Server Farm to Table

If we know where our fresh food comes from, will we believe that it’s really “fresh”?

A sticker next to the label on my package of raspberries reads, “See where and when I was grown. Enter the code at” An arrow on the sticker points at a sixteen-digit alphanumeric code on the label. I’ve learned to ignore stickers on food—so much visual noise accompanying phrases like “New look! Same great taste!” and “10% more. Free!” Normally I would have lumped this particular code in with the many bar codes, QR codes, and other machine readable gibberish that litter products today. But as a researcher focusing on agriculture and technology I took the bait, visited the website and entered the code.

A HarvestMark label (HarvestMark)

The results were underwhelming. I noted that the raspberries’ “safety status” read “no issues reported.” Then I felt naive for assuming that the food I bought was determined to be free of “issues” before it was put on the shelf. Was food safety now the responsibility of the consumer? What happened to the USDA, the FDA, and the A&P?

I realized I’d seen the same type of code before, remembering a package of mushrooms I’d bought a few days earlier. I entered that code into HarvestMark too. Again, “No issues reported,” but this time I was given some additional information—their country of origin, state of origin, and packing location. The code from a package of spinach let me know that it had been grown in either Imperaial (sic) County, California or in Yuma or Maricopa counties, Arizona.

How was I supposed to interpret that information? What is my experience as a customer supposed to be? The industrial food supply has long relied on cultivating consumer assumptions of safe, “fresh” food. Revealing even a part of the circumstances of food production can be a contradiction. First, in informing me that “no issues [have been] reported,” HarvestMark is making me aware of the possibility of issues. It’s a declaration prompting questions which wouldn’t have been asked otherwise—like someone selling me a car and telling me that no other owners have had the brakes fail yet. Second, in letting me know where and when my food was grown, HarvestMark is inverting the longstanding relationship between food producers their presentation of the idea “freshness,” a presentation which has often relied on obscuring the details in order to present an ideal. Products may be “farm fresh,” after a refrigerated journey from a site thousands of miles away.

A sample trace of yellow onions (

I spent a month or so researching and testing similar systems, including websites and smartphone apps such as FreshQC, Find The Farmer, and Trace Produce. FreshQC turned out to be a market research dupe. After I entered an alphanumeric code, the website rewarded me with a series of unsolicited multiple choice questions about the quality of my produce, but no information about its quality or origin. Find the Farmer presented more information than other systems, but was limited to producers of grain used in a particular brand of flour.

Data about “food objects” is recorded and analyzed within meat, dairy and produce supply chains to assist with product recalls. This can make it easier, for example, to trace E. coli infested meat back to its point of contamination. As I dug deeper, it became clear that consumer-facing food traceability services were just an attempt to repackage that data and present it in a way that contradicted about a hundred years of food marketing practices.

Most mass-produced fresh food items progress through a variety of steps between farm and table. An apple may grow at a particular tree at a particular location before being picked, processed and packed. Then it may be stored and shipped from and to different locations before being sold by a retailer. Some of these steps can significantly impair a food’s quality. Traceability—the name given to the process HarvestMark and similar services offer— attempts to make all of the steps between farm and table searchable, to aid in identifying and addressing problems after they happen.

Traceability does for food what FedEx or UPS package tracking software does for freight: using a unique tracking number, you are able to determine where something has been, what conditions it has encountered, and where it is going next. Precision and scope vary significantly within existing food traceability systems. While the USDA characterizes the capacity of United States food producers as “enormous,” they add that “Some traceability systems are deep, tracking food from the retailer back to the farm, while others extend back only to a key point in the production process.” Of course, there will be significant variation in traceability because the types of foods produced and their production processes vary. Bacon is produced and processed differently than apples (thankfully).

What should be recorded anyway? Have a look at the choices presented to you in an urban coffee shop to get a feel for the sorts of issues at play here: decaf, fair trade, shade grown, organic. Different attributes can be associated with transformative points in the supply chain. For instance, whether or not coffee is decaffeinated is dependent on processing, but whether or not it is fair trade is dependent on the relationships between growers, wholesalers, and retailers.

A web-based HarvestMark report for baby spinach. "No issues reported." (HarvestMark)

Beyond such product specific variability, some systems are very precise, tracking food products to the minute of production or the exact area of a field where they were grown. Others are less precise, tracking products to farms in a large geographical area, such as the area served by a single grain elevator. Some traceability systems collect and track information on a broad range of attributes, while others track only a few.

Traceability addresses a problem of modernity: the invisibility of food infrastructure and its effects. In a small farming village, a family may know where all of their food comes from because the possibilities are small: they either produced it themselves, or traded it with other townspeople. Only after refrigeration, new modes of transportation, and new agricultural processes were developed did the origins of our food become opaque. Even then, many transactions occurred without any records, as people relied on handshakes and interpersonal relationships.

It would be incorrect, however, to identify traceability as a completely new phenomenon. The Perishable Agricultural Commodities Act (PACA) of 1930 mandated record keeping that included “produce lot identity and all transfers and transactions between shipping point and destination receiver.” Recent initiatives add computation to the picture. So, traceability today is more than the recording of information about “fresh” food—that recorded information can be searched, shared, and put to use in automated decisions or processes. For example, if an E. coli outbreak in one part of the country is determined to result from tainted spinach, the spinach could be traced back to determine if other tainted batches had been shipped elsewhere.

The availability of traceability information via services like HarvestMark inverts our longstanding relationship to the idea of freshness. Traceability can expose the distance that food objects travel and the circumstances of their production. While folks at a supermarket in New Jersey probably know that their mangos aren’t local, exposing the transportation infrastructure’s role in food delivery and instilling the consumer with the idea of freshness aren’t exactly well aligned goals. With meat, dairy, and produce, provenance—whether bona fide, or implied by a picture of rolling fields on side of a package of butter—is, in many ways, the brand. Consumer access to traceability data may yet constitute a fundamental change in the way we think about (or mythologize) agricultural production.

This post is based on research from the Intel Science and Technology Center for Social Computing