Tracking What We Eat

Traditionally food travelled direct from the producer to the consumers. Eggs directly from the chickens, milk from the cows, vegetables from the garden. This still happens in many rural areas and has become more common again within developed cultures through farm and speciality shops. The supply chain quickly becomes more complex. A miller grinds from several farmers. A baker will buy from different millers. Following this pattern through to modern processed foods means that the supply sources of a dish such as a ‘ready meal’ lasagne are varied.

In Ireland in 2013 it was discovered that horse meat had been used instead of beef in some processed foods. Although it was easy to identify the butchers who had supplied the wrong meat tracking down exactly which meals this ended up in was tricky. Every single suspect meal could be tested but as this cannot be done without needing to re-package the meal this option is not viable. The simple solution was to recall all suspect batches. A process that almost certainly involved the recall of some items that contained no horsemeat.

It is quite common for food producers to advertise that an ingredient is from some named farm but in many cases this is an advertising ploy. The named farm is unlikely to exist. If it did then it would need to produce a much greater quantity and consistency of product than the food processing requires.

Some foods are labelled with their true origin. This is a protection against adulteration and a guarantee of their quality. It is most common where a product sells for a premium due to its place of origin or means of production. By evidencing the origin the producer makes it more difficult for competitors to label lower quality goods with a counterfeit origin. This will maintain prices and customer confidence in the quality of the food. A successful implementation is the New Zealand ‘Trace My Egg‘ program. A 5-digit code is stamped on each egg. This identifies the egg production process, such as free range, together with the farm the egg comes from.

IBM are convinced in the potential of tracking the origin of foodstuffs having launched their ‘Food Trust’ service. Take up so far has been limited but some players including the Carrefour Supermarket group have begun to use the service in more than a limited trial. The customer can scan a QR Barcode to find details about the origin of a product. Carrefour are able to track the origins of components within the foodstuff.

The IBM implementation runs on Blockchain and is offered as software as a service. Blockchain is slow compared to traditional database implementations and requires considerable processing resources (something that IBM would not be short of). The principle of collecting data as foodstuffs travel along the processing and distribution chain and presenting an access code to the user is not restricted to a particular technology. Systems must share a common back end store which could be Blockchain, SQL data storage, Big Data or some other solution. User access should never depend on the back end. Some solutions present the customer with a barcode to scan but that solely serves to redirect the scanner to a web portal.

Customer access to a web portal can provide additional information on the product. The ‘Thank My Farmer‘ implementation directs the customer to details about the origin of the product, its production and related environmental issues. It is, however, relatively generic, ‘Thank My Farmer’ does not provide information on the specific producer’s farm. The data is only as reliable as the portal address provided. A counterfeit pack can direct the customer to the same portal as the genuine goods.

Lessons from the FMD project can be used here. Any pack labelling needs to consider tampering of that label. Any data to be scanned should be integral to the pack not on a label that might be replaced or overprinted. This is not always viable, placing a code on a part of the pack that cannot be accessed without breaking the outer label benefits the manufacturer but not the consumer.

Adding additional information to the label will make it harder to spoof that product. Again looking at FMD each pack had a different code. Copying a single legitimate code many times would not then create any new legitimate codes. Modern packing lines have no issues with generating unique codes as each item is packed. The process will, however, create many more records than if a single code is used for 1 batch. Using an algorithm to link the codes within a batch can help as only the algorithm needs to be recorded but the solution depends on keeping that algorithm secret. FMD was able to cope with recording large numbers of unrelated codes in a traditional database. Blockchain would be less successful with larger volumes of codes because of the increased number of data access requests required to handle them.

With food products created from many sources then the origin of each component in the recipie can be identified together with any movements of those products within the supply chain. This will make product verification or recall a simple task for anyone who as access to the underlying data. The various producers, freight operators and processors would need to share the same or compatible data origins. The business models of suppliers such as the large supermarket chains are already creating an infrastructure where this can be put into place.