Heineken’s event-driven architecture key to becoming a connected brewer

Gary Flood Profile picture for user gflood August 4, 2022 Audio mode
Summary:
Twin combination of move to SAP HANA and use of Solace EDA seen as crucial for achieving Heineken’s EverGreen digital transformation ambitions

Image of green Heineken bottles on a production line in a factory
(Image sourced via Heineken )

Netherlands-headquartered brewing giant Heineken says real-time global connectivity and an event-driven architecture (EDA) areits best bet for better support of the 350 global and local beer and cider brands it sells in 190-plus countries.

The technology is also expected to be a major contributor to the brewing giant attaining the ambitious goals of its 2021 ‘EverGreen’ digital transformation drive.

Guus Groeneweg - Heineken’s Product Owner of Digital Integration, and who works as part of the company’s global IT organization at its head office in Amsterdam - says EverGreen centers on driving cost efficiencies across the business to enable it to become the world’s “best connected brewer”. Groeneweg explains: 

Becoming the best connected brewer means in everything - true application connectivity that leads to real-time analytics use cases. 

EDA is really helping us to realize that ambition, because events are about connectivity. For me, from an integration point of view, events are an important part of achieving our ambition of becoming a best connected brewer.

Improving an insufficient solution

To get there, there are several challenges to solve. For one thing, Heineken’s global IT team had to figure out a way to integrate the company’s many local business units, which are composed of multiple operating companies.

As many of these units were acquired, that means many still maintain the original business systems they operated pre-Heineken. Another challenge is that the data that Heineken wants to “better connect” comes in from a wide range of geographically dispersed sources.

The data also tends to come in peaks and troughs, with large surges connected to the launch of a new Heineken brand. That risks overwhelming older approaches to data integration platforms that rely on point-to-point communication via synchronous, REST-based APIs - which is the modern default for a lot of modern applications today.

As this could possibly result in duplicate or lost orders, lost revenue, and therefore unhappy customers, a better way of connecting lots of operational data at global scale was called for. Groeneweg says:

One of the main reasons we started looking at EDA was because our integration platforms were not able to handle some of the, let's say ‘blocky,’ moments where a lot of business messages get exchanged.

This eventually led the organization to consider ‘events’ - which to some extent turn away from using Internet technologies to more of a middleware-oriented way of working.

That’s because ‘event driven’ is a software design pattern in which loosely or even decoupled applications communicate by publishing and subscribing to events via an event ‘broker.’ 

In this model, an event is at its most basic just a notification of something that's happened or changed - a change in state of an application or system, like a new order getting processed or a customer’s debit card getting swiped, says Groeneweg.

These events will then get sent to a broker and onto any other system or application that's subscribing to that event. An appropriate response can then be triggered, such as alerting a technician to check a system in a factory.

In Heineken’s case - that’s a lot of events. Groeneweg says:

If a forklift truck driver in Singapore needs to get his picklist on time and so wants to know which truck to load, that's all based on messaging. All the 120,000 invoices we process globally per day, and which also must be processed on time, are too. And if a bar owner in Mexico or Brazil or New Zealand orders beer from us on their Salesforce app, that’s also a message that needs to arrive on time - otherwise they don't get their product on time.

25 million business process messages a month

Also a factor was the parallel IT decision - again, seen as an important basis for achieving EverGreen - to implement a new, global S4 HANA SAP ERP system across the whole group. Groeneweg explains: 

Doing that means we don't want to rely anymore on point-to-point integrations. Say a customer in Mexico changes his address: our invoice system needs to know that our ordering system will need to know that, a lot of systems want to know that.

In the past, we would have written one-to-one, point-to-point interfaces to all the systems that need this address update of that single Mexican customer. But with EDA, instead of building all those separate interfaces, we just use the integration pattern, publish this customer data records change once, and all the systems that are interested in knowing this just subscribe to the main system and will get their updates automatically.

In total, Groeneweg and his team need to ensure at least 25 million such business messages (events) get delivered per month which all need to be delivered on time and in full, otherwise Heineken gets incidents.

So, to achieve EverGreen targets, the company’s IT team has been using EDA for just over a year to better connect over 4,500 internal Heineken business-critical applications in areas like payments, logistics, inventory management, and more.

EDA is being supplied to Heineken by a vendor called Solace, whose ability for handling this volume of data and messages was key to its selection.

Twelve months into use, EDA has brought much better burst management, Groeneweg says, such as when customers across the world choose to pay at the same time. It has also started to smooth out production interruptions, which leads to better overall efficiency and customer service.

Once HANA is online in about six months, Groeneweg predicts even more efficiency, as use of EDA will be spread to more and more business workflows. He adds: 

EDA will be used in more geographies, and probably a couple of more types of flow - so not only customer material data, but also more transactional flows, like payments, logistical warehouse movements, purchase orders, maintenance orders, and preventive maintenance.

That’s where we want to be.

Loading
A grey colored placeholder image