How does Starbucks use machine learning (ML)?

First published on February 23, 2022

 

3 minute read

John Patrick Hinek

Growth

TLDR

Starbucks uses machine learning (ML) to recommend relevant products to their customers and improve business function. Their success shows ML’s power in transforming any business. 

Outline 

  • Intro

  • Reinforcement Learning

  • ML for Business Strategy

  • Closing

Intro

Discussions about machine learning (ML) use cases often center on a tech company's ability to leverage ML to improve their software service. While ML has dramatically changed the way tech companies operate, its use cases expand far beyond purley tech products. Making use of their giant database, Starbucks is on the forefront of ML integrations.  

Starbucks' entry into big-data collection came in 2011 with the introduction of their mobile app. Initially building traction with the introduction of app-based rewards points, or stars, the Starbucks app now processes over 100 million transactions a week. As a result of their app’s success, a consistent collection of data is able to be generated daily. Deep Brew, Starbucks’ artificial intelligence (AI) initiative, looks for tech-based solutions to optimize their business function and customer experience. 

Reinforcement Learning

The data Starbucks gathers about their customers aids them in creating a personalized user experience through their app. Starbucks deploys reinforcement learning to offer up product recommendations. Reinforcement learning is a type of ML which rewards actions by learning from its past successes and mistakes. Through trial and error, reinforcement learning finds the best path to a reward. For Starbucks, this reward is a purchase. 

Built on Microsoft Azure, Starbucks deploys reinforcement learning into their app where it gains powerful data-driven insights on its customers. The ML will register each transaction put forward by the customer as a data point. It will then use this data to begin recommending products that will best suit the buying habits of the customer. While reinforcement learning doesn’t need much data to start recommending products, the more data points it has, the more accurate its recommendations will become. 

In use, if the buying habits of a Starbucks customer show a strong preference for tea, the ML is able to push recommendations for tea-based products. This not only keeps the customers' preferences top of mind when they open the app, but also introduces them to new products which data suggests they might enjoy. Starbucks senior VP of analytics and market research, Jon Francis, said “just like their relationship with a barista, customers receive the same care and personalized recommendations when it comes from our digital platforms.” 

ML for Business Strategy 

Data gathered on the Starbucks app informs the company about more than just customer’s drink preferences. The Starbucks app is able to track a product’s popularity, the exact time orders were made, and locations garnering the most traffic. Aided by AI and ML, Starbucks uses all of these metrics to improve their business strategy. 

Data gathered at Starbucks can give insight on the popularity of certain stores during operating hours. Starbucks uses this information to gauge inventory needs and optimize scheduling by hiring more or less people depending on the history traffic of a store. Starbucks claims the success of this implementation has allowed its employees to focus on building customer connections, instead of scheduling and calculating inventory. 

Starbucks uses their data and AI to make revenue projections and determine the best location to open a new store. Data from Starbucks mobile app gives them quick insight on what stores are the most profitable. Analyzing store traffic and the location of the order placed, Starbucks is also able to pinpoint the optimal location for a new property without eating the profits of surrounding locations. Starbucks uses data gathered from surrounding stores and past revenue history to implement AI predictions on how profitable a new location will be. 

Closing

Even when selling a physical product, Starbucks demonstrates ML’s power in generating effective business decisions. Along with the ease of ordering via mobile, personalized product recommendations appearing every time a customer opens the app undoubtedly has elevated Starbucks to becoming the coffee, and ML data-driven company they are today. Starbucks creative use case for ML shows how powerful it is in improving any business. 

Start building for free

No need for a credit card to get started.
Trying out Mage to build ranking models won’t cost a cent.

No need for a credit card to get started. Trying out Mage to build ranking models won’t cost a cent.