The power of prediction: how AholdDelhaize is using data and AI to provide more personalised shopping experiences

Ahold Delhaize is one of the world’s largest food retail groups. Albert Heijn, an Ahold Delhaize brand, is the leading food retailer in the Netherlands and also has stores in Belgium and Germany.

Can an algorithm help you bake a cake? For customers of Dutch grocery retailer Albert Heijn, the answer is yes. The company, part of global conglomerate Ahold Delhaize, is harnessing Microsoft AI technology to create rich, ‘hyper personalised’ shopping experiences. As a result, it can bring convenience and innovation to customers online and in store while using data-driven decision-making to ensure shelves are fully stocked.

As the largest Dutch supermarket chain, with almost 1,000 stores employing 100,000 staff, selling great food has been at the heart of Albert Heijn’s business for over 130 years. The business continues to evolve to meet changing shopper preferences, and part of that evolution is harnessing the power of the cloud and big data through an omni-channel retail approach—a business transformation enabled by Microsoft Azure.

Albert Heijn first store purmerend

Unleashing the power of big data

Like many across the retail sector, Albert Heijn knew its business was changing rapidly, but its legacy technology didn’t allow it to evolve quickly enough to meet shifting consumer preferences. The company had multiple on-premises data warehouses and a huge landscape of applications, making it difficult to get data into the hands of internal business owners, including analysts and data scientists. It was therefore crucial that a central, cloud-based location for current and historical data was made available, to provide better synergies for data users across the business and a platform for delivering new services through data science. The openness of the Azure platform enabled Albert Heijn to easily adopt all new technology developments, including open source tools like those running on Linux, while unlocking the opportunity to do more with data.

We decided to move to Platform as a Service in the cloud. “It was important that the platform was as open as possible to adopt new and emerging technologies, such as machine learning and AI. For us, that platform was Microsoft Azure.

– Caroline van Zeggeren, Director, IT Development BI & Data Analysis

In order to better serve customers and optimise internal processes, Albert Heijn put together a shopping list of its; setting out the goals it wanted to achieve by moving its data to the cloud: To make use of big data to create seamless shopping experiences that provide customers with the right product, price and promotion, via the right channel.

• To provide a platform to drive fact-based decision-making by providing better access to high-quality data from internal and external sources.
• To empower the company’s internal group of data scientists to develop algorithms that deliver new, highly personal and relevant shopping experiences.
• To optimise internal processes, reduce complexity and be more cost-effective.
• To ensure data is protected and compliant with national and global data regulations, such as GDPR.

From hindsight to foresight

The migration of Albert Heijn’s data platform to Microsoft Azure Data Lake had an immediate business impact, enabling data engineers, data scientists and analysts to collaborate better and speed the path of data to those who need it. As a result, it is now easier for business owners to retrieve their own data sets and quickly create custom dashboards. These new capabilities unlock the value of existing internal data and enable it to be combined with external sources to produce insight-based recommendations for the business. A reduction in the support needed from the IT team now means teams can develop their own data experiments at speed, breeding a testing culture that enables the company to quickly create new solutions and achieve competitive advantage.

Caroline van Zeggeren sees this as a crucial component of the transformation. “We can provide data sets our people can trust, support and which are available,” she says. “Users can extend and connect external data sources such as survey or market share information, and they can do that themselves, instead of depending on IT.”

This speed and improved access to data has meant that rapid progress is already being made within the business, with enhanced insight enabling better decision-making and exciting opportunities for the future. “We’ve gone from knowing something happened, to understanding why it happened,” says Mike Blay, Vice President of IT Development. “The next stage is being able to predict what’s going to happen next. From hindsight to foresight.”

Predict My List: hyper-personalisation at scale

As a result of the migration, Albert Heijn is harnessing data in new ways to provide a better service to consumers and, in turn, boost sales. Predict My List is one such solution; a tool that populates customers’ online shopping lists. Using insights from past shopping lists combined with data such as location, local events and seasonality, it delivers recommendations for alternative or forgotten grocery items, available promotions and more.

