The 28% gap is a comparison between groups of workplaces, not a before-and-after inside any canteen. The study examined how levels of readiness relate to the climate footprint of the food actually served [15], and the same analysis reports that workplaces where staff eat less meat in their own lives are the ones more open to change [7]. Part of that gap therefore describes the people, not a menu anyone rewrote. Inverted, the resistant sites serve lunch with about 39% more emissions than the ready ones [13], and they are also where traditional Danish dishes built around meat remain standard [5].
That puts the available abatement in the least cooperative places. Across 62 sites, the reported split works out to roughly eight change-ready workplaces against 26 resistant ones [16]. A caterer that starts with the willing eight is working the part of its client list where the footprint is already lowest, while the tonnage sits with the 26 [13][16].
One vegetarian day a week is the outer edge of the published evidence on tolerance [9]. On a five-day lunch week, that bounds the menu change at 20% of meals [14]. What 20% does to a site's footprint cannot be read off these figures, because the 28% comparison is between whole menus at different kinds of workplace rather than between one substituted day and four unchanged ones [4][15].
The emotional numbers need care about units. The annoyance and anger figures come from a nationally representative survey of individual employees who get lunch at work [8]; the 42% resistance figure counts workplaces in one Danish region [3]. The two sit close enough to look like mutual confirmation and they are not measuring the same object. Lund's own reading is that some of the reaction is instinctive, since satisfaction with canteen food does not track whether plant-based food is served, up to that one-day limit [9].
So the operational input here is a screening instrument, not a recipe. Readiness has observable correlates in sector and staff composition [6][7], which means a caterer or an employer can predict where a menu change will cost management time before spending any. The carbon side of the ledger was never the hard part: the difference between the top and bottom tiers already exists in the data, unassisted [4]. What the study prices is the other side, and it prices it as concentrated rather than average. The partial results are in npj Climate Action, with the survey work reported separately in Danish [1][8].
