We are committed to helping organizations everywhere stay connected and productive. Read more

JJ Food Service
ING Direct


Shopping carts that already list what you’re going to order? Now that’s customer service. And it’s the service that JJ Food Service customers get every time they place their orders online or by phone. This is the power of Microsoft Azure Machine Learning and Microsoft Dynamics AX.

Operating a food delivery service is like running a 24/7 relay race. Every morning, orders start rolling in. The logistics team takes over to route and sequence orders, and passes them to the warehouse for filling and loading overnight. The next morning, delivery drivers step in—just as customers begin placing orders for the next day.

JJ Food Service knows the drill. As one of the UK’s largest independent food delivery services, it provides 60.000 customers with everything they need to run a food-service business—from fresh, frozen, and dry food to items like paper products and cleaning products. And the company runs like a well-oiled machine.

But that wasn’t enough for Mushtaque Ahmed, Chief Operating Officer for JJ Food Service. He wanted to do more. Like anticipate customer orders, recommend additional products, and, above all, cut the time it takes customers to place orders, so they can spend more time on their food-service businesses.

The constant quest for improvement

JJ Food Service has always been at the forefront of using technology to streamline its processes. “We have lots of competitors,” Ahmed explains. “You can’t raise pricing because customers will just go elsewhere. The only way to stay successful is to make ourselves more relevant to customers and streamline operations on the back end.”

The company has done an enormous amount of work to operate more efficiently—no mean feat for a company that services 60,000 customers with more than 4,500 products from eight central warehouses. It started by implementing Microsoft Dynamics ERP planning and Microsoft Dynamics CRM solutions in 2004. Over the years, it has refined its operations, and Microsoft Dynamics AX now powers the entire operation—from HR, procurement, and sales to inventory/warehouse management and order processing.

But Ahmed recognized another benefit: JJ Food Service was sitting on a rich vein of customer data just waiting to be mined—and which could boost its customer service even further. “You have to keep asking yourself what is possible, what we can do next,” he says. “You can’t just stay the same, because then the competition catches up.”

The company started looking at data sets for business intelligence to see if it could improve customer service. However, there was so much of it that the costs of staffing the project was huge. And that’s when Microsoft Azure Machine Learning (ML) entered the picture.

Amazing customers with predictive shopping lists

Many companies want customers to spend a lot of time engaging with them. Not JJ Food Service. Its success hinges on making it fast and easy for customers to get what they need—and spend the extra time in the kitchen. That’s why it started wondering about anticipating customer orders, rather than just being ready to place them.

But anticipating orders was easier said than done—until Azure ML. The biggest challenge was how widely customers varied in type and size, and what they purchased in each order. For instance, a restaurant may order salad greens every day, flour every two weeks, and cooking oil once a month. “To be successful, we needed to be relevant for that week, that day, that exact point in time,” Ahmed explains. “If we know that a customer has ordered coffee on the first day of the month, we can add it to the order for the first day of the next month.”

JJ Food Service realized that Azure ML could help resolve these issues, and started working with the Microsoft Azure team.

The first thing the company did was to write code for its website to capture customer behavior. Next, it provided three years of transactional data to train the data model. Finally, it integrated the Azure ML data and recommendations into its call center environment, so that call center reps see exactly the same thing that customers see on the website. That means online and phone customers get the same great service.

The process took only about three months to implement, and it immediately surfaced key customer data within the existing Dynamics AX infrastructure. Today, when customers log on or call in, the system surfaces predictions based on analysis of when previous purchases were made. And it fills out the order pad automatically.

The result? Customers spend less time shopping, which is exactly what Ahmed wanted. “With Azure Machine Learning, the wow factor is huge. Customers are amazed that we can predict so accurately what they need.”

Surfacing recommendations for a more personal touch

When customers log on, they see not only a predictive shopping list but also recommendations for things they might like to order. Let’s say a fish and chips shop orders batter; the system will also ask whether it needs particular spices. As Ahmed explains, these are two separate products that work better together.

Even better, the system reviews the total order right before checkout to see if the combination of items indicates a possible need for other products. For example, a fast-food restaurant orders meat, poultry, vegetables, and some drinks. Does it also need cooking oil? Is its supply of paper cups running low?

It’s an approach that works. JJ Food Service estimates that these recommendations currently make up about 5 percent of the total shopping cart. This number may not seem large—and Ahmed expects to see it go down as the system gets smarter and more able to predict orders accurately—but when you consider the company’s size, it can really add up. Plus it adds a personal touch for customers.

Radically improving cross-selling

Originally, JJ Food Service took orders in its call center. (Remember, the company is 25 years old!) Over time, about half of its customers switched to ordering online—and in the process, the company lost the ability to cross-sell. “We needed a way to capture the impulse to buy online,” says Ahmed. “Customers that weren’t talking to the call center weren’t being asked key questions about products they might need to purchase.”

But by using Azure ML, JJ Food Service can now make recommendations to customers ordering a particular item. This feature is vital to promote new products or bring customers’ attention to products that they currently go elsewhere to purchase.

Targeting new customers more effectively

An established company that continues to grow and expand, JJ Food Service knows there’s no better way to capture business than by making itself indispensable to new customers the moment they log on.

Using Azure ML, the system displays products that similar companies have purchased—a strategy that works well for many online businesses. “We can surface data from previous customers or similar customer communities,” says Ahmed. “We can show them immediately what products are important to them.

And it’s just the beginning …

By surfacing Azure ML data within Dynamics AX, JJ Food Service has simultaneously increased sales and customer satisfaction, and will—over time—streamline back-end operations even more. For example, knowing what products will be in demand can help the company stock warehouses more efficiently.

For Ahmed, however, it’s just the start of the journey. He’s already looking around the next corner to see what else is possible. For instance, the company is exploring how to use Azure ML to power promotional activities and create product launches targeted by company type.

“I don’t know of another food delivery service company that’s doing what we are. Microsoft technology has played a key role,” Ahmed concludes. “Microsoft Dynamics AX works hard for us, automating processes. But we also need to make these processes intelligent—and that’s where Microsoft Azure Machine Learning is vital.”