The FDA and Food Safety
I promised some folks in the ag world who follow this newsletter I’d discuss food safety at some point. Today’s the day!
Recently, the FDA proposed new requirements around traceability for companies who produce, process, pack, and ship certain foods as part of the Food Safety Modernization Act (FSMA).
The TLDR of the new rule is that stakeholders at each step in the supply chain will have to track information related to the product (like location or point of origin). The objective is to make it faster to trace the lineage of products in case of a recall and reduce risk to consumers.
Here’s what this looks like:
If you receive raw products from a farm via truck or ship, you now have to track the location and ID of where that product came from, entry number (if imported), location and ID of where the product was received, quantity of product, product ID and description, a point of contact for whoever generated the product ID, reference records and reference numbers for the receipt of the product (eg invoice numbers), and transporter information (who transported the product). These are called Key Data Elements (KDEs). Each step in the supply chain has a unique collection of KDEs they have to track.
So let’s say there’s a recall. In theory, you should now be able to quickly identify where the product came from based on the affected ID. The originating farm can then also recall any affected product that was shipped to others. They can trace this forward because the farm will have to maintain a record of where product goes. It should be no mystery who receives product and where it goes.
Here’s an example from the FDA for Romaine lettuce.
For anyone who doesn’t work in the food/agriculture industry, this is called trace forward/trace back protocol. The changes being implemented require additional tracking events and establish recordkeeping requirements, but the overall philosophy remains the same. We’re trying to mitigate foodborne illness by maintaining provenance for products. This helps us identify where products came from and where they’ve been sold faster.
This is all pretty neat and it’s is the kind of innovation I love to see. There’s nothing technological about it at all. By simply tracking an ID code at each step of the supply chain, you can identify when things go wrong and help stop it from spreading.
I love low-tech solutions in low-tech industries. They’re relatively easy to implement and they work. You don’t necessarily need AI and blockchain, for example, to solve the traceability problem in agriculture. There are some pretty incredible first steps we can take as an industry that are ultimately transformative.
The proposed requirements aren’t without their limitations however. Here are a few:
No standards for IDs - this is one of the big issues in food/ag and stretches far beyond this regulation. We have very little standardization for nomenclature and identification of products. There are lots of types of ID formats (GS1 is the big one), but there’s nothing that is ubiquitous. The proposed regulation states that each party in the supply chain can choose their own format. They just have to let the FDA know what that format is. This is a good first step, but I would love to see us implement some set of standards across the supply chain to make identification easier and better. If the ultimate goal is to stop recalls quickly, standardizing IDs is going to help stop blockage when there is confusion as to what each ID format is and means.
Lack of visibility at the farm level - FSMA and all of the food safety regulations do little to track data at the farm level. While this proposed regulation starts to add some visibility at the farm level, there is still room for better origination data. Instead of just tracking what was harvested and where it goes, we could do better by thinking about what happened while plants grew. If we begin to track which inputs are consumed, which workers are touching plants, which SOPs are followed, we can start to become more proactive about risk.
Reactive Risk Control - these requirements are still reactive. We are essentially creating lineage so that in the case of a recall, we can stop further damage. I get way more excited thinking about how we can prevent incidents from happening at all. What if we used the traceability information, production data, and data science to analyze risk real-time? Then we could start preventing outbreaks from the get-go.
Food safety is so important and it’s a really exciting area for innovation. I’m glad to see the FDA take steps toward a more traceable system, though I think it’s only the beginning.
What I’m reading
I just started reading No Rules Rules. I’m only on Chapter 4, but I can already say it’s worth a read for any entrepreneur. Reed Hastings is notorious for doing things differently at Netflix and it’s cool to see how he went about creating the culture he did. I love hearing why entrepreneurs choose the paths they do.
Up Next
I’ve got a few topics I’m going to write about over the next few weeks (most have been requested by folks reading this newsletter):
Communication styles for high performance teams
How to ask someone to be an advisor
How to build a great pitch deck
GTM/pricing strategies
What else would you like to hear about? I’d love to get your feedback by email or in the comments.