A client asks what you did to bring catering emissions down this year. “We sourced locally where we could” is the answer most event teams reach for, and it is the kind of answer that sounds like progress without being able to prove any. Local sourcing is a real lever, but it is a small one next to the two decisions that sit directly inside a caterer’s brief: what goes on the menu, and how much gets produced against a headcount that was never going to be exact.
Both are measurable. Both are decisions an organiser makes or negotiates, not something that depends on an attendee’s personal choices the way travel does. That combination, large and controllable, is what makes them worth getting right before the next menu goes to print.
Catering sits in the band you actually control
Ranking event carbon reduction levers by degree of control puts most categories into one of three bands: decide, negotiate, or influence. Attendee travel mostly falls into influence, because you do not book anyone’s flight. Catering falls into negotiate. You do not cook the food yourself, but you set the menu, the guarantee terms, and the production brief the caterer works from. That is a stronger position than most teams use it for.
The GHG Protocol’s guidance on Category 1, purchased goods and services, backs this up from the measurement side: for most food-service operations, more than 90% of emissions sit upstream, in what was grown, raised, and shipped before it reached the kitchen, not in how the kitchen ran that day. The two decisions that determine what got grown, raised, and shipped in the first place are exactly the menu and the volume ordered against it.
Lever one: what’s on the menu
Protein choice is the single largest swing available inside a catering brief, and the gap is not close. Producing a kilogram of beef generates roughly 60kg of greenhouse gases; a kilogram of peas generates around 1kg, according to the meta-analysis of 38,700 farms across 119 countries that underpins most food-carbon comparisons in use today.
At the scale of a full event menu, that gap compounds. A 2019 Center for Biological Diversity study of catered event menus found that switching from conventional to mostly plant-based options cut the carbon footprint of breakfast, lunch, and dinner menus by 60% to 85%, alongside habitat-loss reductions of 21% to 93% and water-use reductions of 46% to 72%, depending on the meal. No other single catering decision comes close to that range.
This is not an all-or-nothing swap. A menu that shifts the default plate toward plant-forward options, with meat available rather than assumed, captures most of that reduction without turning the brief into a negotiation about whether meat is served at all. The point is that this is a specification decision, made once, at the menu-approval stage, rather than something that depends on what any individual attendee happens to choose at the buffet.
Lever two: how much you cook for a guarantee that never shows up
The second lever is less visible and, on current evidence, at least as large: how much gets produced against a headcount nobody is ever fully certain of.
A 2026 pilot study from WWF and ReFED, covering 20,000 meals across six US business events, found that 42% of food prepared for buffets went uneaten. Of that uneaten portion, 29% still met the standard for donation; the remaining 13% went to compost or landfill. One plated dinner in the same study had a banquet event order calling for 1,200 entrées; the kitchen produced 1,440, a 20% overage built in as a buffer against a guarantee that was always going to be approximate.
The same study found the fix is neither exotic nor slow. Right-sizing serving vessels, tightening the special-dietary count against actual registration data rather than a “just in case” margin, and cutting sauce and dressing portions by half for a 1,000-person lunch, saved 72 pounds of food and just over $1,000 on that one course alone. Applied across the targeted categories, the pilot sites reported a 55% reduction in waste, and 80% of participating planners said they were very likely to keep the practices after the study ended.
Extrapolated across the roughly 1.9 million business events held in the US each year, the study estimates the national cost of matching its baseline waste rate at 55,000 tons of food and more than $330 million, a number that exists because the guarantee-to-production gap is treated as normal rather than as a line item anyone owns.
Overproduction has a carbon cost independent of what happens to the surplus afterwards. Every uneaten plate still carries the full upstream footprint of growing, processing, and shipping the ingredients on it, whether it gets donated, composted, or landfilled. Donation and composting reduce the disposal impact. Neither undoes the production emissions already spent making food nobody ate.
Why “we sourced locally” doesn’t answer the question
Local sourcing addresses transport emissions, which are a real but comparatively small share of a plate’s footprint next to what was grown and how much of it got made. It is also the easiest catering claim to make without any data behind it, because “local where we could” does not specify a percentage, a supplier, or a baseline it improved on.
Menu composition and production volume are different. Both can be specified in a contract before the event, both can be measured against what was actually served, and both produce a number that answers the question a client is actually asking: not what you sourced, but what you changed and what it saved.
What a defensible catering reduction claim requires
Making either lever count in a report means treating it the way any other measured category is treated: a baseline before the change, a specification that states the change, and a measurement after the event that shows whether it held. A menu brief that specifies a plant-forward default, a guarantee-and-overage policy that ties production to registration data rather than a flat buffer, and a post-event count of what was actually served against what was produced, together turn “we tried to do better” into a number a sponsor or auditor can check.
That is also where the two levers connect back to measurement. Recipe- and supplier-level data is what lets you show which menu swap moved the number and by how much, rather than asserting that a plant-forward menu probably helped. Without it, a genuinely large reduction and a marketing claim look identical on paper.
Two decisions, not a hundred small ones
Sustainable catering checklists tend to list a dozen items: local sourcing, reusable service ware, seasonal produce, compostable packaging. Most of them help at the margin. Two decisions, what the menu defaults to and how much gets produced against an uncertain guarantee, carry most of the weight, and both belong to the organiser, not the attendee.
Specify them, measure against them, and the catering line in your next report stops being an estimate and starts being an answer. Book a carbon strategy session to work through what a measured catering baseline would show for your next event.
