Big Data Implementation at Starbucks
Starbucks corporation is an American multinational chain of coffee shops and roastery. The world largest coffee shop chain with more than 32,000 stores worldwide (at the end of Q3, FY 20), was founded in the year 1971. They have modelled and remodelled their strategy to remain a market leader in the coffee business. Starbucks’ CEO, Howard Schultz, in his second stint was not hesitant in adopting a data-driven strategy. He supervised the closure of hundreds of stores and also planned the opening of the next 1500 stores, all through a data-driven approach.
Starbucks has implemented Big Data throughout its value chain, directly or indirectly (through feedback from one component to another).
Starbucks has implemented Big Data heavily in Operations, Marketing and Sales and Services, and the remaining components (Inbound and Outbound logistics) of the value chain receive data-driven inputs due to the tight integration of the entire value chain.
Operations
Starbucks has an asset-heavy business model; it operates through 32,180 stores in around 80 countries, of these 16,271 (50.56%) is owned by Starbucks.
The decisions about real-estate became entirely data-driven after Howard Schultz took over the company in 2008. Before 2008, Starbucks had a linear growth in the number of stores across the world-wide, which impacted its profitability, after 2008 we can see that the opening of new stores plateaued before gradually increasing 2012 onwards.
Aim: The objective behind this data-driven real estate investment approach was to minimize costs after the financial crisis of 2008. The operating margin dropped significantly from 11.2% in FY 2007 to 4.9% in FY 2008. Hence, there was a need to remodel the corporate strategy and improve margins for the coming years to retain the trust of the shareholders.
Methodology adopted: Howard Schultz approached the real estate investment decisions analytically and ensured that the Starbucks stores were strategically located, in places which would fetch high operating margin. He shut down stores which didn’t align with this strategy and kept the decision of opening new stores on hold.
To exact to its service needs of identifying locations to open new stores and to close those which hampered the profitability, Starbucks partnered with Esri, an international supplier of geographic information system software. Esri provided them with a technological platform, Atlas, to analyse maps and retail locations; the data (also known as Spatial data) under analysis were population density, average income, demographic distribution, traffic patterns, construction status, transportation node, etc. This approach was not only limited at a localised geography but instead, it was implemented at the corporate level across all geographies. (Rachel, 2020)
Analytics experts implemented a combination of predictive and prescriptive analytics. Atlas provided the spatial data on a map. Through a detailed spatial analysis, the experts could determine the average estimated customer count, average ticket size and average customer spend using a predictive model. The economic viability of opening a new store was determined by incorporating the cost structure in the model.
Spatial Analyses using Atlas Spatial data is provided on a Map, and these data sets are layered, with each layer giving information about a specific category.
Analyses can also be performed on a combination of layers using different operators and functions provided by the tool.
The spatial data had four components
· Geographic — feature data of a given location, the latitude, longitude and elevation.
· Attributes — description of the location, demographic information.
· Geometric — shape of the various objects in that location.
· Topologic — spatial relationship like connectivity, adjacency, inclusion etc.
Analyses steps:
· Data collection and layering.
· Reclassification of the layers into a meaningful range of values.
· Suitability index (of the same range) in each re-classified layers, to determine the extent to which the location satisfies the criteria under question.
· Overlaying the layers so that a weighted overall suitability index is calculated for a given location.
Marketing, Sales and Services
Aim: The objective was to optimise the menu, increase product offerings, expand the arena of business (from stores to other avenues like retail chains, grocery stores, etc.).
Menu offerings: The spatial analysis also helped Starbucks in making flexible, non-standard menu offerings based on the demand, weather conditions and customer preferences. The average ticket size information helped Starbucks in deciding how much capacity to install in each store, thus optimising its costs. (Rachel, 2020).
Product development and assortment: Apart from spatial data, Starbucks also collected customer information on how they ordered their beverages and their preferences, from the store, through consumer research and through social media websites to decide on product development and assortment. These data also helped Starbucks in making some of its products available at grocery stores so that customers can consume them at home. (Marr, 2018).
Personalisation: Starbucks has 14 Million+ customers registered in its ‘Reward-Loyalty Program.’ This program enables Starbucks to understand their consumer behavior and make product offerings as per their preferences. Starbucks has also used targeted marketing based on menu preferences and frequency of consuming from Starbucks. (Rachel, 2020).
The Starbucks mobile application also has a virtual barista feature, which takes orders from the customer through the voice commands, this is an artificial intelligence enabled feature, but the data to train and model this system comes from the enormous data collected by Starbucks.
