Beijing, China

I’m at Walmart’s global food safety collaboration center, housed in a nondescript office tower downtown, on a rainy Beijing summer day. I meet with Frank Yiannas, VP of food safety at Walmart. He’s fresh off a flight from Bentonville, Arkansas—Walmart’s headquarters—and is in Beijing for a quarterly meeting of the board that oversees the center. He’s keen to underline China’s importance in the world’s future food supply landscape. “It’s such an important and expanding economy in the global 21st century food system,” he says. “The 21st-century food system—China will play a huge role in it.”

And Yiannas is right. Chinese take their food safety seriously. After dozens of scares involving “fake” food—from eggs to infant formula—Chinese consumers place a premium on food products that come from what are perceived to be trusted sources. That’s why there’s a roaring trade in baby milk powder, from Hong Kong and Australia, on the mainland. It’s so big that the Chinese government changed tax rules in 2016 to clamp down on the grey market for imported baby milk formula.

That’s partly why Yiannas chose to launch Walmart’s blockchain ambitions in Beijing, with a pilot project that tracked packages of pork—China’s most popular meat—from farm to supermarket shelf. The pilot was successful enough to trigger a second trial, this time with mangoes in central America, which was successful as well.

We’re talking in the food safety collaboration center. Instead of a lab filled with scientists testing meat or vegetables for toxins, it’s really just a series of meeting rooms, much like any co-working space across the Chinese capital. “We certainly could have built a lab,” Yiannas tells me, but Walmart was less interested in using the center to focus on the science of food safety, and more on the links in the chain that get food from place to place. “Some companies will do food-safety work in a silo,” he says. “We reach out to the suppliers, the consumer groups.”

The Beijing center, the first of its kind for Walmart, opened in 2016. The retail giant committed $25 million over five years to the facility after a number of food safety scandals—from mislabeling of organic pork to selling expired duck meat—plagued the country. Walmart’s goal is to convene the world’s food safety experts to figure out ways to track the provenance of a package of food—be it mangoes or pork ribs—with new techniques and systems.

One of Walmart’s grandest projects is an attempt to graft a blockchain, that immutable cryptographic ledger first used by bitcoin, onto the world’s complex food supply chains. Walmart has roped in some of the industry’s biggest players, among them fruit producer Dole, consumer goods giant Unilever, and Swiss water and food conglomerate Nestle, to form a consortium of 10 food producers and retailers to make it a reality. They’re building the technology with IBM, which has been among the most active technology firms pushing blockchain solutions to corporate technology departments. If Walmart is successful, the project could fundamentally alter the way information is secured, stored, and shared across the food and retail industry, ushering in an era where an item of produce can be tracked in real-time from farm to table, by producers and consumers.

The problem with supply chains

To hear logistics experts describe the links of warehouses, container ships, trucks, and cargo planes that deliver our food to us, it’s a wonder anything makes it onto our shelves at all. It’s a mess of phone calls, hand-written forms, and, yes, the much maligned fax machine. Retailers and food suppliers can spend millions of man-hours a year on food traceability, says Yiannas.

Food supply chains are complex affairs that have only become more intricate over the last decades. “The fundamental problem with supply chains is that information is captured in silos,” says Steve Rogers, an IBM supply chain expert. “The globalization of supply chains has just made that worse over the last 15 to 20 years.”

Collaboration—or the lack of it—is the reason why current food safety systems take so long to trace a product’s provenance. Yiannas describes a landscape for food safety systems that’s fragmented, with each producer or retailer using its own systems to track products. The problem begins when an item leaves one producer’s system and is entered into another’s, with no traceability between the two systems. “It’s very linear data capture. It’s siloed in Walmart’s system, or another retailer’s system, and there’s no connection,” he says. “With blockchain, it’s different. It’s fundamentally about networks.”

Rogers describes the complexities of even a basic food supply chain: “Think of a farm that may be selling to three or four different wholesale buyers. They may then be distributing to multiple different packagers, who eventually package it for different brands. And then another set of logistics is going on, and eventually it gets to a store.”

In January 2016, a listeria contamination caused at least one death and dozens of hospitalizations in the US. This came 10 years after an e. coli outbreak from contaminated spinach, which infected 199 people. The first cases of the e. coli outbreak were reported in mid-August, but it was only in early October that the Centers for Disease Control issued a warning. Rogers says greater supply-chain transparency would have tightened the “containment ring” and minimized infections. “When you don’t have the information, your containment ring becomes extremely large and costly. With knowledge, you can draw a much tighter containment ring and hopefully you can isolate it quickly,” he says.

Rogers and Yiannas are betting on blockchains to do this. But it didn’t come easy for these veterans of the logistics and food businesses. “I was a naysayer, I was so skeptical,” Yiannas says of the time IBM first gave a presentation about the technology to him. “I became a believer when I started to understand how blockchain worked,” he says. “It solves not only digitizing food information but it addresses the social issue of how that information is shared.”

