This is an English adaptation of a FoodBud historical article originally published on February 12, 2022.
FoodBud, citing Starbucks financial disclosures, reported that Starbucks China generated US$3.62 billion in revenue over the prior year, equivalent to about RMB 23 billion.
As of January 2, Starbucks China had 5,557 stores in China. Its pace of new openings slowed in the prior fourth quarter, mainly due to the pandemic, and same-store sales in China fell 14% in that quarter.
Within China’s chain beverage market, Starbucks remained far ahead on revenue scale, whether compared with tea chains or coffee chains. Luckin Coffee had not yet released its Q4 2021 results at the time; its total revenue for the first three quarters of 2021 was RMB 5.53 billion. Because Q4 was also affected by the pandemic, FoodBud estimated Luckin’s full-year 2021 revenue at RMB 9-10 billion, still far below Starbucks China.
Starbucks tracks weather changes and uses them in operations. On cooler rainy days, higher-calorie, more indulgent products tend to sell better. In hot weather, new products tend to perform better.
Starbucks had previously planned to deploy 4,000 AI-enabled coffee machines in its U.S. stores, with the expectation that all U.S. and Canadian stores would have AI-enabled coffee machines by 2022.
The system, named Deep Brew, was developed with Microsoft Azure. Starbucks intended it to create a smarter customer experience. Based on reinforcement-learning models, Deep Brew could recommend products according to a customer’s taste, the most popular choices at the local Starbucks store, and even current weather conditions.
The same system also supports more precise inventory management by using real-time inventory data and forecasts of future inventory changes.
Starbucks also turned its machines into IoT devices that generate dozens of data points. Data from individual machines helps the company maintain product standardization and quality across stores.
For example, if the water temperature or espresso pump pressure falls outside expected parameters, corrective action can be taken. More importantly, Starbucks began building predictive-maintenance models from IoT data to forecast when a machine may go down or need servicing. That allows replacement parts or machines to be sent proactively, reducing downtime, customer frustration, and lost revenue.
Starbucks can use systems to identify store availability and idle time slots. That is one reason it partnered with Meituan in China to explore rentable space services and improve store utilization.
Since its founding, Starbucks has embedded the “third place” concept into its brand: a place outside home and work where people can spend time.
For a space-led operator like Starbucks, the question is not only how to sell good products, but also how to improve the commercial efficiency of store space.
In the prior year, Starbucks China tested several formats. The first Starbucks shared-space concept store in mainland China opened in the office tower at Shanghai Raffles City. The store covered about 200 square meters, had nearly 100 seats, and was divided into four zones: a sofa area, leisure area, semi-open individual work area, and paid meeting rooms.
The sofa and leisure areas resembled traditional Starbucks stores, but the layout also included long tables, two-person tables, four-person tables, and sofas to support mobile work and small-team collaboration.
Along the windows were semi-open individual work areas and paid private meeting rooms. The semi-open design preserved some privacy while helping customers focus. The store also had three four-person discussion rooms and one eight-person meeting room. The four-person rooms were priced at RMB 50 per hour, while the eight-person meeting room was RMB 180 per hour, making them high-frequency-use spaces.
Starbucks China then announced a partnership with Meituan to launch the exclusive “1971 Living Room” space service and added the “Starbucks Delivers” function. Starbucks also became an early user of Meituan’s new “Super Store” function, which was planned to be fully available to users by the end of 2022.
Before the Meituan connection, Starbucks’ delivery orders came 50% from Ele.me, 30% from its own app, and 20% from WeChat or Alipay mini programs. That mix was expected to change after Meituan delivery integration. According to Starbucks’ latest financial disclosure at the time, digital orders already accounted for 38% of Starbucks China’s business in the latest quarter.
Starbucks’ product portfolio can be understood as a ladder. Frappuccino and Caramel Macchiato serve as entry-level products. Latte is a mid-stage product. Black coffee and Starbucks Reserve coffee target more professional coffee consumers.
In Starbucks China, beverages accounted for 70% of total revenue, while Frappuccino accounted for 30% of overall sales revenue.
In May 1994, two employees who had worked in soft-drink stores experimented with existing ingredients and created a high-quality blended beverage. That same year, Starbucks acquired Coffee Connection and kept the name of its milkshake-style drink, adapting the formula under the Frappuccino name.
In summer 1995, Starbucks introduced Frappuccino across all U.S. and Canadian stores. The product sold 200,000 cups in its first week, and Frappuccino represented 11% of total revenue that year, according to financial reporting. In 1996, Starbucks also launched bottled ready-to-drink Frappuccino.
Frappuccino expanded Starbucks’ product matrix, attracted consumers who did not previously have a coffee-drinking habit, and increased revenue outside coffee’s main dayparts, generally morning and afternoon.
Starbucks uses A/B testing in marketing. For example, for a push campaign targeting 10,000 people, it can first test two groups of 500 users, select the better-performing version, and then send it to the remaining 9,000 users.
A/B testing is not only used by internet companies. It is also widely used in fast-moving consumer goods. Earlier email-marketing programs often used personalized marketing strategies; retailers such as Watsons also needed A/B testing to evaluate campaign plans, identify the most effective version, and then scale promotion.
Starbucks’ ready-to-drink products also help the company test the consumption potential of future store markets. Store operations and retail products can reinforce one another.
By broadly placing ready-to-drink products in convenience stores and other retail channels, Starbucks can use that retail network to assess the potential and spending capacity of covered markets. This can help Starbucks make more efficient site-selection decisions.
The article also noted that Black Gun Coffee, a coffee brand that had launched a few days earlier, could use stores to drive growth in nearby DTC business, including online direct sales of coffee beans.
Whether in coffee or tea chains, Chinese brands such as Heytea and Nayuki were also experimenting with retailized products. Retail products and store businesses can complement each other: they allow brands to reach target customers more frequently, serve different occasions and price layers, and build a more three-dimensional business line.
Note: Revenue estimates, rollout targets, and 2022 plans above are historical figures from the original February 12, 2022 article.