Aston Martin Lagonda Drives Sales Increase with Data Reply Using AI on AWS

Aston Martin Lagonda (AML), a British manufacturer of ultra-luxury high-performance sports cars and sports utility vehicles (SUVs), increased customer engagement, provided dealer partners with better customer insights, and improved predictions by 85 percent using a propensity-to-buy model powered by artificial intelligence (AI) on Amazon Web Services (AWS). Previously, AML’s customer data had been siloed across systems, making it difficult to access and share. As part of a migration to a new Salesforce instance, data analytics firm Data Reply cleaned and integrated the data using AWS services. Working with Data Reply, AML has successfully restructured its data, establishing a global customer view crucial for international clients.

Aston Martin Lagonda (AML) designs, creates, and exports its vehicles in 56 countries around the world, building on its rich heritage of more than 110 years. Due to the global nature of the brand, this has presented challenges in ascertaining customer insights from data across its key markets. Traditionally, AML collected data from multiple sources, including Salesforce, customer surveys, and aftersales interactions. The data was then stored in different siloed systems, making it a challenge for Aston Martin to garner a holistic view of its customer base.

Many of AML’s high net worth customers have multiple homes in different locations and have multiple vehicles in different regions. They have relationships with several dealers which caused duplication and variations in the quality of data collected. As a result, AML had lots of data about its customers but wasn’t always able to effectively use it to build on those existing customer relationships. “We were interested in using data to know our customers better,” says Andrea Senso, global head of data science and analytics at AML. “We wanted to understand them and what their journey is so that we could make our marketing mix more effective. We wanted to give each customer the high-quality experience they deserve.”

The British brand saw the opportunity to solve this problem when it migrated to a new instance of Salesforce set up by AWS Partner Data Reply. AML asked Data Reply to help with cleaning the data and improving its accuracy in the new instance. Data Reply used AWS to clean the data and merge multiple accounts into one using entity resolution management and other data-cleansing techniques. “It was a challenge actually identifying the same customer across multiple relationships with multiple dealerships,” says Senso. “When this data cleaning was done, it was the perfect time to bring everything together for an integrated 360-degree view.”

After the data was cleaned and integrated, Senso asked Data Reply to help create a single 360-degree customer view powered by that data. “Data Reply was already working with AML on the Salesforce transformation, and I had a good impression from previous projects,” says Senso. “We wanted to develop a minimum viable product (MVP) to see how our data could improve our marketing to customers.”

Meeting regularly, AML and Data Reply built a prediction model that uses artificial intelligence (AI) to analyze customer data. It takes customer data from Salesforce, runs it through Amazon AppFlow—a fully managed integration service to transfer data between services such as Salesforce, Google Analytics, and Amazon Redshift—and hosts all of the data on Amazon Simple Storage Service (Amazon S3), object storage built to retrieve any amount of data from anywhere. The solution uses AWS Glue—a serverless data integration service that makes it easier to discover, prepare, move, and integrate data from multiple sources for analytics, machine learning (ML), and application development. In addition to these AWS services, Data Reply built many of its own features.

The ML model was built using Amazon SageMaker, which allows the user to build, train, and deploy ML models for any use case with fully managed infrastructure, tools, and workflows. This used models like Amazon SageMaker Random Cut Forest (RCF)—an unsupervised algorithm for detecting anomalous data points within a data set. This allowed AML to use different data points, like time since last test drive, time since last service, and service history of the vehicle. That information could help reveal what led customers to purchase a vehicle. These insights were then used to predict which other customers were likely to purchase in the near future. “We built models specifically for AML on AWS to create an MVP,” says Thomas Silk, Business Unit Manager at Data Reply. “It was a combination of proprietary tech we built on AWS and ready-made components from Amazon SageMaker. A bespoke solution meant AML could get the most out of its data.”

The MVP that AML and Data Reply built was tested during the launch of the Aston Martin DBX707, Aston Martin’s high-performance SUV. AML wanted to use data and Data Reply’s AI modeling skills to help dealers successfully reach customers who would want to buy the vehicle. “This was a great chance to build a focused model and see how it performed before applying it across the organization,” says Senso. “As an ultra-luxury brand, we have a specific customer base we’re targeting, and being able to identify which customers are most likely to purchase serves their needs and ours.”

Using the predictions supplied by the 360 model, AML was able to refine its approach mid-way through the launch campaign—thanks to better understanding of customers. The result was a 35 times higher conversion rate from prospect to purchaser. In just the first 6 months of operation, the solution allowed AML to improve its customer predictions by 85 percent. “For me, the thing that was most valuable is the level of engagement that we saw from dealer partners,” says Senso. “They really see value in it. It’s great to bring new people to the brand, but it’s also very important that we continue to provide a great experience to existing customers, and now we can. One dealer said to me that this was the best use of Salesforce data they had ever seen.”

The propensity-to-purchase model is being rolled out to all vehicles and will be incorporated into a new application for dealers to help them increase sales and cleanse data. Data Reply also engaged AML business and technical stakeholders in the AWS Data-Driven Everything Program (D2E)—in which AWS partners with a company to accelerate the journey towards becoming data-driven. The plan was developed as part of the D2E Mobilize workshop. It includes innovations using generative AI and business intelligence. “Data Reply has identified lifetime value and propensity to purchase, which is important, without losing sight of the customer experience,” says Senso. “We’re becoming more and more sophisticated. Data Reply is our data science partner and, enabled with technologies from AWS, has helped improve our global sales and marketing organization.”