The “collaboration” bit of Walmart’s food safety center in Beijing also goes some way to explaining why a blockchain is a good fit for it. Blockchain tech is designed to solve the problem of coordinating members of a group who don’t necessarily trust each other. In the case of bitcoin, it allows anonymous strangers to send digital money to one another while eliminating the possibility of one party cheating another by spending funds that they don’t have, for example. That principle has been seized on by other industries, from finance to shipping, as a way to solve coordination and issues of trust globally.

There are trials from the shipping line Maersk and work by Chinese e-commerce giant JD.com. The latter has a blockchain, live since May of 2017, tracking the delivery of frozen beef from the steppes of Inner Mongolia to supermarket shelves in China’s megacities of Beijing, Shanghai, and Guangzhou. Fellow e-commerce giant Alibaba has announced plans to do something similar, but with beef products from Australia.

“Companies have been digitizing information on the supply chain for two decades,” he says. “That’s not new. What we didn’t believe—but what we saw—is that the fundamentals of how blockchain works are different.” He talked about the glut of horse meat fraudulently mixed into processed meat products across Europe in recent years. “[Blockchain] is like shining a light on each point of the food system that could deter those types of incidents.”

What’s a blockchain, anyway?

At this stage, some explanation of what exactly a blockchain is is in order. First of all, there is no single, monolithic blockchain. Different cryptocurrencies, such as bitcoin or ethereum, have separate blockchains. When people refer to “the blockchain” they may be referring to the oldest and most popular one, which is the bitcoin blockchain, or using it as a generic term for blockchains in general.

In the case of a cryptocurrency like bitcoin, it might be helpful to think of a blockchain as a byproduct of bitcoin transactions. How does this happen? Each bitcoin transaction is recorded in a ledger—its blockchain—that is shared among all the computers that make up the bitcoin network globally. Bitcoin’s ledger is made up of chunks of data—blocks—that when strung together form a history of every transaction ever made by anyone with bitcoin. This string of blocks is bitcoin’s blockchain. (Here’s our in-depth explainer on how transactions are added, or mined, in bitcoin.)

Since bitcoin was created nine years ago by the still-anonymous Satoshi Nakamoto as a tool to side-step banks and state-issued money, it has increasingly been embraced by the very institutions it was designed to disrupt. From Wall Street to the world’s biggest corporations, executives tasked with bringing innovation to their companies began talking about “blockchain technology” as an idea distinct from the cryptocurrencies that introduced it.

The idea is to reverse engineer a blockchain and extract the best features while leaving the rest behind for the anarchists and libertarians of the bitcoin world. The challenge is this: How do you create a blockchain without the speculation and mania linked to bitcoin? Instead of a blockchain being merely a record of bitcoin’s transactions, executives wanted a blockchain that could deal with existing, real-world assets to suit their own needs. That might be streamlining supply chains, automating banking back-room operations, or tracking diamonds.

Marrying blockchains with supply chains

The reason corporations think a blockchain could solve their supply chain problems can be summed up in one word: trust. It’s what blockchains were designed for—getting a bunch of counter-parties to agree on a set of transactions. That’s why bitcoin users, for example, can stay semi-anonymous.

In the same way, corporate blockchains would allow a vast network—from farmers to trucking companies to warehouse staff—to trust that a product has passed through a link in the chain and been processed in a particular way. Blockchains used by corporations are closed and can avoid the energy-intensive expense involved with “mining,” or doing the computing necessary to solve the cryptographic puzzle associated with bitcoin.

Blockchains for corporations have another critical property: greater privacy. While it may sound counterintuitive, corporations are even more sensitive over transaction data than, say, a dark web user who’s buying illicit drugs with bitcoin. When corporations first began to explore blockchain technology, they were keen to build in a layer of privacy over transaction data. That’s what consortia like R3, or the open-source Hyperledger project, have developed in the blockchain solutions they market to companies.

What Walmart and IBM are doing

To understand what Walmart’s experiments with blockchain technology looked like, I went to IBM’s research lab in Zhongguancun, which has been called China’s Silicon Valley because of the sheer number of technology giants headquartered there. There, I met with the team of research scientists and executives who worked with Walmart to develop a blockchain solution based on the open-source Hyperledger codebase.

Jin Dong, the associate director at IBM Research who led the effort, walked me through the process of putting Walmart’s pork on a blockchain. Data about the pork product, from the farm inspection report to the livestock quarantine certificate, are digitized by an “industrial personal digital assistant,” which basically looks like a smartphone in a rugged case. It emits a laser to scan barcodes rapidly. These data are then uploaded in real-time to Walmart’s blockchain. The farm is operated by a company called Jin Luo, and this particular farm is located in Lingyi city in northeastern China.

The other key link in this process are the drivers. They are charged with not only transporting the product, but bringing along the necessary documentation for each batch of products. With the blockchain pilot, they were asked to take photographs of the various certificates as they transported them from place to place, where they would be stamped. This verified that the products were moving along the supply chain. “If someone is looking for specific documents, it takes a lot of time to locate them,” Dong said. Digitized documents stored on a blockchain and linked to a particular batch of products removes that pain point, he said.

