Biarri Builds Optimization Engines on AWS to Maximize Customers’ Profits

In the digital, software-driven era, companies have more efficient, modern alternatives to spreadsheets for complex calculations. Businesses with revenues surpassing $1 million should no longer rely on multi-tab spreadsheets to make critical sales decisions, asserts Evan Shellshear, head of analytics at Biarri Group. Biarri started 12 years ago in Australia with a mission to help businesses make better decisions on resources and pricing. The group’s team of mathematicians and data scientists applies commercial mathematics to build optimization tools that deliver significant savings.

Biarri has built over 100 web-based optimization engines for its customers, including Australia Post. The group also works with small to medium-sized businesses, using its proprietary Workbench analysis framework to cost-effectively build cloud-based analytics tools tailored to customer requirements.

Biarri has been running its business using Amazon Web Services (AWS) since day one, selecting and staying with the platform for its reliability. Biarri reports 99.9999 percent uptime, which is integral to supporting the hundreds of complex queries its customers run every day on its tools. “The services and compute power AWS provides are like water and electricity; we expect them to work and switch on any time we need them. We’ve never had an issue with outages on AWS,” Shellshear says.

The core technology Biarri uses across solutions includes Amazon Relational Database Service (Amazon RDS) for streamlined database administration, and AWS CloudFormation for quick provisioning. It also relies on AWS Auto Scaling to maintain application performance when computing algorithms in parallel.

One of Biarri’s longest-running customers is Alliance Group, a New Zealand farming cooperative that is a leading supplier of lamb and other meats across the globe. Alliance was using a sophisticated yet overly complex spreadsheet to drive core sales and operations decisions around product allocation to customers and the cuts of meat that optimize revenue. Only one person held the keys to that spreadsheet, so allocation decisions were not transparent or understood across the sales teams or wider business. Sales optimization could take up to 24 hours because employees were located in different time zones, and there was serious risk of the spreadsheet being corrupted due to its complexity.

Alliance came to Biarri in 2018 seeking a nimble solution that could run on every salesperson’s desktop and deliver analysis results within minutes. Tim Hannan, sales transformation manager at Alliance Group, explains, “We’re dealing with large quantities, so even small changes in pricing and volume can have a multimillion-dollar impact on revenue.” The web-based optimization tool Biarri built for Alliance is completely dockerized, using Amazon Elastic Kubernetes Service (Amazon EKS) for orchestration. It uses Amazon ElastiCache for Redis for task query handling and Elastic Load Balancing to improve scalability. 

Alliance now uses the optimization engine Biarri built daily. “The tool offers excellent transparency across our organization, so we are no longer looking at different volumes and customer sales plans, cuts, and markets in isolation. We can adjust prices and instantly recognize the impact it’ll have on revenue,” Hannan says. To ease the transition out of spreadsheets, Biarri’s engineers designed a simple, user-friendly interface for Alliance. Despite its straightforward, clean layout, the tool offers a high level of granularization, including over 850 different sales scenarios.

The Alliance team has full confidence in the fairness of results yielded by the optimization tool. Every member of Alliance’s sales teams, both in New Zealand and at its satellite offices in the UK and Asia, have equal access to the live data. Historically, the organization had experienced some friction across teams regarding allocation decisions on its finite supply of product each season. Now, when a team member isn’t allocated the amount they requested, they can see that the price point they offered simply didn’t reach the required threshold.

“It’s not only taken away the friction but also sped up our decision-making process because everyone is seeing the same information and logic,” explains Hannan. The black-and-white results and accompanying logic the tool brings have also given Alliance’s sales teams leverage with customers. When negotiating prices, they can use reports from the optimization tool to demonstrate the value each cut of meat is demanding in the market. “The computer doesn’t lie. Having solid numbers to rely on has put us in a very strong starting position to make more optimal sales decisions,” adds Hannan.

Alliance continues to find new use cases to apply the optimization tool. Recently, when a customer asked for a further processed premium cut, Alliance could factor in labor requirements and resource constraints to determine a higher price point to justify the work required. The business stopped offering the cut until the market was willing to pay that price. The agility to shift gears regarding pricing and cut selection has proven invaluable during times of market uncertainty. In 2020, when disruption to supply chains halted orders from abroad, Alliance reacted quickly to adjust production and reroute products.

Shellshear says, “With the tools developed on AWS, our customers benefit from better planning and the ability to manage market volatility. They can forge a full-scale digital transformation with centralized control of their core business processes that plugs all employees into a digital engine that facilitates global collaboration and expansion.”

Word of mouth is the top source of referrals for Biarri, so keeping customers such as Alliance happy is a priority. Many businesses only realize the power of the optimization engines Biarri can build after experiencing them first-hand. “The customer has to succeed,” Shellshear emphasizes. “We try to simplify our solutions to get our customers to value as quickly as possible. On AWS we can solve exciting mathematical problems in an efficient manner and do the hard work so our customers don’t have to.”

 To learn more, visit aws.amazon.com/containers.