Taylor Farms is one of the largest suppliers of fresh-cut produce, sandwiches, deli salads, and other products to North American consumers. The company operates 28 processing facilities that source from multiple growing regions and commodity types. In this Q&A feature, Drew McDonald, Senior Vice President of Quality and Food Safety at Taylor Fresh Foods, discusses the company's food safety efforts, learnings, and progress on its journey to achieve compliance with FSMA 204, the U.S. Food and Drug Administration's (FDA's) Food Traceability Final Rule.

FSM: What does traceability mean to a value-added produce processor?

DM: Traceability can mean a lot of different things to different people, depending on commodity and where they are in the supply chain. I suspect that a group of produce suppliers, buyers, auditors, and government officials sitting around a table would all have slightly different definitions.

My definition is broader than the Critical Tracking Events [CTEs] and Key Data Elements [KDEs] that the FDA identifies in the Food Traceability Rule. For me, in the simplest terms, traceability means that one can identify not only the source of the ingredients that are in the product, but also a whole set of inputs, process activities, growing and harvest events, etc. In other words, traceability shows the entire history of what went into making the product. From a single production code, one should be able to connect ingredient source, production line and time, wash system performance, cooling time, receiving time, receiving checks, and harvest information—including, but not limited to, harvest time, harvest crew, farm lot, etc.

One should also be able to expand beyond this to other events surrounding the variables identified, such as sanitation records for specific equipment, as well as anything else involved in producing the product. But these are not all one-to-one connections. There is a misconception that in value-added produce, one can simply scan the bag code and it will tell you what farm it came from. The fact is that not all value-added, processed products come from a single farm. The product code should allow the producer to identify which farm lots the ingredients came from. For example, a "spring mix blend" comes from a fresh-cut processing facility. The spinach, arugula, and mizuna items come from farms to the processing facility, where they become value-added products.

FSM: How does produce traceability differ from traceability for shelf-stable foods?

DM: On the one hand, the idea of traceability—using the production lot code on the value-added produce package or shelf-stable product—for both types of products contain information that indicate the production date and manufacturing location. However, the regulatory, consumer, and customer expectation of what can be derived from the value-added produce production code tends to be greater than the expectation for shelf-stable foods. For example, there's no societal expectation of readily available traceability to the wheat grown that went into the flour; but with produce, there is an expectation that the processor has visibility to the producing farms.

With shelf-stable foods, from a food safety standpoint, the original source of the ingredients is not as critical to know, assuming the shelf-stable product's ingredients have gone through a number of transformation steps that are "kill steps." Although this mindset is changing with a new focus on non-pathogenic contaminants (e.g., chemicals and metals), the expectation of traceability for fresh produce is more extensive, encompassing using traceability to reveal details on-farm, near-farm, farming inputs, irrigation, packing, harvest, cooling, and other details.

"With regard to tracking inputs, the simple answer is that most inputs are traceable and, in fact, are already traced. So that's really important… This can be done."

FSM: The "easy" part of traceability is determining where products have been sent—but how can the inputs (i.e., which growers the products came from) best be traced, particularly for commingled product?

DM: Actually, we don't always know where our products have been sent. True, it is easy for us to see where we directly sent a product, but unless third parties are capturing our lot code information each step along the way and sharing this information with us, it becomes very difficult to determine specifically where and when products have been further distributed.

Nevertheless, with regard to tracking inputs, the simple answer is that most inputs are traceable and, in fact, are already traced. So that's really important. This can be done. Now, whether it is always done by every company is a separate issue, and I definitely have a view here. One basis for this tracing stems from finances. For example, when a company buys products (inputs), there is likely transactional information involved. If this information is captured electronically, then it is easier to "trace." If it is captured via paper, it is not so easy to trace. So, the good news is that inputs are traceable, and I see opportunities to modernize. The natural question, then, is: Which inputs are important to trace? I'm not sure we know this entirely yet, and more research is needed to understand residual risk across the broad range of environmental, indirect, and direct inputs, and other factors.