As the pork products wend their way to a Walmart distribution center, and ultimately, the shelves of a Walmart store, they have passed through dozens of hands. But the blockchain pilot allowed the progress of each batch of products leaving the farm to be tracked, in real-time, across the country. Once the products reach individual stores, they are repackaged and distributed across the shop floor. The pilot doesn’t track individual retail packages, IBM said, although the level of detail in the data is sufficient to track a particular retail package back to the farm. Changrui Ren, an IBM Research executive who worked on the project said, “The nature of blockchain ensures the data put into a blockchain can hardly be changed, and can be stored forever,” he said.

The complexities of sliced mangoes

After the Chinese pork pilot, Walmart ran a larger scale test tracking mangoes from central America. “We picked mangoes because it was more complex,” Yiannas says. “It was sliced mangoes.” The fruit is grown by small farmers, transported to a “packing house” where they are washed, then moved across the US border. They are then stored in refrigerated warehouses, and eventually end up at a Walmart supplier where they are sliced and packaged. “You can see how many layers it takes to get those mangoes from farm to table,” he says.

Current rules for food traceability require a retailer to be able to see one step forward and one step back in the chain. That means Walmart would only be required to know which importer handled a particular package of mangoes, but wouldn’t necessarily know which packing house it came from, or which farm. “It’s almost impossible to have a clear view of where this food came from,” Yiannas says.

Yiannas decided to put his team, and the blockchain they had built, to the test. He bought a packet of sliced mangoes and brought it to the office. “I told them to trace it back to the farm,” he says. After numerous phone calls, faxes, and emails, the results were in—six days, 18 hours, and 26 minutes later. “Those records had to be pursued,” he says. He then did a live demonstration of the blockchain system in front of the media, tracing the same packet of sliced mangoes. “We put in the code, and we traced it back to the farm in 2.2 seconds,” he says. “So we’ve gone from seven days to 2.2 seconds. This isn’t a lab. This isn’t theory. This is the real world.”

Every second matters, Yiannis says, because in a real food contamination situation, it could mean the difference between many more sick consumers, and farmers being forced to stop supplying their produce for fear they were the source of the contamination. “Imagine if there was a food scare and you didn’t know where the contaminated mangoes came from,” he says. “Seven days is a long time.”

Even the best blockchain architecture would be useless if producers and retailers couldn’t agree on how to use it. That’s why Yiannas pushed to create a consortium with other food companies—something that’s easier to do when you’re the world’s largest retailer by sales. “The blockchain solution we’re working on is the foundation to creating a smarter food system for the future,” he says. He envisions packing pallets with thermometers that transmit their cargo’s temperature instantly, and all that data stored on a blockchain can be accessed by customs officials, regulators, and consumers.

But in order for that data to be make any sense when it’s shared across silos, it has to be captured and formatted in standard ways. “Any companies working on [blockchain tech for food chains] need to base it on established standards,” he says. “Some new standards will need to be created, but I don’t think it will be that hard. The food industry has been working on this.”

Some skepticism

Some industry observers think blockchain’s potential in logistics has been overhyped. Gartner’s Ray Valdes, who covers blockchain projects for the research firm, says that many supply chain projects could just as easily be executed on a traditional database. “Ninety percent of enterprise blockchain projects today don’t need a blockchain, and would be better off without one because they are not aligned with what blockchain can actually do,” he says. Valdes says he advises most clients against using a blockchain in their technology system. “If I can save them $1 million on a pointless blockchain project, that’s our value,” he says.

When it comes to supply chains, Valdes believes that blockchains could play an important coordinating role. “If you have a fragmented business ecosystem, with many parties who don’t know each other but need to do business, then they could collaborate through a blockchain,” he says. But there’s a catch. “It’s a ‘boil the ocean’ problem,” he says, meaning that it’ll take fundamental shifts in an industry for adoption to take place. Optimistically, he says, it would take a decade for the industry to rearrange itself so that everyone was logging interactions on a blockchain.

Valdes argues that a company as dominant as Walmart doesn’t need its suppliers on a blockchain. It can simply ask its vendors to use whatever system it chooses. “They have been very successful because over the years they have built a robust system of record for their supply chain,” he says. “If you were a supplier to them, you would happily accept their centralized version of the truth.”

But in Yiannas’ mind, blockchains solve the very thing that prevent databases from being built and shared between companies: trust. He told me about his quest for the perfect food traceability system. “I’ve looked at them all,” he says. “I’ve been in pursuit of the holy grail.”

Blockchains, because they are designed for collaboration, could change all that, and not just for the Walmarts, Unilevers, and IBMs of the world. “It’s a new way of democratizing how data is captured and shared,” Yiannas says. “It’s a game-changer in that it’s adding value to all the participants.” Yiannas puts blockchain tech on the same scale as cellphones when it comes to digital disruption. “Imagine a digitally transparent system connecting farmers and buyers,” he says. “One of the best things that has happened to farmers in small economies is just giving them a cellphone so they can connect with buyers.”

The unlikely factor aligning a Walmart food executive and a mango farmer in central America, then, may turn out to be an accounting device that was first given expression in a stateless cryptocurrency, but which may turn out to be critical in resolving one of the food industry’s most vexing problems.