THE LETTUCE NOBODY RECALLED

A position paper on FSMA 204, traceability, and the commercial value of specificity.

Amy L. Gates

Chief Executive Officer, Silo Technologies

Updated August 2026

Position

Traceability should be treated as more than a regulatory obligation. For fresh produce and specialty crops, its greatest value is the ability to get to a precise answer quickly: to identify implicated products, protect consumers, and just as importantly, demonstrate when an operator is not part of the problem.

This summer added something to that. Getting to a precise answer quickly is no longer enough on its own, because consumer confidence is now moving faster and further than the recalls that trigger it. The industry's next step is not simply more data. It is one source of truth that runs from field to finance, so the answer exists before anyone asks for it rather than being rebuilt by hand every time.

---

Software at the Speed of Produce.

The Lettuce Nobody Recalled

Earlier this month, a Salinas Valley grower chopped up roughly 300,000 pounds of romaine hearts and plowed them back into the dirt. Nobody recalled that romaine. Nobody tested it, and no regulator named it as part of this summer's Cyclospora outbreak. It was grown roughly 2,000 miles from the central Mexico operation implicated in the investigation, but it went into the ground anyway because retail and foodservice demand for salad had fallen. [1]

I've spent more than 25 years in produce, and I know what it takes to get a crop to the point of harvest. I've run a grower-packer-shipper importing from Mexico and South America, overseen food safety and FSMA compliance for a global importer, and spent years volunteering on our industry's food safety and traceability efforts. So when I hear about an operator disking under an unimplicated crop because of an outbreak somewhere else, I don't just think about the waste. I think about the people, the investment and the months of work that went into growing it. I also think we need to ask why, after everything we've learned over the last 20 years, this continues to happen.

The recall itself was actually quite specific. By historical standards, the investigation moved quickly. FDA's traceback converged on Taylor Farms de Mexico, and the July 17 recall identified recalled iceberg lettuce sourced from central Mexico. Yet the story did not really settle, and that matters because the market does not have the luxury of waiting indefinitely for perfect information. [2]

A category manager deciding whether to keep an item on the planogram is not going to spend hours sorting through the difference between a recalled lot, a growing region and an entire commodity. Neither is a consumer standing in front of the salad case. When people do not have enough information to make a narrow decision confidently, they make a broader one. Buyers reduce exposure, consumers stop purchasing and suddenly an outbreak tied to a particular source becomes a problem for an entire category.

Where the investigation stands now

As of August 17, FDA had stated that the shredded lettuce implicated in the larger of this summer's Cyclospora outbreaks was removed from commerce. That is about as close to saying an outbreak is over as FDA is generally able to get. The agency has moved into an intensive root cause phase that includes inspection and sampling. Acting FDA Commissioner Kyle Diamantas said Americans should feel confident eating fresh produce, including leafy greens, and HHS described the outbreak as contained. Michigan returned consumers to routine lettuce and salad advice while continuing to advise against specifically recalled product. [3]

At the same time, FDA continues to investigate several smaller Cyclospora clusters for which no food vehicle has been publicly identified. The largest is roughly 172 cases, with smaller active investigations of approximately 25 and 11. North Carolina deserves its own attention. NCDHHS has reported more than 900 cases since May, with roughly half concentrated in Wake County, and the state has said its increase does not appear to be associated with the iceberg lettuce outbreak. [3]

I want to be careful with the numbers here, because a lot of people have not been. The roughly 30,000 cases being aggregated from state reporting are not 30,000 shredded lettuce illnesses. Precision matters in both directions. When we are imprecise about what is implicated, good product goes into the ground. When we are imprecise about how large a problem is, confidence goes with it.

The number that should stop all of us

Here is what bothers me most about where we are right now. The implicated product is off the market. The federal government is telling people the food supply is safe. And consumers still are not buying.

