The food industry often points to the aviation model when talking about how we can work to improve food safety. Specifically, many articles refer to the National Transportation Safety Board (NTSB) and the Federal Aviation Agency (FAA) as the gold standard for data sharing, process improvement and public safety confidence. And on the surface, it works and is a wonderful parallel for what we need in the food industry for improved food safety. There is neutral oversight, confidential reporting, industry-wide learning and a cohesive commitment to learning from failures to create one of the safest systems in the world.
Despite that, we’re often focused in our discussion on only part of the model, and we routinely gloss over one of the most critical reasons aviation safety has improved so dramatically. It’s not just the sharing, data governance and learning from failures. It is the significance of one little black box.
In aviation, every incident is backed by continuous, high-resolution operational data. Flight recorders (black boxes) don’t just capture failure, they capture streams of data about the entire system. With that, we capture the critical information that clearly defines normal operations and that leads to the context necessary for offering insight into anomalies.
What happened before, during and after the failure is just as critical as the failure data points themselves. The black box captures baseline of normal operations so that when failures occur, the deviations can be identified, evaluated from all angles and eventually understood well enough to be shared and preventative for the future.
Aviation didn’t just build a system to share data. It built a system to generate the right data. Food safety generally hasn’t done that.
We’ve built robust compliance systems, audits and testing programs. And increasingly, we’re exploring data sharing platforms to speed our efforts on efficient food safety. But most of what we collect is episodic, fragmented, and designed to answer one question: “Did you meet the requirement?” and not the question of “What actually happened?”
When an outbreak occurs, we reconstruct the past with incomplete information. Environmental monitoring results may exist, but without context. Water data may exist, but without linkage to weather, hydrology or upstream conditions. We may know a pathogen was detected, but rarely enough around that positive to know why it was there and even less about the system dynamics that allowed it to persist or move within an area.
That’s not a black box. That’s a snapshot of compliance data, and at best, infrequent data around a presumed failure that we then look to explain post-event. Snapshots are rarely (if ever) able to explain failure in complex systems.
As it stands now, food safety is often trying to adopt the sharing model of aviation without first building the data foundation that makes sharing meaningful – the baseline information captured in the black box.
If your system isn’t capturing continuous, context-rich data during baseline operations, then sharing it won’t materially improve risk understanding. So the question isn’t only, “How do we share more data like the NTSB/FAA model?”. The additional question we need to be asking ourselves is “How robust is our black box?”
Are you always capturing environmental conditions in real time? Logging all your activities, equipment and processes to know what normal is? Building depth around inconsequential data elements (aka – negatives)? Detecting signals of increasing risk before they become failures?
Without enough baseline data on the normal system, it becomes nearly impossible to distinguish normal variability from emerging risk. In the absence of baseline data, outbreak and failure investigations become educated guesswork. “Root cause” becomes what we can explain, and/or presume and not what actually happened.
Aviation didn’t reduce accidents by auditing pilots harder. It reduced accidents by instrumenting the system and by building a data layer that is always captured in the black box. That has made failure legible and actionable to be preventative.
It’s time for food safety to build the black box.