Preemptive Freshness Management: Part One--Empowering Workers to Improve Delivered Freshness
on Mar 14, 2017
Preemptive Freshness Management takes a proactive approach to reducing waste and spoilage in the produce supply chain, from grower to retailer. A process model that incorporates product characteristics, customer requirements, and resource constraints--informed by continuous situational awareness--provides field, packhouse, and DC workers with early warnings, clear guidance on goals and tradeoffs, and prescriptive corrective actions to maximize freshness and minimize loss.
Full Article Below -
( This article is excerpted from the complimentary report Preemptive Freshness Management Empowering Workers to Improve Delivered Freshness,
available for download here )
Most approaches to managing produce freshness today are forensic and reactive. That is, they rely on after-the-fact forensic evidence, typically a strip-chart temperature recorder or other vehicle-level monitoring of temperature in the reefer. This evidence may help the retailer (somewhat)1 in deciding whether to accept or reject a load, but it does little to prevent or reduce spoilage in the chain up to that point. Similarly, at the various process steps in the cold chain—from harvest to pre-cool to storage to distribution—decisions are generally made on a reactive basis, after it becomes apparent that things are not working to plan.
Rather than waiting until the fire starts and then frantically trying to put it out, a preemptive approach systematically takes measures to prevent the fire in the first place. A preemptive freshness management system continually monitors field, pack house, DC, and store operations, thereby knowing much earlier when there will be capacity constraint. It then makes more intelligent decisions on prioritization and sequence, to maximize freshness and minimize loss. This approach results in vast improvements compared to the on-the-fly human judgment calls typically made in produce operational settings.
Empowering the Worker
Workers in the produce cold chain—from the field, to packhouse, to distribution center (DC), to the retailer—make the best decisions they can, every day, based on the tools they have and the goals they are being measured against. This often leads to suboptimal results because the tools are rudimentary, such as just knowing the time and temperature in a pre-cool unit, and goals are simplistic, such as ‘you can go home as soon as all these cases have been shipped.’ To empower smarter decisions, a system can be implemented that gives the worker three things:
Early Warning Alerts Example
One example of an early warning notification is the way Zest Fresh analyzes outstanding current purchase orders (POs) for a grower and quickly determines how to best allocate the current harvest against demand. Typically, a grower will slightly undercommit, but actual actions and conditions in the field and yard—such as pallets not coming in from the field as quickly as planned, or too many pallets coming in at same time—can create bottlenecks, congestion, and slowdowns during the day. These bottlenecks can erode the grower’s ability to meet their freshness commitments. As operations get congested and backed up, workers start making tradeoffs, such as processing the closest pallet first, even when other pallets have been waiting for pre-cool longer. By analyzing the current situation against outstanding PO commitments, Zest Fresh can generate alerts and recommend corrective actions as needed.
Early warning of upcoming issues—The system should alert the workers, with as much advanced warning as possible, whenever impending issues are expected to prevent them from processing the produce in the proper, optimal way (i.e. it alerts workers and supervisors ahead of time if produce will be late going into pre-cool, will not be cooled adequately, operations are going to back up and fall behind, and so forth).
Clear guidance on goals and tradeoffs—Helping the worker to clearly and easily understand which customers, products, and shipments have priority. Instead of handling cases in strictly FIFO2 or ‘whatever is nearest to me’ basis, the worker can take into account the priorities and tradeoffs. For example, perhaps that shipment of romaine lettuce has three more days of travel or requires more days of freshness than the iceberg lettuce shipment. Without the system telling them that, the worker has no idea. Armed with that knowledge, the worker may pre-cool the romaine first, even though the iceberg is at the ‘front of the line’ for precooling.
Prescriptive corrective actions—Ideally, the system itself will be able to do a ‘what-if’ analysis, weigh the trade-offs, and propose the optimal preemptive or corrective actions to the worker. The system should make it easy for the workers to make the right optimal decisions and take the right actions every time, without thinking. It is impractical for a worker out in the field, at the packhouse, on the truck, or at the distribution center to stop what they are doing, run an analytics program, and then consider the tradeoffs on various choices. It is much better if they are simply instructed what to do next, even while the system is providing workers and supervisors with unambiguous clarity on what the goals are, current progress towards meeting those goals, and alerts when things are not going to plan.
Timeliness is a critical aspect of empowering smarter decisions. Workers can’t be sitting waiting for answers from the system before they can do anything … or worse yet told after-the-fact what went wrong, when it’s too late to correct. Alerts, guidance, and prescribed actions need to happen in near real-time, as the work is actually happening, to effectively guide the work and avoid waste.
Recommending Corrective Actions Based on Current Situation and Requirements
To make intelligent recommendations, the system needs full situational awareness and contextual knowledge.
This includes A) current conditions and their impact, B) product requirements, C) management preferences,
and D) physical or resource constraints.
Conditions and Their Impact
Examples of the types of conditions the system needs to be aware of include the temperature at harvest time, expected temperatures later in the day, temperature the night before, whether or not it rained recently,3 and how
many days before or after peak maturity the product is being harvested. In addition, the type of produce being harvested greatly affects the impact of all these variables and other aspects of freshness. If the system does not take into account all of these conditions and their impact on operational decisions, then it cannot prescribe the optimal actions to workers.
At any given time, the grower’s operations (field, precool, packhouse, shipping) has a queue of orders they are fulfilling. Within that queue, different shipments have different requirements; in part because some buyers require longer shelf life at delivery, while others (such as a restaurant or food processing plant) are OK with shorter remaining shelf life. Even the same buyer may have some slow-moving and some fast-moving stores, where the slow-moving stores require longer shelf life. Another important part of the requirements equation is the transit time from grower to customer. All these requirements need to be taken into account when prioritizing and optimizing the steps in the operation.4
Some customers will be given higher priority than others, perhaps because they are more strategic, have higher volumes, larger life-time revenue for the grower, or other considerations. In any case, the system needs to allow the grower to enter these kinds of adjusted weightings and management preferences, including the ability to specify rules, such as ‘Customer A is more important than Customer B, except when the order for Customer B is a special promotion.’
The system must have knowledge of resource constraints in order to make the tradeoffs, and recommend actions that are feasible and can actually be executed. Constraints include things like the capacity and rate of transport from the field, storage space in the yard and packhouse, pre-cool units’ capacity, number of trucks that can be loaded at once, hours of sunlight, and so forth. This also includes fluctuating variables such as the amount of labor available or transit time from various fields to packhouse. Things like pre-cool unit capacity vary depending on the type of produce, as well as external conditions, such as field temperature. Without this knowledge, recommendations are meaningless, as they actually may not be feasible to execute in the real-world operations. Furthermore, the solution must be able to calculate overall capacity and throughput in order to forecast whether or not demand can be met, and thereby generate alerts that provide alternatives, to optimize processing and allocation in an overcommitted, under-capacity situation.
In part two of this series, we cover how produce cold chain processes are mapped, how those maps are used to do predictive analytics, and the need for solutions at each process step.