Throw-A-Weigh

Throw-A-Weigh is a smart cart technology that measures, manages, and reduces food waste for grocery stores.

Our solution is to minimize food waste through data, by addressing root causes of why food goes to waste and diverting food that is still edible from compost and landfills to donations and sales. Throw-A-Weigh is a data collection and analysis tool comprising both hardware and software components. The hardware is a no-touch smart cart that recognizes produce with AI and categorizes it into donation and compost through a series of sensors, saving time and energy for what would’ve been a very manual data collection process. The system collects data such as weight, PLU code, gas emissions, etc and provides classifications for why the produce is being thrown away and quantifiable remaining shelf life. Then, the software analyzes this data and provides instantaneous reports with the results and recommended action items. This information can help grocers classify food as still sellable and additionally prevent future waste through understanding why some of the current produce went bad. By providing accurate, real-time data, we are empowering produce managers to make data-driven decisions and filling the knowledge gap for food waste reduction strategies.

Food waste is a big problem - up to 160 billion pounds of food in the U.S is wasted every year. It occurs at all stages of the supply chain, ultimately leading to 30-40% of all food produced going uneaten. Just focusing on produce alone, grocery stores in the US lose up to 5.5 billion pounds of wasted produce every year, equivalent to $16 billion. We’re tackling the 13% of food waste occurring at the grocery store level for systematic change across the 40,000+ grocery stores in the U.S. The impact of food waste not only cuts into the slim margins of grocery businesses, but also is a lost opportunity to solve hunger/food insecurity, conserve resources, and mitigate climate change.If food waste was a country, it would be the third-largest emitter of greenhouse gases, after China and the U.S. Food waste is also the single largest category of material (24%) in landfills, where landfills contribute 20% of total U.S. methane emissions. According to Project Drawdown’s modeling of “plausible and economically realistic” climate change solutions, reducing food waste is the #1 solution when it comes to reducing carbon dioxide emissions (projected 87.45 gigatons CO2 reduced between 2020-2050). With the urgency of climate change and the ever-growing necessity to avoid the 2°C global temperature increase, we believe it is imperative to implement any and all solutions that will slow down greenhouse gas emissions.

Our target customer is the corporate buyer of retail grocers. They work at a high level, making purchasing decisions for multiple stores. These buyers care about price points and savings and need accurate data to analyze costs, sales, and projections. Not only will our solution improve their sustainability efforts and increase tax benefits by providing granular data on donated items, we will also help to better the quality and quantity of their donations to feed more people in need. Our end users are produce managers who make purchasing decisions for the store they are in charge of. They care about their floor employees and store performance in terms of efficiency and slim margins. By default, produce managers carry the responsibility of reducing food waste for their store, which can be difficult without knowing the what, where, and why. Grocers struggle with collecting shrink data because they are short on labor and time. Floor employees also dislike the tediousness of measuring and logging shrink, leading to low adherence. By automating the data collection and analysis processes and delivering the insights produce managers have been looking for, Throw-A-Weigh provides the information without adding on labor or time for employees. Grocery stores expect high turnover in their products. While their biggest goal is to simply sell more product to make up for the lost margin on the wasted product, Throw-A-Weigh offers the opportunity to sell more of the product they have already ordered and have in their inventory. 

Through more than 60 customer interviews we have conducted at over 50% of the grocers in Austin, TX, each store shared that fresh produce waste is a problem for them, with the average medium size store budgeting for a $7 million loss annually in fresh produce alone. We have been conducting a concierge pilot testing model at Wheatsville, an Austin Co-op, for the past 4 months to collect data on their food waste and providing reports and recommendations to gather more user feedback on our solution and process. Additionally, we have been in contact with the Emerging Technology and Sustainability team at H-E-B, a Texas-wide grocery retailer, for future pilot testing relationships. We’ve also had conversations with the Central TX Food Bank and Save the Food, an Austin food rescue non-profit, and discussed the opportunity to increase the quality and quantity of the donations they receive, as currently a portion of all donations is inedible and must be sent to compost instead.