New receipt data from Fetch, which analyzes verified receipts representing $212 billion in annual U.S. retail sales, found that shopping trips containing fresh lettuce dropped 10 percent over a two week period compared with the two weeks before it. Fresh salad mixes fell 23 percent. Packaged salad kits fell 27 percent. Measured by individual buyers rather than trips, the pullback was worse. Twenty five percent fewer shoppers bought fresh lettuce, and shopper counts for salad mixes and packaged kits dropped by roughly 40 percent. Frozen alternatives rose 17 percent in the same window, and Fetch saw no meaningful change in promotions that would explain it, which tells you the shift came from consumers and not from merchandising. [14]

Numerator found that 87 percent of produce shoppers were aware of the outbreak, that more than a quarter of households said they stopped buying salads and fresh greens entirely, and that one in ten cut back their spending. Numerator put roughly $280 million in consumer produce spending at risk for every month the outbreak continues. More than 6.5 million U.S. households stopped buying salad mixes and kits in the month ending July 26 compared with the month before, and per-household spending in the most affected states dropped 10.7 percent. Consumers did not fully replace what they stopped buying. [15]

NielsenIQ scanner data reported by Quartz showed dollar sales of prepackaged salads down 14 percent over a four week period at the peak of the recall, with weekly unit sales of fresh head and leaf lettuce down 9 percent. Then the market did what markets do when demand disappears. Bureau of Labor Statistics Consumer Price Index data showed retail lettuce prices fell 16.4 percent in a single month, the largest one month decline ever recorded for the category. [14]

I want you to notice that 27 percent. In 2006, five months after the spinach outbreak, retail sales of bagged spinach were still down 27 percent from the year before. This summer, packaged salad kits fell 27 percent inside of two weeks. [4] [14]

Those two figures do not measure the same thing, and I do not want to pretend they do. One measures how long the damage lasts. The other measures how fast it arrives. Read together, they tell you something more useful than either one alone. The hit shows up within two weeks, and twenty years of evidence says it can still be sitting on your P&L five months later.

There is one encouraging number in all of this. Among consumers who stopped buying salad mixes and kits because of the outbreak, 82 percent said they plan to buy them again once it is contained. The demand is not gone. It is waiting on confidence. [15]

I do not believe we rebuild that confidence with reassurance. We rebuild it with proof, and proof is a records problem before it is anything else. A shopper's question and a category manager's question are the same question our industry has always struggled to answer quickly: what is this, where did it come from and where did it go? Everything I say in the rest of this paper comes back to that.

So did traceback fail us?

I have heard people look at those numbers and conclude that traceback failed this summer. I do not believe that is right, and I think getting it wrong will send our industry after the wrong fix.

Traceback worked. FDA converged on a supplier and identified recalled product, and by the standards of most of my career that was relatively quick. The people who did that work deserve credit for it.

But it is worth looking at what a traceback is actually designed to do. Investigators work backward from where people got sick, through the restaurants and retailers and distributors that served them, until the records point at a supplier those chains have in common. The operation most likely to turn up as that common name is the one selling into the most places. That is arithmetic more than detective work, and it is a sound way to find the center of an outbreak. What it does not do, and was never built to do, is clear anybody at the edges.

The edges are where most of this industry lives.

A traceback answers one question: which product is implicated. The market asks two questions at the very same moment. Which product is implicated, and which product is not. We have spent 20 years building for the first question. Research, standards, sampling, PTI, whole genome sequencing, the traceback methodology itself. We have built almost nothing for the second.

And here is the asymmetry that bothers me most. The companies most likely to be identified quickly are the ones with the scale to be identified. The companies least able to answer for themselves are everyone else. A large processor has a food safety department, a QA team and an ERP that was implemented by consultants. A family operation running crews through the season has a bookkeeper, a clipboard and a spreadsheet somebody built in 2015. When the phone rings, the obligation is identical. The ability to answer it is not.

When the second question goes unanswered long enough, the market answers it on its own. It assumes the worst about the entire category. That is what a 27 percent decline actually is. It is not a verdict on an implicated product.

And I want to be clear that this is not for lack of willingness. Not one operator I talked to this summer was holding anything back. Every one of them would have handed a buyer their entire supply chain that morning, gladly, if they could have put it together. They could not assemble it fast enough for it to matter, and by the time they could, the decision had already been made without them. That is not unwillingness. That is a company that wants to answer and has no way to do it on the clock it is being held to.

So I want to be precise about what I am arguing, because it would be easy to hear this as an argument against speed. It is not. Speed to recall and speed to exoneration are the same capability. Connected, lot-level data, produced under pressure. The difference is who can actually reach it.