Now, regarding "commingled" products, the fact is that many produce items are commingled by their very nature; shelf-stable products are commingled, too. The harvesting of many different produce items is, by its very nature, commingling. A bag of spring mix salad is commingled. A tray of cherry tomatoes picked from multiple plants could be considered commingled. Tracing the inputs to commingled rows in a harvest lot is achievable. Tracing the agricultural inputs to each lot of multiple lots inside a bag of spring mix is achievable. Knowing which lots from multiple, individual ingredients are in a finished product is achievable. This is what is important.

FSM: Why is it difficult for regulatory agencies to pinpoint the vehicle of illness during foodborne illness outbreak investigations, even with the use of advanced technologies like whole genome sequencing (WGS)?

DM: Outbreak investigations are difficult, and WGS is only a tool, not an answer. WGS only tells you whether people are ill from a common strain. It does not tell you the source or sources of the illness. Even historical reviews of the WGS in the public database may not tell you the source. They may suggest where to start looking, but this could also be a red herring.

Epidemiology is also imperfect. It works better when investigating obscure products or ingredients, but is more challenging when the products are popular and consumed frequently. For example, it's no surprise when a majority of people say they ate leafy greens. In fact, the CDC's [Centers for Disease Control and Prevention's] population study suggests that in a given week, over 80 percent of people eat leafy greens.¹ The difficulty arises in scenarios when you have a clear signal from an illness WGS, but imperfect epidemiological results that rely on individual memory going back many days to weeks, and traceback information across multiple shipment or use dates, especially when very commonly consumed items are in the suspect mix.

Furthermore, as new information comes in during an outbreak, there may be data that conflict with earlier assumptions. If these data—even a single presumed outlier—are ignored or dismissed, it challenges the value of all the data. On the traceability side, I think the difficulty is not with the traceability of the products, but how the data is being gathered and used. Over the years, regulators have learned a lot about the supply chain through outbreak investigations. This is good. While outbreaks are not the best way to learn, over the years, the produce industry has seen more and more germane questions that consider key tracking events along the lines of planting, growing, irrigation, harvest, cooling, packing, processing, etc.

However, there is an opportunity now to use existing knowledge of distribution points to improve the efficiency, accuracy, and effectiveness of an outbreak investigation. The approach right now tends to identify the larger player in a particular market. Odds ratios and case control analysis do not seem to always be considered. I believe there is a significant opportunity to use traceback/trace-forward information in a different way, with FSMA 204 coming. A goal should be incorporate the traceback/trace-forward information as part of the hypothesis-generating process, rather than using it to confirm the hypothesis.

Separate from WGS and traceability, it is important to mention that industry has seen the inefficiencies and challenges that exist during outbreak investigations because of the jurisdictional divides among multiple federal and state agencies. As a result, it can be difficult for one agency to take a holistic approach to an outbreak. Once CDC identifies the potential food source(s), that impacts whether FDA or USDA [the U.S. Department of Agriculture] investigates.

FSM: Under what circumstances is traceback the most useful to an outbreak investigation?

DM: Traceback should never stand alone. However, there are many circumstances where traceback is useful. For example, in a produce-related outbreak investigation where a suspect product was harvested and packed in the field with a label that is specific to that lot, when that product tests positive for the relevant pathogen in a targeted marketplace sampling event, then that is useful information for the investigation. Traceback could also help rule in or rule out a particular food commodity or supplier. For example, a product may be a suspect food vehicle or come from a suspect supplier, but one or more of the matching WGS illness cases did not involve consumption of or exposure to that product. In such cases, the product may not be the food vehicle. Traceback can also reveal whether that product lot was in the marketplace at the time, or not ever.

"If traceback slides away from an investigative, iterative process toward a confirmatory biased approach, then it becomes less useful."

FSM: What limits the utility of traceback in an outbreak investigation?

