Selling goods or services to customers through several channels and to earn a profit is a complex concept of knowledge and skills. Walmart stands first in the world as a retailer serving more than 324 million customers with 20,000 stores in 28 countries. Whether it is in-store purchases or social mentions or any other online activity, Walmart has always been one of the best retailers in the world. Walmart ensures its best service to the customers by keeping customer satisfaction as its highest priority.
On the other hand, it is also well-known for its successful operation in more than 10 websites, being operated all over the world. On knowing these facts, it is evident that Walmart handles an enormous amount of data varying in its types and sources of it.
Though the retailers satisfy demand identified through a supply chain, data plays an important role in making decisions based on facts, trends, and statistics. Thus, the business must be able to float through the noise of the increased size of data and extract the right information so that they can make the right and best possible strategy. Also, data helps to understand and improve the business by reducing the wastage of time and money.
Likewise, modern Walmart’s business marketplace is a data-driven environment and it is in the process of building the world’s largest private cloud, capable of coping with 2.5 petabytes of data every hour.
Here are more details of how the data is being collected, stored, transformed, and used by Walmart in this report.
Types of data:
As mentioned earlier Walmart operates with more than 20,000 stores, serving more than 1.1 million customers per hour. Any data that is extracted out of these stores(online and offline) and customers are stored in a Neo4j(reference 1) database consuming 200 billion rows, ‘representing data of only a few past weeks’. The information collected has more than one source, including meteorological data, economic data, telecom data, social media data, product price and details of local events. Walmart has an immense amount of data at its fingertips. Vast and varied datasets are used to tackle all the data and to generate a real-time solution in microseconds through designed algorithms.
Chart 1: Overview of Data Sources.
As it is observed from chart 1, Walmart’s data universe is filled with varieties of data from three different categories.
1. Online data:
The Walmart registered mobile app users are more than 26 million as of June 2016. It is observed that mobile application users spend 40% more compared to those customers in stores. (reference 2)Another interesting fact is that more than half of the customers enter the Walmart store with a smartphone leading to another opportunity for Walmart to promote its digital and online payments. Therefore, Walmart makes future predictions and decisions based on the records extracted through the data of all mobile and website-registered customers. A record is a set of email ID, contact numbers, addresses, usernames and passwords so all the details provided by the customers are authenticated by Walmart before saving the records.
On the other hand, social media has inexplicable followers for Walmart with 1.1millions on Twitter, 2.3 million on Instagram and 32 million on Facebook. It has become one of the most important platforms for marketing, advertising and preparing a strategy for Walmart’s business.
Overall, mobile applications, websites, and social media contribute to the data stored by Walmart in building its business strategies.
2. Offline Data:
Walmart’s in-store data has a wide variety of sources. Some of them are, sales tracking through the scanned UPCs, Item stocks and inventories through off-shelf and on-shelf items count, data on lost and shrink items, PLU for the cashier’s usage, item discovery and it’s positioning through the tele-zone usage, credit, and debit card track records, Walmart master card registrations, user details collected due to western union transactions and employee everyday login records.
Streamlining customer experience, Walmart also collects data on in-store customers through connection to its Wi-Fi connection spots, and tracks and stores data from call centers, feedback and experiences.
Walmart has a great relationship with weather forecast companies to predict climate data for years so they make likely correlations between weather and store sales on a zip code level. Even when those correlations make no obvious sense sometimes still it might be useful in most of the cases. Walmart makes better business decisions with the available weather data to transform the customer experience.
Some of the facts that declare that Walmart operates with big data:
- a typical supercenter sells approximately 46 million items in store.
- with 12 million items in the store compared to 35 million products in the online store.
- serves 1.1 million customers per hour. [exhibiting the high velocity in ETL of the data]
- operated with the help of 2.2 million employees.
- Walmart is successful in handling such huge data also with the global net sales amounted to about 500.34 billion U.S. dollars being one of the largest employers in the world. reference 3 for the facts and charts.