The central argument

Speed to recall and speed to exoneration are the same capability. The difference is who can reach their own data when the phone rings. We built recall speed into the regulator and the largest handlers. We never built exoneration into everyone else, and everyone else is who the market can hurt.

We do not have a data problem. We have a connection problem.

This is the part of the conversation that I think we sometimes get wrong. We talk about traceability as though the produce industry simply has not collected enough information. After spending most of my career inside produce operations, I do not believe that is the problem. Most companies have an enormous amount of data. The problem is where it lives.

I recently heard about a large distributor pulled into this summer's outbreak that had the traceback information it needed the entire time. Getting to the answer required going through five separate systems. Field records might be in one place, labor in another and inventory somewhere else. Receiving could be in the ERP, shipping information on the BOL and cooling records on a clipboard in the shed. Every one of those systems and processes may be doing exactly what it was designed to do, but none of them tells the whole story.

So when the phone rings, a person becomes the integration layer. Someone has to open each system, find the right records, reconcile them and put the story back together. You do not really feel the cost of that fragmentation during an ordinary Tuesday. You feel it when somebody starts a 24-hour clock and every hour spent assembling information is another hour your buyer, a regulator or the market is left wondering whether your product is involved.

What operators need is one source of truth. One place you go when the phone rings that can tell you where the product came from, where it went, which lots, which dates and which customers. That is what FDA is asking for, and it is the same thing your largest buyer is asking for. It should be one query, not a scavenger hunt across five systems and three people's memories.

I use that phrase deliberately, so let me say what I mean by it. A source of truth is not a report you build after the fact. It is the record the business already runs on, captured once as the work happens, so that the answer exists before anybody asks for it. If you have to assemble it, you do not have one. You have five systems and a very good employee, and that employee is your traceability program.

One source of truth does not mean one system, and I want to be clear about that because it is where these conversations usually go wrong. Nobody is throwing out the scale software, the cooler's system or the labor app they have run for fifteen years, and I would not ask them to. What operators need is a primary source of truth. One system that knows about the others, that the rest feed into, and that you go to first when the phone rings.

That is what we are building at Silo, and I am not going to be coy about wanting your business. But I would rather you prove the problem to yourself than take my word for it, which is why the first thing I recommend in this paper is a mock recall, try a traceback and see what it takes.

I would also say something about the clock. FDA's requirement is 24 hours, and I think our industry has quietly accepted that as the target rather than the floor. Your largest customer does not work on the regulator's clock. Your buyer wants to know before the category review at noon. I am not going to put a number on what the right answer is, because nobody in this industry can honestly deliver one yet, including us. But it is a great deal shorter than 24 hours, and it is not reachable at all as long as the record has to be rebuilt by hand every time somebody asks.

We have been here before

In 2006, FDA advised consumers not to eat fresh spinach during a multistate E. coli O157:H7 outbreak. The outbreak caused more than 200 illnesses and three confirmed deaths. Five months later, the value of retail sales of bagged spinach was still down 27 percent from the same period a year earlier. Growers who had never shipped an implicated leaf felt the commercial impact anyway. [4]

What happened next is one of the reasons I care so much about this issue. The industry did not simply move on. It created solutions. The Center for Produce Safety was founded in 2007 and has since become a major source of produce-specific food safety research. The Produce Traceability Initiative became an industry-led effort to improve whole-chain traceability and standardization across produce. I spent four years on the CPS board and served on PTI's Joint Industry Leadership Council, so I have seen firsthand how much work has gone into the science and standards underneath today's food safety system. [5]

Then, in the fall of 2018, a romaine E. coli outbreak led to a broad consumer advisory before investigators could narrow the risk sufficiently. Researchers later estimated total societal losses from that incident at $276 million to $343 million, with substantial losses borne by processors, shippers and retailers that had no role in causing the contamination. The incident was ultimately traced to romaine from a single farm. [6]

The organisms are different, the outbreaks are different and I do not want to oversimplify any of them. But the commercial pattern is hard to ignore. When investigators and businesses cannot get specific quickly enough, the market gets general, and everyone in the category can end up paying for what happened at one operation. We have spent 20 years building better science and better standards. What we still have not finished building is the connected data infrastructure underneath them.

