Smart People: Machine learning – Winning the Counterfeit

Niall Murphy, founder and CEO of Evrything

Niall Murphy is a man on a mission. “Machine learning is changing how consumer product brands fght back against the $1.2-plus trillion in revenues lost to counterfeit and supply chain integrity issues each year,” he explains – and he wants to be in the thick of the fray.

Born in South Africa, Murphy founded Evrything to provide a digital identity for billions of consumer products on the web. This wasn’t his first start-up venture: he cofounded a pan-European Wi-Fi network, called The Cloud, with technology entrepreneur George Polk, ultimately selling it to BSkyB in 2010. He also founded one of the first ISPs in South Africa in a joint venture with Sprint and three digital media businesses. Long before that, in 1994, he teamed with Daniel Erasmus, a leading futurist, to establish the Digital Thinking Network, billed as a “scenario thinking consultancy.” In 2018, the first app from Evrything was launched at Carrefour supermarkets in Spain. Called Reciclaya, it allows customers to determine how to recycle individual products by simply scanning the barcodes on supermarket receipts. The app separates the products into virtual recycling bins so that customers can deposit them in the closest corresponding containers. Consumers can earn rewards by registering where they are recycling by scanning or tapping identity tags on recycling units.

Digital identity systems generate real-time data and capture everything that happens to every product throughout its entire life cycle

Niall Murphy

Winning the Counterfeit: Evrything CEO Niall Murphy - machine learning


“For the first time, consumers can receive detailed information on how to recycle all products and packaging using a single app – and be rewarded for their participation,” Murphy says. “It’s convenient and simple and, through a gamified experience, it drives a new style of behavior.” Murphy has his sights set on bigger game: counterfeiters. “Counterfeit products and gray-market imports prey on the quality, goodwill, and trust that brands have spent years and billions of dollars building, sucking profits and impacting consumer trust,” he maintains.

The problem is not a new one: “A running battle has been fought for decades between brands and fraudsters,” Murphy says, moving on to describe an ongoing “arms race” where brands invest in special labeling designs, invisible inks, and complex packaging materials in an attempt to outwit the fraudsters. “The good news is that software intelligence and the mass-scale digitization of the world’s consumer products is coming to the rescue.”
Counterfeit products and gray-market imports thrive because of a lack of visibility into the supply chain, Murphy believes. If brands knew where every single product item was as it moves through the supply chain – where it was made, how it got there, when it reached the retailer and, ultimately, when it was purchased by the customer – brands would know if the final product is genuine and if it is in the right market. “Achieving this kind of visibility hasn’t been possible until now,” he says. His answer is the application of digital identity technology in the cloud:

  • Every product can have a web address through a global upgrade to barcodes enabled by the advent of the GS1 Digital Link standard.
  • Brands have the capability to gather logistics data through billions of smartphones already equipped to scan barcodes.
  • Mass-scale, crowd-sourced data from consumers allows brands to continually redefine and grow revenue, based upon real-time intelligence.
  • Cost-effective, real-time data collection throughout each product’s supply-chain journey can be achieved.

“Digital identity systems generate real-time data and capture everything that happens to each individual product throughout its entire life cycle, from manufacturing to recycling,” Murphy explains. By creating a dynamic digital ecosystem around the world’s physical products, brands have an opportunity to change the integrity-management game.

The challenge, until now, has been the cost of detection. Previously, the only way brands could track products and protect brand integrity through the supply chain required cost-prohibitive hardware and a lot of people – teams of brand protection experts – to identify problems in the market.

Machine Learning and algorithms

The arrival of software algorithms with predictive intelligence can help brands not only to protect against but also to stay a step ahead of counterfeiters at scale. Data gathered from products as they’re manufactured, distributed, retailed, and consumed is aggregated in the cloud, while machine learning algorithms, trained to look for patterns, can apply this data to identify suspect events. With this intelligence, companies can take a proactive stance against brand integrity issues.

Real-time gathering of data and giving each product a unique digital identity dramatically increases the amount of data that needs to be processed. The best way to do this, Murphy feels, is through machine learning (ML), where a machine programs itself by learning from data, finds answers in the data collected, learns patterns of what is normal and what is not, and learns to identify when things happen outside of these norms.

Cheap and abundant processing power, big data, and improved algorithms have all contributed to the practical application of machine learning and now it can be applied with product data at massive scale, he says. Unlike traditional software programming, which is limited to the rules and vision of the software coders, machine learning reduces the programming time for problems involving a complex network of rules and offers the ability to attack seemingly “unprogrammable” tasks that go far beyond the human brain’s capacity.

“If a brand inspector were to be present at the manufacture of each product item and then accompany it throughout its life cycle from distribution to recycling, problems of illegitimate production, diversion in the channel, and counterfeits at the point of retail would be eliminated,” Murphy claims. Clearly that would be delivered at absurd cost, not to mention the additional carbon footprint of the airlines involved. That said, product digitization coupled with machine learning can do the same job as a human brand inspector – but much faster and better. Each unique digital identity in the cloud accompanies every product on its journey and a machine-intelligence brain scrutinizes each step along the way. This provides unprecedented data analytics and real-time traceability.

Product identify with machine learning

If a product is headed toward the wrong channel of distribution, it is detected. If a product identity appears in the wrong market, or the pattern of events surrounding a product is wrong, it is detected. Every consumer engagement becomes a data point to support integrity enforcement in the supply chain and, because every product item is uniquely digitally identified, the source of any suspected problem is rapidly identified.

According to the Organisation for Economic Co-operation and Development (OECD), 2.5 percent of global imports are counterfeit – with US, Italian, French, and Swiss brands being most affected. According to Murphy, protection is now available to any brand, consumer, or retailer by using a regular smartphone and industry-standard product codes. “For companies losing tens of millions of dollars every year, software is changing everything,” he concludes.

The Smart People 1/2019


Leave a Reply

Your email address will not be published. Required fields are marked *