Based on the above statistics Walmart stores massive volumes of data for their operations and forecasts.
Types of big data analytics techniques:
The analytics systems at Walmart cover millions of products and customers from different sources. Thus, the system analyses around 100 million keywords on a daily basis to optimize each keyword. implements the following technologies in handling big data.
A single day’s data in Walmart account to multiple terabytes of new data. And, historical data sums up to a few petabytes making it complex and large data to be analyzed.
Hence, to handle such massive data, the big data ecosystem came into the picture with the collection of infrastructure, applications and analytics. The main objective of analyzing this ecosystem is to optimize the shopping experience for customers when they are in a Walmart store, or browsing the Walmart website or mobile devices.
Here we understand how Walmart brings data analytics and data mining techniques in practice.
Firstly, it is a fact that Walmart made use of Big Data since the technology came into existence. A migration from 10 node Hadoop cluster to 250 node Hadoop cluster by Walmart in 2012 was a stepping stone to have a transition from its aged methodologies to newborn technical knowledge. This transition corroborated to combine more than five different websites into one single website successfully to let a new latest cluster collect all the generated unorganized data.
A transformation from the relational database to the Neo4j database(reference 1), a graph database was much in need since the relational database did not satisfy the requirements, due to the complexity of the queries.
To know more about Neo4j, it is a graph database that could quickly query customers’ purchases and real-time recommendations. With Neo4j, Walmart could substitute a heavy batch process with a simple and real-time graph.
In the very beginning of Walmart’s journey with Hadoop, they developed few applications as mentioned below.
- A mapping application that lets its users identify even the tiniest product’s location in the store along with the store’s location.
- eReceipts application provides customers with electronic copies of their purchases.
- Savings Catcher application alerts the customers whenever its competitor reduces the cost of an item the customer already bought. This application then sends a gift voucher to the customer to compensate for the price difference.
Map update application:
To process and generate big data with parallel, distributed algorithms on a cluster. A cluster here is a set of connected computers that work together which can be viewed as a single system.
Map update application workflow
Walmart developed Mupd8 to tackle issues like performance and scalability while dealing with real-time processing applications which could emphasize on the quality of generated data.
Mupd8 allows developers to write applications easily and process them using the Map Update framework, an easy way to express streaming computation. Writing an application as a combination of customized map and update operators, big data developers can focus on the business logic of the application and let Mupd8 handle load and data distribution across various CPUs.
To address high availability, low latency and scalability, the current Mupd8 architecture leverages open-source solutions, including Cassandra. To conclude, at Walmart Mupd8 platform has already supported more than a dozen sophisticated stream applications processing over 300 million status updates per day, gathering real-time information (reference 4).
Data mining helps Walmart find patterns that can be used to provide product recommendations to users based on which products were bought together or which products were bought before the purchase of a particular product. Effective data mining at Walmart has increased its conversion rate of customers.
Moreover, Walmart tracks and targets every consumer individually. It gathers information on what customers buy, where they live, and what products they like through in-store Wi-Fi. The big data team at Walmart Labs analyses every clickable action on Walmart.com. To conclude, all these events are captured and analyzed intelligently by big data algorithms to recognize meaningful big data insights for millions of customers.
Actions taken by Walmart using Big data – (references 5 & 6)
1. Social media survey:
Big data helped Walmart launch new products into market analysis based on data collected from social media. For example, social media data let Walmart understand that the users were interested in ‘Cake Pops’. This data allowed Walmart to increase its sales by stocking the Cake Pops quickly in stores. Likewise, the likes and dislikes of customers concerning every product were well understood through big data analytics.
2. Scan & go:
A focus by Walmart in the past 2.5 years has been to improve the checkout process for 140 million customers each week. The Walmart Pay app rolled out nationwide last year allowed the consumer to scan their phone at checkout instead of having to pull out cash or credit cards. Scan and Go allows consumers to skip the checkout lane and is being tested in Wal-Mart’s U.S. stores, and is completely rolled out at Sam’s Club.
