This article first appeared in the Autumn 2015 issue of OnWindows magazine.
Founded in 1960, Domino’s is the recognised world leader in pizza delivery and digital ordering technology, with a significant business in carryout pizza. It ranks among the world’s top public restaurant brands with a global enterprise of more than 11,700 stores in over 75 international markets. Emphasis on technology innovation helped Domino’s generate approximately 50% of US sales from digital channels at the end of 2014, and reach an estimated run rate of US$4 billion annually in global digital sales. Domino’s features an ordering app line-up that covers nearly 95% of the US smartphone market and, in June 2014, debuted voice ordering apps – a true technology first within traditional and e-commerce retail.
Last year, the company’s CEO set forth a strategic directive that might have been a straightforward data management initiative in many businesses, but with Domino’s broad and varied sales technology profile, the directive presented some specific challenges. It required that Domino’s uniquely identify each of its customers; however, the data about who those customers are and what products they consume is scattered over, and embedded in, the customers’ order history with the company – comprised of more than one billion transactions from a period of several years; big data by any definition.
According to Domino’s CIO Kevin Vasconi: “In order to identify the customers who conducted the transactions, we would need to comb through each transaction, pull out any identifying customer information, then take all of that customer data, clean it and match it up across transactions to create a master record for each customer.”
This was additionally challenging as Domino’s transactional data consisted of orders submitted by phone, online and via mobile technologies, with potentially varying order delivery locations that might represent a combination of home, work, and/or school addresses, as well as carry-out orders placed at restaurant locations. The full range of these transactional variables across multiple orders might still represent a single customer’s purchasing activity. Customer identity stitching – being able to isolate a customer’s unique identity and link it to their entire ordering history – would enable Domino’s to analyse the customer’s buying preferences, in order to better target its marketing initiatives.
For this complex, high-volume, strategic initiative, Domino’s looked at several leading providers of master data management (MDM) solutions. Based on responses to a request for proposal, Domino’s invited each of the top responders to prepare a proof of concept (POC) of their respective matching capabilities, using Domino’s data, comprising more than three million records. The matching results produced by Maestro were the most accurate of all of the competing POCs.
According to project consultant Capgemini: “Profisee’s demonstration of superior matching capabilities; professional execution of the evaluation process; on-time, as-requested responses; attention to detail; and highly competent staff, proved that Profisee could meet Domino’s requirements, and help them achieve a successful implementation.”
Profisee collaborated with Domino’s business and technology stakeholders throughout the course of the initial implementation, through a Master Data Services Accelerator engagement. Working jointly with Microsoft, Profisee designed the accelerator to support the education, skills transfer, initial adoption and application development associated with Profisee’s Maestro MDM platform, built on top of Microsoft’s Master Data Services product. This collaborative approach expedited the design and development of the customer MDM solution and enabled Domino’s staff to take ownership of the solution once the initial implementation was completed.
Initially, the implementation team created a ‘sandbox’ environment to test matching strategies and perform address verification on a subset of 4.5 million records. Maestro’s support for multi-step, layered matching strategies was key to the success of the sandbox testing.
With the success of the initial test, the Maestro production environment was configured with the proven address verification and matching strategies, and approximately 550 million customer records (distinct customers inferred from existing order transactions) were loaded into the production system to be cleansed, verified and matched. From those, Maestro created 100 million ‘golden’ customer records – thus identifying the unique customers whose buying history was previously buried in the data scattered across one billion order transactions.
Domino’s can now identify unique customers from among billions of order transactions, and ensure that it has complete, accurate information on each customer, without duplication.
“Having cleansed customer master data to provide to our big data analytics platform allows us to understand individual buying patterns and product consumption. That allows us to create highly targeted marketing campaigns to optimise sales and profitability,” reports Vasconi.
Looking ahead, Domino’s will take advantage of Maestro’s incremental matching and look-up-before-create functionality to integrate the MDM solution into its operational systems, and plans to expand the solution to identify households in addition to specific individual customers.
With the success of the customer master data implementation, Domino’s has also expressed an interest in looking into a Maestro solution for mastering its product data, and potentially expanding the Maestro solution internationally.
Share this story