This summer also showed us what happens when specificity works

At almost the same time as the Cyclospora investigation, another outbreak involving a Food Traceability List commodity unfolded differently. CDC and FDA identified fresh jalapenos from Sinaloa, Mexico, distributed by Coast Citrus Distributors, as the source of a multistate Salmonella Javiana outbreak. Traceback identified a common grower in Sinaloa as the likely source. Coast Citrus agreed to recall the implicated product, Chipotle and QDOBA switched away from the affected jalapenos, and public health officials stated that given those actions they did not consider those establishments to represent a current ongoing risk to consumers. FDA updated its outbreak advisory on August 10 with recalled product information and has initiated sampling. [7]

That distinction matters. Companies had been publicly connected to a national outbreak, but because traceback was specific enough, they could act and public health officials could then communicate what was implicated and what was not. To me, that is one of the most important business cases for traceability that we do not talk about enough: exoneration.

Most of the conversation around traceability understandably focuses on how quickly we can find contaminated product. That has to stay the first priority, because faster traceback shortens exposure and saves lives. But most companies in this industry will never be the source of an outbreak, and every one of them still gets asked. For them the same capability points the other direction: the ability to prove, quickly and with evidence, that the problem is not theirs.

When a buyer calls and asks whether you are implicated, saying no is an assertion. Being able to immediately produce the relevant lot codes, growing regions, harvest dates, receiving records and shipments is evidence. The faster we can move from a commodity, to a region, to a farm and ultimately to a specific lot, the less unnecessary damage gets spread across everyone else.

There is one more thing this summer taught me about exoneration, and it is the part I had not fully appreciated before. It is not something a company can finish on its own. Your traceback is only as fast as your supply chain partners can produce. If I can pull my lot codes in an hour but the grower I bought from, the cooler that held it, the distributor I sold to or the broker in the middle cannot pull theirs, my answer is still incomplete. Exoneration is a chain problem, not a company problem, and that is why I keep coming back to connected data rather than simply better recordkeeping inside any one business.

What FSMA 204 is really asking us to do

The Food Traceability Rule implements Section 204(d) of FSMA and is codified at 21 CFR Part 1, Subpart S. For foods on the Food Traceability List, the rule requires records containing Key Data Elements associated with specific Critical Tracking Events. Covered entities must be able to provide required information to FDA within 24 hours, or within another reasonable time agreed to by FDA. The current compliance and enforcement date is July 20, 2028. [8]

For produce companies, the practical work is maintaining an unbroken relationship among events such as harvesting, cooling where applicable, initial packing, shipping, receiving and transformation. The Traceability Lot Code is assigned at certain events defined by the rule, including initial packing of a raw agricultural commodity other than food obtained from a fishing vessel, and it generally moves through the supply chain unless another event requires a new code. [8]

I do not think operators should look at 2028 and assume they have two years to start thinking about this. Your customer's deadline may come much sooner, and commercial expectations can move faster than regulation. Walmart, for example, set enhanced food traceability requirements with an August 1, 2025 date for covered suppliers. Whether or not every buyer uses the same approach, the direction is clear: large customers increasingly want standardized, lot-level data to move with the product. [9]

That is also why I do not put much comfort in the argument I hear from operators whose commodity is not currently on the Food Traceability List: this does not apply to me. Maybe not technically today. But if your customer decides that lot-level traceability is going to be its standard for receiving food, the regulatory distinction becomes much less important.

The rule is also not the only thing moving right now. On August 11, FDA finalized its Guide to Minimize Biological Hazards in Ready-to-Eat Fresh-Cut Produce, with the Federal Register notice published the following day. It supersedes FDA's 2008 fresh-cut guidance, and while the guidance itself is nonbinding, much of it explains existing mandatory requirements under 21 CFR Part 117. Two pieces of it are worth reading closely if you handle fresh-cut product. FDA now treats wash water as a potential process preventive control whose parameters should be validated under actual operating and worst-case conditions. And FDA states plainly that it is not aware of an antimicrobial suitable for fresh-cut produce that would significantly minimize or prevent Cyclospora cross-contamination, pointing instead to other preventive controls including supply-chain controls. [16]

Read that alongside this summer and the message is hard to miss. For some hazards, there is no kill step and no chemistry that solves it for you. What you know about your supplier, your grower and your lot is the control. That is a traceability question wearing a food safety hat.

