Asia's Source for Enterprise Network Knowledge

Wednesday, May 22nd, 2019


Brewing up the best consumer experience

ZTE introduces LoRa-based smart meters

Anheuser-Busch InBev SA/NV (AB InBev)
is a Belgian publicly traded transnational beverage and brewing company, with a heritage that dates back more than 600 years, spanning continents and generations. It is considered one of the largest fast-moving consumer goods (FMCG) companies in the world. AB InBev has a diverse portfolio of well over 500 beer brands, including Budweiser, Corona, Stella Artois, Beck’s, Hoegaarden and Leffe.

Integrating systems and data from acquired companies

When companies grow via external acquisitions, integrating the systems and data from acquired companies is always a challenge. For AB InBev, that challenge included a hybrid environment with both on-premises and cloud systems such as Salesforce, 15 SAP instances, 27 ERP systems, 23 ETL tools, and a host of brewers operating as independent entities with their own internal systems.“As you can imagine, that made it extremely difficult to obtain

a single, unified view of our business, and there was no single source of truth,” says Harinder Singh, Global Director of Data Strategy & Solution Architecture at AB InBev. “Also, we’re operating on six continents and we needed to become GDPR-compliant, and that required global visibility into all our data assets.”

Because executives at the company couldn’t readily get access to the enterprise-wide information necessary for strategic decisions, Singh says he knew AB InBev had to take a different approach.“We recognized that we needed one central repository for our data assets,” he says. “Our internal customers— like data scientists, operations teams and business teams—were struggling to pull together data from over 100 source systems, analyze it, and make timely decisions on product development, supply chains, marketing campaigns and more.”

Also, like other alcoholic beverage producers, AB InBev must abide by strict regulations regarding gathering consumer information. “So, we collect external data, such as geographical data and purchasing trends,” says Singh. “But we still need to standardize and integrate that data, which was another aspect of our data challenge.”

Why Talend?

As part of its process to evaluate potential solutions, AB InBev conducted a months- long proof of concept (POC) with Talend. “We knew we wanted to embark on a cloud journey, and Talend was built in that world, enabling cloud and on-premises systems to talk to each other in a secure manner,” says Singh. “During the POC, Talend performed the way we wanted and expected it to, and it was clearly the best product for our need. In addition, Talend pricing was more competitive than that of other options.”

At the moment, says Singh, AB InBev is primarily using Talend Data Integration to bring together the disparate data sources inside and outside the company, and using Talend Data Preparation for data discovery. “We’re a multiple company, that is, all our data management work has to be done for multiple companies under the AB InBev umbrella, so we’ve built a reusable framework using Talend,” says Singh. “That way, when we pull data from one brewery, we can reuse the code we created and save a great deal of time.”

In the architecture AB InBev has built, Talend extracts data from a range of sources—real- time and batch, cloud and on-premises, ERP systems, data from IoT devices—and stores it in a landing zone, which is part of
a data lake, or data hub, that resides in the cloud on Microsoft Azure. That data is then processed and archived before going into a golden layer, from where it’s consumed by AB InBev internal users. Talend also powers a sandbox for data scientists, which they use to test various data models. In addition, the AB InBev architecture includes Hortonworks for Hadoop, and related technologies such as Hive, Spark, Hbase and Kafka.

“Among the biggest benefits of the new
IT architecture are simplification of the infrastructure, and reusability of processes to rapidly extract and provide access to data,” says Singh.“Because we have reusable code, what used to take us six months now takes us six weeks. That translates into faster decisions and reduced time to market for decisions, campaigns, products and more.”

Singh cites cost savings as another major benefit. “Now, instead of paying for and managing 23 different ETL tools, we’re moving towards managing only one by standardizing on Talend.”