“By using predictive analytics, stores can anticipate demand at certain hours and determine how many associates are needed at the counters. By analyzing the data, Walmart can determine the best forms of checkout for each store: self-checkout and facilitated checkout,” Wal-Mart noted in the post.
3. Customized recommendations:
Walmart’s big data algorithms provide recommendations to its customers based on the purchase history analysed through credit card purchases.
4. Social media big data solutions:
A big part of Walmart’s unstructured, informal and generally ungrammatical data-driven decision is based on social media data- Facebook comments, Pinterest pins, Twitter Tweets, LinkedIn shares and so on. WalmartLabs is leveraging social medial analytics to generate retail-related big data insights.
Through the Social Genome analytics solution which combines public data from the web, social media data and proprietary data like contact information, email address and customer purchasing data, this data helps Walmart is reaching customers or friends customers who tweet or mention something about the products of Walmart to inform them about the product and provide them a special discount.
5. Inventory management at Walmart using predictive analytics:
Inventory plays a vital role in Walmart’s triumph. Walmart reduces losses, risk and wastage that takes place due to overstocking. Thus, big data helped Walmart to have a well-planned inventory management system that helps it stay stocked according to the in-demand products.
This, in turn, facilitated suppliers to plan the inventory supply based on the real-time data. So, both the retailer and supplier are benefited by saving funds.
6. Mobile big data analytics solutions:
Mobile users in Walmart account for 1/3rd of its traffic every year and Mobile phone customers are extremely important to Walmart as smartphone shoppers spend 77% more in-store.
Walmart is leveraging big data analysis to develop predictive capabilities on their mobile app. The mobile app lets a customer use the geofencing feature of Walmart’s mobile app. This app asks the user to enter into the “Store Mode” when the shopper enters the store. The store mode of the mobile app helps users to scan QE codes for special discounts on products they would like to buy.
7. Consumers in the produce department:
Walmart is using big data and IoT sensors to find out how long people laze around in the fresh produce department. This analysis has helped them find that if the fresh produce looks fresh enough then people loiter for longer and this is the secret to make customers buy more things from the Walmart stores.
It’s an old practice that people make paper lists before heading to the shopping. Walmart attempts to translate paper lists to digital tools which have not yet been widely embraced by shoppers. However, this new list feature of Walmart using data analytics lets users enter items in a natural language like ‘popcorn’ and then this app pushes a limited number of brands and lets the user choose a specific product. And the lists are integrated to the new store maps, each item can be located down to the four-foot-wide section on the shelf.
9. Restore & returns:
The shopper purchase data is integrated and managed in the app, giving authority to the user to scan the paper receipt using which the shopper will be able to initiate a return process within the app, select an item purchase from the history and create a barcode on the mobile device. Where in which this code can be presented at the customer service counter and simply drop off the return. Then the customer gets acknowledged regarding the return. However, there is no extensive usage of this application currently.
10. Use market basket analysis to classify shopping trips:
Walmart enhances customer experiences by understanding their store visits based on different trip types. Regardless of whether a customer is on the last minute run or taking a stroll down the store. Classification of this kind helps Walmart enhance the customer shopping experience. However, these trip types are created based on previous purchase history.
Conclusions & views:
Minimizing the operating costs: Data analytics in Walmart helps in its planning, scheduling, and control of activities which are involved in transferring goods from its warehouses to the stores. Thus, data analytics assists Walmart to best position its warehouses nearer to its stores minimizing operational costs.
Advanced supply chain management: By Walmart’s best practices in supply chain management, and technological advances such as the barcode and RFID, it has assured that it remains competitive regarding its prices in the market.
Dealing directly with suppliers: Walmart stores details of how many products need to be shipped and produced. Walmart presented it as a way for suppliers to partner with it to improve efficiency in inventory management and meeting customer needs. Walmart shifts its shipping responsibility of inventory management to the supplier.