What 2011 taught me

At the time, I was vice president at Frontera Produce. We bought and sold cantaloupe grown by Jensen Farms, a fourth-generation operation in Holly, Colorado. The facility had gone through a third-party audit shortly before the outbreak and received a superior score of 96 percent. [10]

Then the FDA went into the packing shed and found Listeria monocytogenes in the environment. The outbreak ultimately involved 147 illnesses across 28 states, 33 outbreak-associated deaths and one miscarriage. It remains one of the most consequential foodborne illness events in modern U.S. produce history. [11]

I have sat with what that means for a very long time. I have seen the families, and I want to be careful here because whatever happened to my company afterward is not the story of that outbreak. No software company, and certainly not me, is entitled to turn 33 deaths into a marketing lesson.

But there is a part of that experience that is mine to tell. Frontera did not plant, grow, harvest or pack that fruit. We bought it and sold it, and we followed the verification practices the industry prescribed at the time. When the grower went bankrupt, litigation looked for everyone else who had touched the product, including the distributor, auditors and retailers. We spent five years keeping a business alive while working through something for which there was no real playbook.

The detail I have never forgotten is that the audit passed. The accepted protocol had been followed, yet it did not identify what was happening inside that facility. That experience changed the way I think about verification and traceability. An audit tells you something about the conditions observed during an audit. A test tells you something about the product or environment that was sampled. Those are important tools, but they answer different questions from traceability. Traceability tells you what you handled, where it came from and where it went.

Traceability would not have prevented the Jensen Farms outbreak. A packing shed contaminated with Listeria is a food safety and operational failure, not a data problem. Where traceability becomes critical is everything that happens after something has gone wrong: how quickly implicated product can be identified, how precisely businesses can define their exposure and how much of someone else's failure lands on an operator who cannot demonstrate where their product began and ended.

There is a live opportunity attached to this one. FDA held a public meeting on Listeria monocytogenes prevention on August 18 and 19 and has asked for stakeholder input on sanitation and environmental monitoring, testing and sampling, corrective actions, root cause investigations, facility and equipment design, supplier controls and emerging technologies. Comments are due November 2, 2026. Fifteen years after Jensen Farms, FDA is asking our industry what actually works in a packing shed. If you have operational experience worth sharing, this is a place to spend it. [17]

Traceability helps us get to the truth

There is another reason this matters beyond recalls and commercial protection. Good traceback also helps investigators find the actual root cause. In August, IFPA asked CDC to broaden its Cyclospora hypothesis-generating questionnaire to more systematically consider drinking water and other environmental exposures alongside food. It would be easy to interpret that as the produce industry asking investigators to look somewhere else. I do not see it that way. We should want an investigation to be correct wherever the evidence leads. [12]

The current phase of this summer's investigation is exactly that kind of work. FDA has moved into inspection and sampling to determine root cause, and several smaller Cyclospora clusters still have no identified vehicle. North Carolina investigators noted that recent patients reported eating parsley and cilantro more frequently than expected compared with population surveys, though neither state nor federal officials have implicated either product. I am not suggesting anything about parsley or cilantro. I am pointing out that investigators can raise a hypothesis like that only because they have data, and they will only be able to rule it in or out with better data. [3]

The 2018 romaine outbreak is a good example of where this leads. Investigators ultimately found the outbreak strain of E. coli O157:H7 in sediment from an on-farm water reservoir in Santa Maria, California. FDA reported that the farm had turned up repeatedly across the traceback investigation and concluded that water from the reservoir most likely led to contamination of some romaine lettuce consumed during the outbreak. [13]

That sequence matters. Traceback gets you to the ranch. The ranch gets you to the reservoir. The reservoir helps you understand the potential route of contamination. You cannot sample the well until you know which field.

Environmental root-cause work depends on specificity. Traceability is not only about finding contaminated lettuce. It helps put investigators on the specific ground where the cause may exist, whether that turns out to be agricultural water, adjacent land use, wildlife intrusion, sanitation or something we have not considered yet.

