The data economy is making new demands of Australia’s, and indeed the world’s, business models, with rapid advances and availability of technology redefining markets and reshaping industries.
AI is already helping business owners better understand their customers, supply chains and emerging opportunities. There are numerous companies (large and small) using predictive analytics and machine learning to analyse their data.
One company that applies AI-powered solutions to help retailers generate a return on retail space investment is HIVERY, a big-data start-up using cutting-edge AI tech to inform product recommendations, price learning, space rationalisation, and promotions.
Founded in 2015 and spun out of CSIRO’s Data61 the same year, HIVERY uses cloud-based AI-powered solution based on machine learning to gather and convert data findings into easy-to-action measures.
From small experiments with dozens of vending machines in Newcastle Australia, to counting Coca-Cola as a client and exporting five products globally, HIVERY have expertly leveraged the power of AI and machine learning to grow and expand their business.
How it works
Retailers need to know which products to stock, how much of that product to stock, and how to price/promote it all in the most cost-effective way in different locations and markets.
“Take Coca-Cola - they have 500 brands and around 3,000 beverage types or SKUs (stock-keeping unit, most often a scannable barcode printed on product labels),” explains HIVERY COO and co-founder, Franki Chamaki.
“Coke in a can, Coke in plastic bottles and even different variations of Coke: Coke Diet, Coke Cherry and not to mention Coke in 200ML, 250ML, 1.5 Litre, etc. How do you determine what SKU should be placed in a specific outlet? What is the best?
Optimal space is just a derivative of the right assortment and the right price/promotion. In order to have optimal space, you need to work out what optimal SKU/assortment and as well as price/promotion. This is what we solve.”
The role of AI and Machine Learning
Self-learning algorithmic models that use AI and machine learning provide the mechanics behind HIVERY’s unique product ranking method, while the operations research element (the application of advanced analytical methods to help make better decisions) determines how much of the recommended product should be supplied.
This process also simultaneously incorporates the specific requirements and rules distinctive to individual businesses.
SAM, HIVERY’s Space Assortment Manager is the world’s first AI category management application. It can develop a planogram (a visual representation of a store’s products or services on display) in several minutes - a process that would normally require 150 steps, require 15+ staff members (a visual merchandiser, category manager, space and assortment planners, etc, working together to design the planogram) and take six months.
The same methodology can be applied to vending machines. “It takes a vending planner around 45 minutes to optimise one vending machine, and that's just space, not assortment,” says Chamaki.
“With HIVERY’s solution, this is done seconds and we are able to optimise for both space and assortment while incorporating merchandising rules all simultaneously in one click of a button. Literally one button.”
AI can be used in any business and any industry, says Chamaki. “It has the potential to create new competitive advantages provided the goals are well defined.
AI is a general-purpose technology, meaning it can be leveraged by any entity, just like the internet or electricity before.”
The Results
In four years, HIVERY has grown their product suite from one to five, have four international offices (Sydney, Japan, USA, and Mexico), and count The Coca-Cola Company, Walmart, Red Bull, Diageo and TRIAL Co as clients.
They recently won the Export Council of Australia’s Technology and Innovation Award, Deloitte's Top 500 Technology Award (2018), Business Chamber’s Excellence Export (2018), and are representing New South Wales in the 57th Australia Export Awards.
Over the course of the next four years, HIVERY plans to continue scaling their AI and operational research products, with an outlook to generate US$20 million in recurring revenue.
CSIRO’s Data61
Chamaki credits CSIRO’s Data61 with a critical role in fostering HIVERY, lending specialised knowledge in AI and skills to develop unique intellectual property, acquire talent, and access grants.
“Our recent product in the trade promotion space was actually co-developed with Data61’s research assistance. Here, we needed to demonstrate to our client that we could generate a more profitable promotional schedule, observing a number of constraints.
A Data61 researcher worked with us in developing a unique prediction model to produce an optimised promotional schedule. The promotional schedule generated results that are 40 percent better than the current business method,” explains Chamaki.
“Given over 80 percent revenue of most consumer-packaged foods revenue comes for trade channels like supermarkets, any improvement can generate significant commercial benefits and competitive advantages.”
AI, ML and data science revolutionising industry
According to Chamaki, AI technology has the ability to reshape and optimise not just Australia’s retail industry, but industry in general. “It’s a general-purpose technology, hence it will impact all and every aspect of personal life, society and industry just as electricity or the internet did.”
“If you’re not leveraging or starting to think how to leverage AI through experimentation, rest assured that your competitors will be. Worse, your whole industry and you will be left behind, so don't be a dinosaur," - COO of HIVERY, Franki Chamaki