Outcomes
Operations (Real estate decision outcomes): The spatial analyses helped Starbucks in opening stores which would fetch high margin and high return on assets. There are more than one store close to each other, but they are strategically located such that cannibalisation of sales doesn’t occur from the different nearby store(s). The operating margin grew from 4.9% in 2008 to 19.59% at the end of Q1 FY 2016. The Return on Assets (ROA) improved from 3.07% in 2008 to 21.03% in 2016. While there are some inconsistencies in the growth of operating margin and ROA, the overall trend has been an increasing one.
Marketing, Sales and Services: The Digital Menu Board enables Starbucks to display a dynamic and non-standard menu; this allows Starbucks to change its product display on the board and drive sales strategically.
The Loyalty program is an effective sales booster. Starbucks attributes 40% of its total sales to this venture. This program is available on the Starbucks mobile application, the membership been growing since 2008, and the vast data that is generated further serves as an input to improve the offerings, for example, the products to be displayed on the digital boards. The loyalty program also gives the customers ad-on benefits like meber events, send gifts to friends, etc. The most significant impact of the loyalty program is that Starbucks has gained a loyal customer base and that base also serves as a data source for its future digital strategy. (In, 2020).
Insights
Starbucks adopted technology for its strategy formulation at the right time. It has become an icon for a highly efficient supply chain and a customer centric organisation with the help of big data implementation, the cafés are located at all the location that anyone would expect, and the coffee shops are strategically located to minimise cannibalisation and maximise demand fulfillment.
The use of spatial data to decide on real estate investment, tells about the extent to which data analytics, particularly big data could help in revamping the entire business. Howard Schultz was able to foresee the future of data-driven decision making, and he implemented it to retain the competitive advantage that Starbucks has enjoyed for a long time.
Transactional data and other consumption related data has helped Starbucks in being updated with all the new data-driven marketing techniques, and personalisation methods so that it maintains its core value proposition of customer-centricity.
Reccomendations
1. Bundling of products: A conjoint analysis of product preferences of cutsomers will reveal appropriate basketing of products which will improve the sales significantly.
2. Inbound and Outbound Logistics: Big data implementation can be expanded to inbound and outbound logistics; this will enable Starbucks to lower the lead times further. SKU will be refilled timely, and inventory will be optimised. This will also minimise silo formation.
3. Prevention of data silos: With the vast amount of data under management across various functions, the formation of data silos is highly probable. Silo formation affects the analyses results and the output might not align with the goals of all the functiona divisions.
4. Use Case Management: The vast data can help in formulating in many use cases, some of which might be redundant, and hence use cases must be comprehensible and cover the entire value chain.
5. Data driven strategy formulation: To reap the maximum benefits of Big Data, it should be at the center of strategy formulation. And minute strategical changes must be made regularly with the help of newer insights that come from new data.
6. Data Consistency: Starbucks outsources some of its data management (spatial data) and uses in-house capabilities to manage other sets of data (transactional and customer related), this might lead to inconsistencies in data formats and data compatibility.
References
1. In, L. O. G. (2020). Assignment : Digital Winners and Losers Starbucks : Winning on rewards , loyalty , and data. 1–11.
2. Marr, B. (2018). Starbucks: Using Big Data, Analytics And Artificial Intelligence To Boost Performance. Forbes, 2018–2021. https://www.forbes.com/sites/bernardmarr/2018/05/28/starbucks-using-big-data-analytics-and-artificial-intelligence-to-boost-performance/#e8c71f365cdc%0Ahttps://www.forbes.com/sites/bernardmarr/2018/05/28/starbucks-using-big-data-analytics-and-artificial-i
3. Rachel, B. (2020). Assignment : Competing with Data Challenge Starbucks ’ secret ingredient : data analytics. 1–8.
4. Team, E. P. (2018). How to Perform Spatial Analysis. Www.Esri.Com, 1–6. https://www.esri.com/arcgis-blog/products/product/analytics/how-to-perform-spatial-analysis/
5. Wheeler, C. (2014). Going Big with GIS. 1–7. https://www.esri.com/about/newsroom/arcwatch/going-big-with-gis/?rmedium=arcwatch&rsource=http://www.esri.com/esri-news/arcwatch/0814/going-big-with-gis
6. Statista.com, https://www.statista.com/statistics/266465/number-of-starbucks-stores-worldwide/)