Better investigative questions widen the hypotheses we are willing to test, while better traceability gives investigators the information to test those hypotheses and, just as importantly, rule them out. I am not arguing that this summer's outbreak was caused by something other than the implicated lettuce. The FDA and CDC investigation supports the recalled iceberg lettuce as the source. My point is that all of us should want the answer to be right, wherever it leads, and the industry's responsibility is to make sure our data is complete and connected enough that investigators do not have to guess. [2]

What we are building at Silo

Silo is a software company that serves the fresh produce and specialty crops industry. Our customers are growers, packers, shippers and distributors, and our job is to track and trace produce across the operations and transactions that move it from the field to the invoice. That is the lens I bring to everything above, and I want to be clear about what exists today and what we are still working toward. This industry does not need another technology company promising something it cannot actually deliver.

Silo already sits where most of this information is created. Purchasing, inventory, receiving, shipping, invoicing and accounting are transactions that happen inside the platform as a matter of doing business, not as a separate compliance step. That is why it is the natural place for the thread to land. We are not trying to get operators to keep a traceability record. We are trying to make the traceability record the same record they already use to run the company.

For 40 years, Pet Tiger has captured what happens in the field, including crew, piece rate, pick date, field and GPS information. Growers using it already generate much of the origin information that becomes important for traceability. What we are building now is a structured connection between that field-level information and the operational and transaction data in Silo, so harvest can connect to cooling, packing, receiving and shipping without losing the identity of the product along the way.

Much of the rest is integration, and that is deliberate rather than a shortcut. Connecting to the accounting system an operator already uses, moving EDI documents with trading partners, pulling field and labor information out of Pet Tiger, and accepting what other systems already produce instead of asking anyone to produce it twice. Every one of those connections removes a place where somebody currently retypes something, and every place somebody retypes something is a place your record can break without anybody noticing until the day it matters.

Automation is the part I care most about, and it is the part I think our industry underestimates. A traceability program that depends on discipline will fail on the worst day of your year. If capturing the lot requires somebody to remember, it will not be there at four in the morning during a wet harvest with a short crew and a truck waiting. The record has to exist because the shipment got created, not because someone was diligent. That is the difference between a system of record and a filing habit.

Cooling is particularly interesting to me because I think it will expose one of the bigger operational gaps for applicable commodities. In many operations, that information still lives on a clipboard or in a process disconnected from the rest of the business. Solving that is not simply a software problem. Cooling happens on a dock or in a cooler, sometimes at four in the morning, with somebody wearing gloves who already has a job to do. If capturing the data creates another cumbersome step for that person, we have missed the point.

Shipping and receiving are further along because those transactions already happen inside Silo. The remaining work is making sure the right traceability information becomes a discrete part of those records and can travel with the shipment to the next trading partner. Ultimately, I want that information to move as a natural byproduct of doing business rather than as a separate compliance exercise.

Then there is what I think of as the 24-hour answer. Today, a Silo customer facing an FDA request can assemble the information they need, but it still requires more than one screen. The one-click, filterable, FDA-formatted sortable export we want is on our roadmap. It is not in customers' hands today, but it along with supply chain integrations are being built and will be ready in coming months. 

And because exoneration is a chain problem rather than a company problem, the piece we are working toward beyond our own four walls is the Silo Network, where traceability information travels with the product from one trading partner to the next instead of stopping at the edge of each company's system.

The larger vision is actually pretty simple. The information FDA cares about is largely the same information operators need to run their businesses: lot, origin, date, quantity, location, inventory movement and customer. Capture it once, connect it across the operation, and compliance stops being another job and another system. One source of truth, one unbroken thread, from field to finance.

What I would do now

When I spoke at the Western Food Safety Conference in May, I said I believed roughly 80 percent of suppliers in this industry were not ready. I got a lot of nods, and I do not think that is because operators are resisting compliance. Most of the people I talk to are simply overwhelmed by it. They do not know which KDEs they already capture, which ones they are missing or where all of that information currently lives.

My first recommendation is still not to buy anything,, until you have done one thing. Walk your chain. Start at harvest and follow a real lot through cooling, packing, shipping and receiving. Identify what information you already capture at each point and where it lives. Most operators discover that nearly everything they need is already being captured somewhere. The problem is that there is no single place to go and get it.

