top of page

Food Safety Compliance And the Data Silo Problem

Updated: Jul 18, 2023

Staying Afloat in an Ocean of Data Silos

Big data has been with us for some time but its use in food supply chain compliance is only beginning.

There is an ocean of data available on the internet, within your company and in specialised database applications behind paywalls. The first challenge is finding the data that is useful for you. The next challenge is transforming the data into a useful configuration for your use case. Then there is the problem of ensuring that the data is accurate, up to date and of high quality. These are all very significant problems.

In almost all cases the data we need for food safety compliance is stored in multiple places, in different formats, and languages and none of these data sources ‘talk to each other’. These data silos are the most significant issue, and these barriers slow down the rate of digitisation. The table below provides examples of these silos and some of the problems this causes:




Raw material specifications

Supplier, Company emails, spreadsheets in network folders, separate systems

  • Data might change and third-party holders of data may not be aware of critical changes, e.g., allergen risk.

  • Data is locked up in email and network folders and valuable information is hard to extract, e.g., List of materials with ingredients that have a certain Country of origin.

Traceability data

Supplier, your Company, email, spreadsheets, paper, network folders, separate systems

  • Time consuming to carry out trace back or forward leading to higher recall costs and risks.

​Supplier / raw material approvals

Your Company, email, spreadsheets, paper, network folders, separate systems

  • Data is locked up in email and network folders and valuable information is hard to extract, e.g., List of suppliers who are approved for use with certain customer products.

Supplier alerts and recalls, import alerts and border refusals

Various regulatory databases throughout the world, e.g., US FDA, UK FSA, Canadian CFIA, Australian ACCC, EU RASFF

  • Data is in different formats on public web sites and hard to access.

  • Interpreting and understanding multiple data sources is difficult as there is no standard to how alerts and recalls are described.

  • Suppliers may not be named so it is difficult to know if you are affected.

  • Analysing the hazards and supply chain risks from multiple data sources is a very big task and not feasible for most companies to do.

  • No globally used method of food categorization making it hard to analyse.

Regulatory / Certification inspection data

​Regulator website, e.g., FDA Dashboard, Standards owner website, e.g., BRC, SQF.

  • Most regulators do not share inspection data (FDA being a notable exception).

  • Where information is available there is no standard format and no easy way to access and analyse this data.

Internal food safety and quality data

Documents and spreadsheets stored on individual PCs, network locations or a siloed software management system.

  • Data may be captured, costing time and money, however little value is realised as the data is not easy to analyse, being locked up in individual files spread throughout the business.

  • Trend information is not real time as staff only capture snapshot information and cannot join it with other data to see the big picture and to spot trends which can be arrested before problems occur.

  • Automated alerts, notifications and escalations are not implemented, defaulting to human responsibility which results in unwanted but preventable problems.

Corrective action data

Typically captured on spreadsheets or a siloed software solution.

  • NCRs, Corrective and Preventive Actions are not related to suppliers, internal processes, people, documents, and customers leading to poor tracking and trending.

  • Little management engagement or visibility of workload leading to important issues not being dealt with and causing problems.

​Risk data

Spreadsheets and third-party systems.

  • Risks are logged in spreadsheets but rarely reviewed over time to establish changes in risk.

  • No single view of risks across the supply chain or business means it is hard to prioritise resources where they are needed to minimise risk.

The examples above provide a glimpse into the problems with big data and show that a major barrier to digitization is disparate, scattered data sources like Excel spreadsheets, Word documents and multiple, individual applications. Interoperability is therefore the real problem which needs to be solved to really get on top of supply chain compliance and risk. Organisations like GS1 see themselves as key to this problem but adoption of GS1 standards by smaller food firms is not trivial and presents a barrier to adoption due to the costs and IT technology resources involved.

GFSI benchmarked and other standards like SQF, BRCGS, ISO22000, AIB and others attempt to solve these problems by taking a standardised food safety management system approach. Complying with these standards takes results in a lot of work, effort, and hidden cost by food businesses. However, the standards themselves often encourage a silo’d approach, by implicitly specifying different standalone ‘modules’, like supplier management and traceability, and failing to advocate a digital and data centric approach to the problem.

GFSI benchmarked standards urgently need to review this approach to help modernize the food sector by encouraging an integrated approach to food safety management through digitization. The FDA’s New Era of Smarter Food Safety is set to be a catalyst for this change as their influence with GFSI is significant.

The Many to Many Problem:

One root cause of many inefficiencies in the transfer of supply chain information, is the ‘many to many’ problem. The best way to explain this is that suppliers in the food supply chain are required to send, or submit through an online system, the same information to many customers. When a food safety problem occurs multiple levels of communication are required to try to trace the source. This is not always easy as traceability standards vary and products are mixed in the process, e.g., Romaine lettuce. The diagram below illustrates this:

The ‘many to many’ problem is driven by using a ‘one up one down’ traceability model, as is the industry standard required by regulators and GFSI standards alike. It creates significant duplication of effort which adds to the overall cost of compliance and ultimately results in higher costs and food prices. This duplication takes the form of:

  • Supplier Approvals

  • Material Specifications

  • Certification

  • Supporting Documents, e.g., HACCP plan

  • Certificates of Analysis

  • Traceability Information

This problem creates additional work, and costs, for suppliers and customers. Suppliers then find it difficult to track the information they sent to customers and vice versa. The risk is that changes can easily happen in the supply chain without customer notification, causing unwanted, costly recalls, product rejection and consumer harm.

This problem could only be solved if everyone used a common solution to send their trace data and documents. In the real world this will never happen without regulatory pressure. Therefore, the FDA’s New Era of Smarter Food Safety represents a big step in the right direction, but it will take a long time to come to fruition.

Meanwhile, blockchain and DLT are making limited inroads to the ‘many to many’ problem through solutions like IBM Food Trust and other block chain solutions. However, silos remain due to IBM working with mainly large retailers leaving smaller producers out of the opportunity due to their limited IT budgets.

A global, centralised system like the one shown below would be effective, but it would need to be very low cost to ensure that most of the food supply chain is able to get on board with it voluntarily.

The ideal solution would be a blockchain network in which all parties to the supply chain could engage and share their trace records, documents and required information. Each party would need to be in control of who could access their information with mechanisms to verify the identity of each party and the authenticity of the information being shared.

It will take a critical mass of the world’s food facilities, something akin to a Facebook moment, before the world’s food facilities engage in this way.

If you would like to learn more about block chain and its novel uses in the food industry, please feel free to get in touch for a quick chat on its use cases and find out how primority is adapting this technology to revolutionise the food safety industry.



bottom of page