Are you aware of Google’s FINDER algorithm? My guess is not yet but you will hear more about it in the future. Check out this article from The Daily Mail UK .
Google has created a machine learning algorithm named FINDER, Foodborne Illness Detector in Real Time, that is capable of pairing search terms like “diarrhea” and “stomach cramps” with a person’s geolocation history to determine which restaurants they have visited recently and proactively detect which restaurants might be operating unsafely in real-time.
Here are some of the interesting facts from the Daily Mail Article:
- Google ran a test across Chicago and Las Vegas in 2016 & 17.
- FINDER detected that the percentage of unsafe restaurants across those cities were 52.3% vs. the health department inspection data which said that 22.7% were unsafe.
- FINDER was more accurate than customer complaints. which are only accurate about 38% of the time. “Researchers believe this is because most people assume the cause of their food poisoning was the last place they ate at, causing them to file a complaint at the wrong restaurant. They point to medical studies that have shown foodborne illnesses can take 48 hours or even longer to become symptomatic after someone has been exposed. “
- “[We] demonstrated that FINDER improves the accuracy of health inspections; restaurants identified by FINDER are 3.1 times as likely to be deemed unsafe during the inspection as restaurants identified by existing methods,’ according to the study.”
First off, I’m so impressed by the FINDER project, because they used readily available disparate data and combined it together to identify causality and to notify the proper authorities to investigate.
Here are my conclusions from the article:
- The idea that a health inspection or an internal audit conducted periodically is enough to identify unsafe restaurants is an antiquated and inaccurate way of managing food safety.
- This further confirms the FDA’s recent study, read our blog to learn more, that the best way to reduce the risk of foodborne illness is an FSMS (Food Safety Management System) that fosters daily active managerial control.
- In the future, we will see an increase in systems that are constantly monitoring and reporting on exceptions. We do that today in our platform and we are continuing to look for ways to expand that exception reporting. To learn more about the OpsAnalitica Platform, click here.
- Restaurant operators that have been highly critical of reviews where a person said they got sick at their restaurant, there is some validity in that criticism.
- Iwaspoisoned.com, which we have lauded as a really cool platform for identifying foodborne illness outbreaks may have a very short lifespan because the accuracy of complaints is only 38% and this platform is more accurate.
The restaurant industry, like all industries, has a responsibility to operate at the highest standard and to police itself for the continued prosperity of all restaurants. Restaurants get people sick on a daily basis, causing their patron’s great physical discomfort and costing society billions of dollars annually.
The FINDER study showed that over half the restaurants in the cities studied were unsafe. People get sick at restaurants all the time but because most restaurants only get inspected 1 or 2 times per year and it generally requires a doctor’s diagnosis to make an official foodborne illness complaint, it continues to get underreported and these unsafe restaurants are continuing to operate without any regulatory consequences.
The FDA needs to mandate through the Food Code a minimum standard of daily active managerial control procedures and digital record keeping for all restaurants. We recommend that they take a Criticals First approach to this standard.
Restaurants need to implement their own FSMS that promote daily active managerial control and then work those systems to run safer operations for their own well being. I’ll leave you with this thought. 80% of restaurants fail within the first 5 years, we in the industry have always equated that to bad operations, locations, and management. Would you ever go back to a restaurant that got you sick? I wouldn’t, I haven’t. I wonder how much of the restaurant failure rate could be potentially contributed to patrons not going back to restaurants that made them sick?