Then run a real mock recall. Pick a lot from 90 days ago and actually start the clock. Do not sit around a conference table and talk through what everyone believes would happen. Open the systems, find the records, build the report and see how long it takes.

If all of the information exists but it takes five systems and half a day to assemble it, you have your answer, and you did not need me to give it to you. You do not have a data problem. You have no primary place to go get your data, and nothing connecting the systems that already hold it. That is the conversation I would like to have with you, and I would like the chance to earn it.

I would also look closely at repacking and any value added mixes, because both can quickly widen your exposure when the documentation is not strong. And before you finish, build what I call your it's not us packet. If a headline breaks tomorrow morning and your largest customer calls, what can you send them by lunch that demonstrates whether your product is involved? Figure that out now, while nobody is waiting on the other end of the phone.

Why this matters to me

After more than 25 years in this industry, I have been on more sides of this issue than I ever expected. I have run the grower-packer-shipper. I have been the distributor in the middle. I have served on boards and committees and spent years working on food safety and traceability. I have also been part of a company trying to survive the aftermath of an outbreak, and now I am building technology for the operators who have to answer these questions when the phone rings.

Through all of those experiences, what I want for this industry has become pretty simple. I want us to save lives first. Faster and more precise traceback means we can identify implicated product sooner, remove it faster and shorten the window in which people may be exposed. That has to be the reason we do this.

I also want us to protect good operators and family farms. The Salinas grower who put roughly 300,000 pounds of romaine back into the ground was not implicated in this outbreak. Neither were countless spinach growers in 2006 or romaine growers who suffered from broad market reactions in 2018. When our industry can be more specific, we can protect the public while also reducing unnecessary harm to people who had nothing to do with the problem. [1]

And I want us to get our customers back. Eighty two percent of the consumers who stopped buying salad said they intend to buy it again once this is contained. That is not a marketing problem and it will not be solved with a campaign. It is a proof problem, and proof is what traceability produces. [15]

I want to be careful about one claim here, because it would be easy to overstate. A software company does not restore consumer confidence. Shoppers do not read lot codes and they never will. What our industry can do is be specific enough, fast enough, that retailers keep buying, that public health officials can say clearly what is not implicated, and that reporters have something precise to write down. Specificity at the trade level is what keeps a shopper from hearing that lettuce is dangerous when the truth is that one supplier's iceberg was. That is the whole chain, and it starts with records.

And ultimately, I want us to build the most reliable fresh food supply chain in the world. We grow some of the healthiest food people put in their bodies. There is no reason we should not also be able to tell them, quickly and confidently, where it came from and where it went.

That is why I have stopped thinking about FSMA 204 primarily as a compliance deadline. Yes, we have requirements to meet, and yes, we need to be ready. But if all we do is comply with the rule, I think we have missed the larger opportunity. We have the chance to build a supply chain that can get to the truth faster, so that when something goes wrong we can find it, contain it and protect people, and when the problem is not ours, we can prove that just as quickly.

Three hundred thousand pounds of marketable romaine going back into the Salinas Valley soil should bother all of us. So should a 16.4 percent one month collapse in lettuce prices after the implicated product was already off the shelf. Not only because of the waste, but because of what both numbers represent: operators who were not implicated still paying the price for uncertainty somewhere else in the supply chain. After 20 years of outbreaks, research, new standards and new technology, I believe we can do better than that. And I believe we should. [1] [14]

Silo: an invitation to help build the next layer

We are opening the beta for Pet Tiger in the Cloud in September, connecting field and labor information directly into Silo as we work toward one connected thread from field through finance. I am looking for a first group of operators who know this work and want a real voice in how we build it.

If you simply want to run a mock recall and find out whether you have one source of truth or five systems and a very good employee, we will walk the chain with you. There is no obligation at the end of it, and if what you find makes the case for us, I would rather it came from your own records than from my slide deck.

Software at the Speed of Produce.

Technology should make it easier to answer the hard questions, not create another place to look for the answer.

References

1. Blue Book Services. Lettuce growers plowing under unsold lettuce crops. August 2026. Also: The Wall Street Journal, California Lettuce Growers Skip Harvest, Plow Their Crops Back Into the Ground, August 1, 2026. Both report roughly 300,000 pounds of Salinas Valley romaine hearts being plowed under amid reduced demand.

