Year 2 at Seneca Valley: Steady Production, and Nine in Ten Trays Come Back Clean
In October, CPIsight recorded every service day at the district’s tray return without adding a single step for cafeteria staff. Phase 1 set the baseline; Year 2 turns it into a monthly report the food service team can act on.
Year 2: October 2026 at a glance
The first full month of the 2026–27 school year in daily production. The system ran on every service day, and most trays came back with no leftover food.
October 2026 monthly report to Seneca Valley School District. Figures come from CPI analysts reviewing CPIsight images: 8,744 of 12,372 disposals, all 21 service days. The clean-tray rate is per tray.
Consistent, every week
The clean-tray rate held between 85% and 92% in every week of October. It was not one good day: most trays come back with no leftover food, week in and week out.
Waste is concentrated, so it is actionable
The leftovers that do come back cluster on a handful of everyday items. That gives the food service team a short list of specific menu options to try, rather than a general call to waste less.
A monthly report, measured month over month
The district now receives a monthly insights report with item-level findings and menu options to trial. Each option is measured in the reports that follow.
The challenge
Seneca Valley School District had no scalable way to measure plate waste — the essential first step toward reducing it.
No visibility into plate waste
The district knew food was being thrown away but had no data on which items, which meals, or how much. Without baseline measurements, improvement efforts had no starting point.
Manual audits too disruptive
Traditional waste audits require staff to sort and weigh discarded food by hand — a process too labor-intensive and disruptive to run at the frequency needed for actionable data.
Need for data-driven menu decisions
With rising food costs and flat reimbursement rates, the district needed objective data to guide menu planning and purchasing — not guesswork.
The solution
CPI deployed a compact, passive monitoring system at the tray return — a camera and scale that captured every tray without disrupting operations or requiring staff involvement.
Non-intrusive hardware setup
A single camera positioned above the tray return surface paired with a scale beneath it. No construction, no IT integration, no changes to the cafeteria workflow.
Automated image capture and classification
Computer vision detected and classified food items across 7 categories — including Plate/Tray, Starch, Protein, Vegetables, Fruit, Leftover Food, and Milk — without any manual input.
Weight data paired with every tray
99.3% of captured tray images were successfully paired with weight measurements, averaging 122.4g per tray — creating a complete picture of what was left behind.
Phase 1 results
Over 18 days and 7 collection sessions (~26 hours total), the system established a reliable baseline dataset with near-perfect uptime and zero operational impact. It measured roughly 335 kg of plate waste — a landfill footprint of about 0.26 t CO₂e, or 1.7–1.9 t CO₂e across the full lifecycle of producing that food.
Phase 1 baseline data collection, January–February 2026. CO₂e: landfill disposal at 0.761 kg CO₂e/kg (EPA WARM v16, food waste, net of landfill gas capture); full lifecycle at 5.09–5.61 kg CO₂e/kg (ReFED, Food Waste by Sector: Foodservice, 2024 — the range reflects ReFED’s unstated metric vs. short tons). The disposal figure is part of the lifecycle figure, not in addition to it. Water and land footprints are not estimated.
Timeline
Week 1 — System validation and initial data collection: 466 tray images captured across 3 sessions
Week 2 — High-volume collection: 2,291 tray images captured across 4 sessions, peak throughput of 289 images/hour
Year 2 — Hardware reinstalled above the dish return for the 2026–27 school year
Camera mount refined on a district in-service day, with no lunch service affected. Daily production from September 22
First full-month report: all 21 service days recorded, 88% of trays returned clean, and menu options for the food service team to trial
Monthly reports continue, measuring each menu change in the months that follow
Our goal is to prevent wasted food while making sure students are getting nutritious meals they enjoy eating.
From the blog
K-12
Does Offer vs. Serve Reduce Plate Waste? The Evidence
USDA says Offer vs. Serve cuts plate waste. A large audit dataset found more waste at offer middle and high schools. The rules, the studies, what to measure.
K-12
USDA School Meal Standards 2027: Will Students Eat It?
What the USDA rule requires for sodium and added sugars by July 1, 2027, what may change, and how to test reformulated menus for plate waste first.
K-12
School Food Waste Statistics: What the Data Shows
How much food US schools waste, which lunch components go uneaten most, and what each number actually measured: sample, year and limits.
Ready to establish your own baseline?
Start with a Phase 1 pilot. CPI’s system installs in under a day and begins capturing actionable waste data from the first meal.