailing industry for two decades with extensive knowledge of customer engagement and marketing, sustainability, and community outreach. Prior to Liquid Barcodes, Schnare worked at NACS (National Association of Convenience Stores) in a variety of roles from event production to membership and most recently as Director of Strategic Initiatives and host and producer of the popular industry podcast, Convenience Matters.
Test Drive the 7-Eleven Denmark App
Powered by the machine-learning algorithm, 7-Eleven Denmark‘s mobile app combines key insights about the user, as well as the time of day they are viewing the app, and other factors such as the season and weather, in order to intelligently decide what coupons should be shown in each moment for each member.