2. U.S. Food and Drug Administration. Investigation of 15-State Outbreak of Cyclospora Illnesses: Iceberg Lettuce (July 2026). Updated August 2026.

3. Centers for Disease Control and Prevention and U.S. Food and Drug Administration, Cyclospora investigation updates, August 2026; North Carolina Department of Health and Human Services cyclosporiasis reporting, August 2026; International Fresh Produce Association Food Safety Update, August 18, 2026.

4. Centers for Disease Control and Prevention. E. coli Outbreak From Fresh Spinach, 2006 archive; U.S. Department of Agriculture Economic Research Service, Outbreak Linked to Spinach Forces Reassessment of Food Safety Practices, June 2007.

5. Center for Produce Safety. About CPS and 2025 organizational materials noting CPS was founded in 2007; International Fresh Produce Association, Traceability, describing the Produce Traceability Initiative as an industry-led whole-chain traceability effort.

6. Spalding, Ashley, Rachael E. Goodhue, Kristin Kiesel, and Richard J. Sexton. Economic impacts of food safety incidents in a modern supply chain: E. coli in the romaine lettuce industry. American Journal of Agricultural Economics 105, no. 2 (2023): 597-623. DOI: 10.1111/ajae.12341.

7. Centers for Disease Control and Prevention, Salmonella Outbreak Linked to Jalapenos, updated August 2026; U.S. Food and Drug Administration, Outbreak Investigation of Salmonella: Jalapeno (August 2026), advisory updated August 10, 2026.

8. U.S. Food and Drug Administration. FSMA Final Rule on Requirements for Additional Traceability Records for Certain Foods and Frequently Asked Questions: FSMA Food Traceability Rule. Current compliance and enforcement date: July 20, 2028.

9. Walmart Supplier Food Safety and Traceability requirements and related supplier guidance. Enhanced traceability requirements for covered suppliers took effect August 1, 2025.

10. United States v. Eric Jensen and Ryan Jensen, U.S. District Court materials, 2013, noting the 2011 third-party audit of the Jensen Farms packing facility resulted in a superior score of 96 percent.

11. Centers for Disease Control and Prevention. Multistate Outbreak of Listeriosis Linked to Whole Cantaloupes from Jensen Farms, Colorado, final update. 147 persons infected in 28 states, 33 outbreak-associated deaths, and one miscarriage.

12. International Fresh Produce Association. IFPA Urges Broader Approach to Cyclospora Investigations, Submits Recommendations to CDC. August 2026.

13. U.S. Food and Drug Administration. Investigation Summary: Factors Potentially Contributing to the Contamination of Romaine Lettuce Implicated in the Fall 2018 Multi-State Outbreak of E. coli O157:H7. February 2019.

14. Dutton, Jill. Salad Kits Down 27% as Consumers Pull Back From Fresh Category, Data Shows. The Packer, August 17, 2026. Reporting Fetch verified receipt data, NielsenIQ scanner data via Quartz, and U.S. Bureau of Labor Statistics Consumer Price Index data.

15. Numerator Total Commerce Panel and Verified Voices survey, fielded July 31, 2026, reported August 2026. Panel tracks purchase behavior from 200,000 U.S. households across more than 44,000 retailers.

16. U.S. Food and Drug Administration. Guide to Minimize Biological Hazards in Ready-to-Eat Fresh-Cut Produce, finalized August 11, 2026. Federal Register notice published August 12, 2026, 91 FR 52011, Docket FDA-2018-D-3583.

17. U.S. Food and Drug Administration. Public Meeting on Listeria monocytogenes Prevention, August 18 and 19, 2026. Comments due November 2, 2026.

About the Author

Amy L. Gates

Amy L. Gates is Chief Executive Officer of Silo Technologies, a modern software platform serving the fresh produce and specialty crops industry. With more than 25 years in agriculture, she has held leadership roles across grower-shippers, global supply chains and agricultural technology, and has worked closely with the systems and processes that support food safety, traceability and compliance from field operations through packing, distribution and final delivery.

She currently serves on the Texas International Produce Association's Food Safety Committee. Her prior industry service includes the Produce Traceability Initiative's Joint Industry Leadership Council, the board of directors of the Center for Produce Safety, and IFPA's Food Safety Council.