End-User

AB InBev Anheuser-Busch InBev is a multinational drinks company and the world's largest brewer, based in Leuven, Belgium. The company has a global functional management office in New York City and regional headquarters in São Paulo, London, St. Louis, Mexico City, Bremen, Johannesburg, and others.

Problem

From a data point of view, AB InBev needed to migrate from a disjointed collection of breweries throughout Africa to a unified Zone where comparisons are made like-for-like across all sites.

Key drivers:

• The need for standardization in regions' routines and reporting.
• Improved efficiency at the plant level.
• Collecting data into a central location in a unified format for rapid SIC.

When looking at the data, one needs to know what really happened, what worked, and what didn't. With this knowledge, AB InBev hoped to create the ability to direct future decisions in a truly data-driven manner.

Pain Points

• Inefficiencies due to employees using different/disparate systems. 
• Data could be lost/corrupted easily.
• No single version of the truth because of disparate data sources.
• Poor standardization of KPIs across breweries. Some report on 10 KPIs, others on 40 KPIs. There was a definite need to streamline KPIs deemed critical to optimal plant operation.
• Poor standardization of KPI definitions. KPIs shared identical names but had different calculations.
• Very difficult to obtain accurate data for Zone HQ for analysis.
• Very slow transfer of data. Data was sent just once a month.
• Time (and resources) spent putting data together in a single format.

Solution

Phase 1: Flow to Regions (Tier 1)

• A Flow instance was created at each plant.
• Appropriate training was provided to each plant from a UI point of view.
• The use of Excel for reporting immediately decreased.

Problems Resolved:

• Employees are now entering data into the same tool onsite.
•  Data entry is automated wherever possible.
•  There is now an accurate audit trail.
• Operators are more efficient with fewer possibilities of duplicate data capture.
• There is a single source of the truth.

Phase 2: Flow as a Template

• Templated Flow instance created at Zone HQ – Template Server (TS).
• Standardized Metrics and Measures/KPIs are created on the template server.
• Department-specific reports, dashboards, and data entry forms were created on the template server.
• All brewery instances have a template server configured. These servers function to pull the needed templates down <where??>.

Problems Resolved:

• Standardization of KPIs across Africa. 
• Standardized KPI. Comparing like-for-like across plants helps focus attention where it is needed. 

Phase 3 – Regional to Zone Replication (Tier 2)

• A Flow instance was created at Zone HQ for data replication and reporting called the Africa Report Server (RS).
• Bulk Replication Configuration was done at each site (100 000 measures).
• Each site posts data up to a corresponding measure within the RS.
• The data posting is triggered when the source data changes.
• The Tier 1 server is responsible for replication to the Tier 2 server.

Problems Resolved:

• Data is replicated to Zone HQ within minutes of changing at any brewery.
• Analysis is rapid, and upper-level management can quickly make the best decisions.
• Data seen at Tier 2 is always the latest version of the truth. If there is a change at the plant, it will be updated to the Tier 2 server.
• Data comes through in the same coherent format into dashboards that have been configured in advance. 
• Middle management no longer needs to spend hours collating data.
  

Results

The Future – Big Data

Data Value through Visualization

• AB InBev has a central Flow server with plenty of data they can utilize.
• AB InBev is looking at using graphical tools within Flow more.
• Flow admittedly does not seek to be a Power BI or QlikView. Flow provides the mechanism to provide its data via data consumers to other visualization systems, so one can use those tools' visualization capabilities while leveraging Flow's data acquisition and transformation strengths.

Data to Cloud

• AB InBev now has a standardized way of posting data to the cloud via MQTT consumer.
• The company no longer relies on Excel dumps, SQL DB replication, etc.
• The data is of a high enough quality to be used by machine learning solutions specifically focused on using data effectively.
• "We are busy interfacing with our global structures about getting data globally into a Sustainable Development algorithm."
• Reverse Osmosis/Water Treatment companies that are seeking to use the data to decrease brewery maintenance costs.

Project Info

• 29 sites
• 4150 events (total)
• 542 500 measures (total)
• Connected to different data sources:
  - Microsoft SQL Database
  - AVEVA Historian
  - Metering Online
  - OPCUA Historian
  - Web Service
• Reports that have been generated so far:
  - 1680 charts (template chart instances across all sites)
  - 721 charts (Zone/HQ-specific)
  - 285 dashboards
• Flow is used to integrate with third party applications including Flow Consumer, Microsoft SQL Database and PostgreSQL Database.
• Flow tiering, an enterprise decision-making solution, is used extensively at AB InBev, where approximately 5500 measures were replicated from each of the 29 sites.
• Central template server from where all sites pull their Flow templates. Sites use Flow templates to ensure standardization and governance.
• More than 500 people across all the 29 sites are using Flow reports and dashboards daily. ‍
  

“When I initially arrived at SAB, my concept of data transmission was an excel file on a USB flash drive. That has changed drastically with the incredible need for easily accessible, accurate and real-time data. Flow has really become a standard tool within AB InBev Africa - all our breweries use Flow for real time information and reporting.” Rowan Ray, Tech Supply Specialist @ AB InBev.