DM: The most obvious limitations are when information is incomplete or does not match other records. For example, a company may have headquarters at one address, but the actual shipping point or growing location is at another address, which may not be captured properly. Additionally, there may be middle, transactional points, such as a broker or other entities that do not physically handle the product, that may not show up in paperwork or electronic traceback records. These details need to be understood and considered in any traceback analysis.

A less obvious effect on traceback utility is ignoring or dismissing any traceback data that contradicts or does not match assumptions in the investigation. If the data shows that a particular product was not available for consumption, then this should not be ignored. If traceback slides away from an investigative, iterative process toward a confirmatory biased approach, then it becomes less useful.

The last thing I will mention that limits the utility of traceback is when there is a data request for information that would not be known by that particular point in the supply chain. For example, asking the retail outlet to identify the farm or shade house where cherry tomatoes were grown is not something the retailer would know.

FSM: Regulators look to traceback as the primary method for solving foodborne illness outbreaks, but traceback alone cannot provide all the answers. What other methods and technologies should be considered?

DM: Regulators talk about the three-legged stool: epidemiology, traceback, and laboratory/testing data. Today, CDC identifies the food and, depending on the food, FDA or USDA tries to find the source with tracebacks. Traceback data should be a primary tool, but we need to look at ways to use it differently. For example, maybe there is a way to better conduct the food history interviews simultaneous to gathering marketplace supply chain information. Either way, whether as a result of FSMA 204 or with the current sortable datasheet requirement, FDA will get data faster.

What are the tools we can use to analyze it efficiently, thoroughly, and effectively, to identify the source and get to root cause? As the breadth of industry data are digitized, we should look at new opportunities to illuminate supply chains to help quickly rule out distractions. Traceability needs to focus on the signal, not the noise.

FSM: What initiatives is Taylor Farms undertaking to make its food safety and traceability systems stronger and more robust?

DM: Our focus has always been on prevention, and I'll share more detail shortly about the numerous steps we've taken, and the innovations we've developed, to reduce the chance that contamination will be associated with our products. We already have robust traceability systems, and our biggest effort has been to standardize our traceability report across all sectors of our operations. We have recognized that although no two traceback requests are the same, the data requested is pretty standard. We have FDA and USDA product lines, and although the regulators require slightly to sometimes significantly different data points, we have identified some common reporting methods.

We also leverage the GS1 system of standards. Our goal is to capture data along the way, at discrete points, to allow us to be as specific as possible. This effort has resulted in considerable leveraging of technology that eliminates human input error. Two-dimensional scanning, vision technology, paperless data capture, and extensive automation are all tools being deployed. The questions answered by the standardized reports are:

1. Which product(s) was involved? When and where was it produced?
2. Which ingredient(s) was involved? When and where was it produced?
3. Where else did these product(s) and ingredient(s) go? It's always a traceback and a trace out.

"Our culture is one in which we share our learnings and insights and actively convene competitors to tackle food safety challenges head on, for the good of our collective consumers."

FSM: What new food safety standards and practices has Taylor Farms implemented over the past decade, and what has provided the impetus for these efforts?

DM: Wow, this is a big question! We know that consumers are more likely to eat a salad processed by Taylor Farms than any other brand, and with our increased product portfolio—producing not just fresh-cut produce items, but also sandwiches, deli salads, and more, out of our 28 processing facilities sourcing from multiple growing regions and multiple commodity types—we know that our customers and consumers rely on us to consistently deliver safe food. This amounts to nearly 265 million servings per week! That's a lot of pressure. Thankfully, our leadership is interested in and very committed to food safety innovation. Our culture is one in which we share our learnings and insights and actively convene competitors to tackle food safety challenges head on, for the good of our collective consumers.

