- Gary Yin
In 2018, Pranshu Bansal, a Cal undergrad in CS, sat down in the university Crossroads dining hall and bit into a slice of pizza with anticipation. As he was leaving after his meal he passed the trash disposal area and saw near identical pizza slices littered in a heap. In that moment it was evident that Cal Dining lacked a feedback mechanism for waste. How were dining halls supposed to know for certain – without looking at the garbage – what items were actually consumed? After all, food taken out of the hot bar did not imply consumption, just a passing interest. Consumers, in particular hungry freshmen, may have bigger eyes than their appetites. I sat down with Paran Sonthalia, Bansal’s cofounder, to dive deeper into the company born out of food waste, computer vision, and the serendipity of inflight conversations.
The food waste narrative is hardly new. A third of all produced food is wasted, enough to feed half the world’s population. Meanwhile, each day nearly a billion go hungry. Food waste occurs at all points in the value chain from farm to mouth. A collaborative effort by the Food Marketing Institute (FMI), Grocery Manufacturers Association (GMA), and National Restaurant Association (NRA) has detailed these losses.
DeWaste serves to tackle these issues. It ran its first off-campus pilot installation in UC Riverside where Pranshu and Paran mounted cameras on the wall near the trash area. The founders trained a custom model with over 100K university images via Google AutoML (a standard convolutional neural network). According to Paran, they were able to identify insights following three weeks of training such as “there is little food waste generated by pizza on the first week of every month. Consider staffing the chef responsible for all future pizza days.” DeWaste worked with Riverside for 6-months before the pandemic hit during which they helped validate the University's tray elimination initiative, Riverside’s then flagship initiative to cut food waste.
Restaurants rarely know what is thrown away, at what time, and in what quantity...
Energized from the pilot, Paran and Pranshu struck up a conversation with their passenger seat neighbor on the flight back to campus. By happenstance they were sitting next to a director for a major supplier of YUM! Brands (owners of KFC, Pizza Hut, Taco Bell). After several follow-up meetings, Paran and Pranshu reached an agreement for DeWaste to enter into a YUM! partnership. In the subsequent meetings with Taco Bell managers, DeWaste found an opportunity to pivot into a kitchen product, one that minimizes kitchen waste before consumers can even make the order. Aptly named DeWaste Kitchen, the new hardware scans and weighs kitchen food waste to identify “behind the counter” savings. DeWaste effectively stumbled into a much larger total addressable market than the university dining hall. With 700K restaurants in the US, DeWaste market sized the waste reduction market at $9.2B conservatively. Paran shares that their customer interviews have confirmed that restaurants rarely know what is thrown away, at what time, and in what quantity. Without this visibility, restaurants tend to overproduce, failing to gather the necessary feedback to change course quickly enough. This is especially the case in fast food restaurants where food is not made to order (e.g., standard hamburger is produced regardless of customers throughout the day). With high employee turnover, subsequent training, and human error fast food solutions tend towards risk mitigation on food spoilage – much is thrown out. Quick Service Restaurants (QSRs) accumulate food waste through the entire supply chain, with kitchen and leftovers producing 40% of total chain loss (by weight of food). QSRs waste 40 billion pounds.
During the UC Riverside pilot, DeWaste received two promising grants, $15,000 from the Y Combinator startup school and $5,000 from Berkeley Haas AMP Accelerator. Subsequently they were featured on the Forward Thinking Founders podcast, providing Paran and Pranshu a larger platform to discuss the urgency of college dining hall waste reduction. Going forward, DeWaste aims to reduce QSR “behind the counter” food waste by ~50% with its kitchen product, constituting 2-8% of total food costs for YUM! Brands. They are slated to begin piloting in a few Taco Bell locations in Canada this December.
DeWaste aims to reduce... 2-8% of total food costs for YUM! Brands
Restaurants are not insulated from the explosion of VC-backed technological products. In recent years, there have been numerous attempts to introduce smarter hardware and software into the commercial kitchen. In recent months, the need for sanitary cooking practices has helped tailwind a surge of robotics for back-of-the-kitchen work (see Miso and Davinci). Some of these solutions have pursued alternative financing for more preferential terms - Miso just closed a crowdfunded Series C through SeedInvest. Meanwhile, DeWaste is not alone in implementing waste analysis through technology. Lumitics from Singapore has developed hardware with image recognition that sits on top of trash bins. Winnow mounts a CV enabled camera on the wall. DeWaste proposes that these competitors have not seamlessly integrated into kitchen workflows, requiring additional work to install, use, and maintain. This past fall DeWaste welcomed 13 new interns to their team to accelerate the development of DeWaste Kitchen. Their top priorities are algorithmic and hardware interface improvements. Keep your eyes our for the team at DeWaste and their pilot progress.
Links below for the DeWaste site, Twitter handle, and podcast.