Our tech team is continuing to build a minimum viable product, which has separate functioning hardware and software parts that we are refining and integrating. In terms of hardware, we have cameras and weight load cells incorporated into the Raspberry Pi for testing. We currently have an image recognition data model that recognizes 5 different produce items and identifies the reason why they were thrown away at 83% accuracy. We are working towards increasing the depth of our scope to identify apples and their throw-away reason codes to 90% accuracy. At the same time, we are running a concierge model pilot with Wheatsville to test the value of the data and hone in on their pain points. To date, we have collected and analyzed over 14,130 pounds of wasted fresh produce at the store, equivalent to $58,823,18.

Our solution utilizes AI image recognition to identify the produce and the reason why it was taken off the shelf through a camera and machine learning models to provide a feedback loop for grocers to understand the cause of their food waste. We will also incorporate biochemical sensors to detect ethylene gas emissions to quantify remaining shelf life of produce items to improve quality of fresh produce offered to the day-to-day shoppers and underserved communities in need. The physical device automatically sorts the produce item in a donation or compost bin depending on the shelf life of the item, avoiding further touch points that would be time consuming and potentially damaging the item. Our device is mobile and adaptable, where a single unit can be used in many stores just as easily as one store can use many units. While our AI technology allows grocers to improve waste measurement to fit the scale of current stores, our software can support any increased analytics a growing store may need. For our product to first start meeting increasing demand, our team will begin with in-house production of several units for pilot testing and pre-sales, after which we will transition to locally sourced materials and manufacturing.

There are 40,000 grocery store locations in the USA alone and over 400 in Texas alone. We currently serve one grocery store chain with two locations and are in conversations with another chain that has 300+ locations in the state. We expect that we will be able to begin a pilot with one or two of their stores within the next year. We are also discussing a potential partnership with a food rescue non-profit to share data and sort produce between donation and compost for them.

We would like to save our current client $177,000 in the coming year and more importantly divert 42,000 pounds of food to feed more people, equivalent to 14,000 meals recovered. As roughly ⅔ of this food is still edible when it is taken off the shelf, our goal is to help our client sell or donate this edible food through quantifiable fresh produce shelf life and automated donation sorting system. As grocers will remain incentivized to continue donations for tax benefits, the food rescue non-profits and food banks will receive a larger quantity of quality donations, minimizing the amount of produce that must be sent to compost and landfills. By achieving this goal, we would help reduce CO2 emissions by 105,000 tons.

We will measure our progress by the amount of food waste in terms of weight and dollars, reduced CO2 emission, and meals recovered by diverting food from trash and compost to sale and donation. This information will come from our detailed data collection through our concierge and prototype pilot tests and later be part of the reporting for our product. We plan to achieve this by increasing the quality and quantity of donations. 

Our current barrier is high turnover in our current client’s leadership. We have already begun developing other potential client relationships and will likely focus more on the newer client relationships in the coming year. As this is a largely networking based field, it takes us longer to develop relationships as outsiders with technical knowledge rather than grocery expertise. We also have financial constraints that are limiting the amount of contribution we can give to hire more experienced engineers to bring to the tech team.

Our team has diverse and complementary skill sets, technical backgrounds, interests, and strengths. With a combined knowledge of computer science, molecular science, engineering, and data science, our team is well equipped to understand and track the detail that goes into producing diversion. Becky’s experience as a computer scientist has allowed us to set up our cloud-based data pipeline and also to create data scraping and automation tools that help both us and our customers. Rosaline’s experience in system design and data science has allowed us to stay organized and to create goal oriented processes. Paige’s experience in biological sciences and project management has allowed us to identify nuanced ways to collect data for grocers and has helped us expand our team. These backgrounds help us understand the science of food waste and why it is difficult to store perishables. Our experiences allow us to cater to grocers' need for a balance between collecting higher volume and quality of data and saving time and money.

We are currently partnered with Wheatsville, a local grocery co-op in Austin, TX, for piloting our solution. In this pilot, we measure their produce waste every single day and deliver reports to their team. Through the pilot, we are also learning about the grocery’s operations, data management infrastructure or lack thereof, and product-market fit.

N/A

Our mission is geared towards solving the environmental issue of food waste and the socioeconomic challenge of food inequity. Our solution can be applied locally within the community and scaled up to global reach. The prize would be used for hardware and software expenses as we test materials and technology in product development, and for recruiting talent.