We have a long list of standards and practices that have been implemented over the years. I will highlight a few:

• In 2006, Taylor Farms opened a pilot plant and dedicated research facility in Salinas, California with the sole purpose of improving food safety and eliminating outbreaks caused by foodborne illness. In this same year, we implemented pre-harvest pathogen testing on all commodities.
• In 2008, Taylor Farms pledged $2 million to co-launch the Center for Produce Safety (CPS), an organization that fosters aggressive food safety research and provides the industry with updates on research outcomes. We continue to invest in CPS at the highest level—millions of dollars—and I've personally served as Chair of the Technical Committee. We use the results to improve our programs.
• In 2009, we began deploying an advanced cleaning system² across our facilities. Antimicrobials are not a "kill step" for produce, but they prevent cross-contamination. We pioneered an innovative chemistry and monitoring system to maintain the highest levels of control.
• In 2010, USDA's Agricultural Research Service (ARS) completed its validation of the advanced cleaning system's² wash aid that fully mitigates water cross-contamination of bacterial food pathogens. Today, it is a separate business² that supports a breadth of fresh produce processors, including many of our competitors.
• Three years ahead of industry, in 2016, Taylor Farms required that surface water be treated prior to use for overhead irrigation of vulnerable crops.
• In 2017, Taylor Farms worked with an equipment manufacturer to revolutionize the sanitary design of spin dryers, which we've incorporated across our enterprise.
• Taylor Farms worked with trade groups, industry competitors, and regulators to establish guidelines in 2019 for provenance labeling on salad bags and shipment records so that consumers could readily identify the production region of romaine blends in the event of an FDA advisory.
• In 2020, the Romaine Rally Research and Harvester Sanitation task force was spearheaded by Taylor Farms along with Commercial Food Sanitation, third-party academic researchers, FDA's then-existing Center for Food Safety and Applied Nutrition (CFSAN), and industry partners. This research was conducted four years ahead of the majority of the industry.
• In partnership with FDA and CDC, in 2023, Taylor Farms worked on industry-wide research and standards around pre- and post-harvest field assessments for Cyclospora cayetanensis on lettuce tissue.
• Lastly, in 2024, we began deployment of 100-liter irrigation waste sampling. Most of the industry tests at 100 milliliters.

FSM: What are some ways a firm can prepare for compliance with the Traceability Rule?

DM: This is an important question. Generally, the most common experience with traceability that companies have may be from mock recall work. This is a good start, but my recommendation on how to prepare for traceability is to work through real-world, complex scenarios. Mock recalls are useful for basic traceback scenarios, but actual traceback requests may be more complex. As such, a first step in preparation is to develop several scenarios to test. 

For example, a very complex, but all-too-common scenario is this: Trace records for a two-week shipping date range from one month ago for products A, B, and C (i.e., pick products that are similar in that they have 2–3 common ingredients, but are not the same product), and provide information on where these products shipped and what other products used these common ingredients, as well as their respective ship-to locations. Provide these data in a sortable, electronic spreadsheet. 

Additionally, for each code date, provide the supplier and receiving dates for the raw ingredient sources for the named products, along with all processing records. This is a real-world type of request. The best way to prepare for this is to really test your system. Coordinate with your IT group and other parts of your operation so that everyone is prepared. In my organization, we have included some suppliers and customers in these exercises, which has resulted in very useful learnings.

In short, the key elements in preparation are real-world scenarios and cross-functional groups. At this year's Food Safety Summit in May,³ speakers will be conducting some tabletop traceback exercises that include these important elements. Similar exercises have been conducted at other industry meetings and have generated useful discussions and awareness. 

Acknowledgment

The authors thank Janet Riley of Janet Riley Strategies, Andrew Clarke of Loblaw Companies Ltd., and Mark Beaumont of Danone for their insights shared at the “Food Safety Culture and Communication” session at the 2024 Food Safety Summit.

References

1. U.S. Centers for Disease Control and Prevention (CDC). "FoodNet Fast." https://wwwn.cdc.gov/Foodnetfast/PopSurvey.
2. SmartWash Solutions. "A revolutionary advancement in food processing." https://smartwashsolutions.com/.
3. BNP Media and Food Safety Magazine. "2025 Food Safety Summit." To be held May 12–15, 2025 in Rosemont, Illinois. https://www.food-safety.com/food-safety-summit/.