Thanks to Microsoft Azure’s big data capabilities, in conjunction with Albert Heijn’s analysts and data scientists, the company can create a rich view of its customers, providing guidance and the right information at the right time through hyper-personalised shopping experiences.

Our primary goal is to focus on the customer agenda and provide a best-in-class experience, from when they shop online to entering a store. We want to be useful and help our customers save time and money.

-Mike Blay, Vice President of IT Development

So how does all this personalised data make a difference to customers’ day-to-day lives? Imagine you’re baking a special cake for a gathering. You’ve created an online shopping list with the ingredients, along with your weekly groceries. You’re about to check out and submit your order for delivery, when a notification pops up, letting you know you don’t have butter on your list a key ingredient in any cake! Along with the notification, you see a special offer for the butter you normally buy. It’s added-value moments like these, based on data and AI, that enable the company to delight its customers, making their lives easier rather than selling to them in the traditional sense.

Albeit still at an early stage, Albert Heijn’s ‘customer experience first’approach is paying off, with data from Microsoft Azure enabling the Predict My List solution to deliver big results. Providing cross-sell recommendations to thousands of customers every day, sales are already increasing, with more products, on average, added to baskets per shop.

And the added value provided by a more personalised experience isn’t lost on customers either… At least, for the most part. Early consumer feedback on Predict My List has been glowing, with many lauding the convenience of the tool and the help it provides.

However, the high quality of the personalisation has also aided those unaware they are being helped. “A lot of people don’t realise that this is AI,” says Blay. “Someone even thought they’d filled out their shopping list and forgotten about it. They didn’t realise it was an algorithm that did it for them.”

A frictionless future for shoppers

And this is just the beginning. From this solid, cloud-based foundation, Albert Heijn is embarking on a transformation journey that will continue to evolve, benefitting the business and its customers whether they shop online or in store.

It is currently testing a new concept in ‘frictionless shopping’ called ‘Tap to go’ in four of its AH to go convenience stores. This smart technology will allow customers to tap their AH to go card or an NFC-enabled mobile app on the products they wish to purchase, and leave without having to stand in line to pay. Customer-focused innovations like these, as well as Predict My List, are fuelled by feedback; asking shoppers what works best in order to adapt and improve the experience as part of a ‘digital feedback loop’.

“We’re continuing to build on our hyper-personalisation,” says van Zeggeren. “That’s an ongoing journey. Replenishment will also be an ongoing process, as we optimise every element in our supply chain.”

Data is increasingly playing its part across Albert Heijn’s stores and supply chain, in addition to improving the online shopping experience. Enhancing the process of making stock management and replenishment decisions, the company is now using big data, AI and machine learning to build algorithms that provide improved visibility, avoiding empty shelves by way of a next generation replenishment solution in the cloud.

Azure gives us insights within the operations that we’ve not seen before. We’re using dimensions of data to see more details about inventory: where we need stock, how much, where it is in our digital supply chain network. This data helps us optimise our processes; getting inventory where it needs to be, more quickly and efficiently.

– Mike Blay, Vice President of IT Development

Combining shopper data with supply chain information and instore knowledge, the company plans to extend its customer-centred personalisation, adding new data dimensions to create even more rich, useful services.

“We’re using real-time dataflows with specific store layout information to deliver new experiences—something we couldn’t imagine two years ago,” says Blay. With this spirit of ongoing innovation already changing the way the company operates and how customers interact with, and shop for, their food, the possibilities for the future are endless. As a result, parent company Ahold Delhaize is exploring more instances of the Microsoft Azure platform across the business, applying the Albert Heijn model to additional brands, enabling algorithms to be co-developed.

Today, taking advantage of the agility and scalability the cloud affords, alongside the power of data and AI, Albert Heijn is prepared for the ever-changing future of retail and looks set to carry on charming its customers—one cake recipe at a time.