Short answer: when one person puts the office lunch order on a personal card, the share that never comes back can be the boss’s. The amount is small. What makes it hard is the gap in power between the person who has to ask and the person who has to notice. Research on power suggests that people with less power hold back. Whether power also makes people overlook what others know is less settled. So the fix is a setup that takes the one-on-one ask out of it: each person pays at checkout, or each share is posted to everyone before the food arrives.
A 2020 letter to the workplace advice column Ask a Manager shows the pattern in one office. The writer worked in an office of about five people that ordered delivery or take-out for lunch “probably once or twice a week.” As “the most junior at the company,” the writer was often asked to place the order, put it on a personal card, and had everyone pay them back. The other colleagues were “quick to ask how much they owe.” The CEO had “never paid for his food or even offered to do so.”
Source: how do I get my rich boss to pay me back for lunch?, Ask a Manager, January 2020 (reader letter; read from the Internet Archive copy).
Why can asking the boss feel harder than asking a peer?
Asking a peer for $14 is a small ask. Asking the CEO for $14 is a small ask carried up a hierarchy, and research on power gives reasons to think that makes it harder to start.
In a 2003 review in Psychological Review, Dacher Keltner, Deborah Gruenfeld and Cameron Anderson proposed that power changes behavior through two systems. Elevated power is associated with rewards and freedom, so it activates approach. Reduced power is associated with “increased threat, punishment, and social constraint,” so it activates inhibition. In their summary, reduced power goes with attention to threat and to others’ interests, and with “inhibited social behavior.”
That fits the letter. The junior person is the one who would have to start the conversation, and the junior person is the one the theory predicts will hold back. The review covers power in general and has no data on lunch orders. It gives a reason to expect the ask to be harder when it goes up the hierarchy, whatever the dollar amount.
Research on speaking up at work points the same way. Frances Milliken, Elizabeth Morrison and Patricia Hewlin interviewed 40 employees for a 2003 study in the Journal of Management Studies. Most had been in situations where they were concerned about an issue but did not raise it with a supervisor. The reason mentioned most often was “the fear of being viewed or labeled negatively,” and so damaging valued relationships. James Detert and Amy Edmondson’s 2011 work in the Academy of Management Journal describes “implicit voice theories”: taken-for-granted beliefs about when speaking up at work is risky or inappropriate. Across four studies, they found these beliefs widely held, and they added to the explanation of workplace silence.
What these studies do and do not show. None of them studied lunch orders or reimbursement. Milliken and colleagues studied concerns about work issues in 40 interviews. Detert and Edmondson studied reluctance to speak up, including about suggestions that would help the organization. The step from “people hesitate to raise problems with the boss” to “people hesitate to ask the boss for $14” is our own inference.
Sources: Keltner, Gruenfeld & Anderson, Power, approach, and inhibition, Psychological Review 110(2), 2003; Milliken, Morrison & Hewlin, An Exploratory Study of Employee Silence, Journal of Management Studies 40(6), 2003; Detert & Edmondson, Implicit Voice Theories, Academy of Management Journal 54(3), 2011.
Why doesn’t the boss notice that someone paid?
The letter writer thought the CEO had simply not noticed. “I honestly don’t think it crosses his mind to wonder who paid for the meal,” they wrote. The repayment conversation happened in an open office, and the CEO “seems to be entirely oblivious” to it.
A 2006 paper in Psychological Science by Adam Galinsky, Joe Magee, Ena Inesi and Deborah Gruenfeld offers a possible lead. Across four experiments and a correlational study, power was associated with “a reduced tendency to comprehend how other people see, think, and feel.” The fronter’s side of lunch, the card, the total and the wait for repayment, is a view the boss would have to take. The paper did not test that situation. Two of its experiments tested a narrower question: whether high-power participants took into account that other people did not have their privileged knowledge. They were less likely to than low-power participants.
A key experiment did not replicate. In 2016 the Many Labs 3 project re-ran Study 2a, one of the two privileged-knowledge experiments. Study 2a used 42 undergraduates and a scenario about a sarcastic email after a bad restaurant meal. The replication pooled 2,969 participants and found that high-power and low-power participants read the email about the same way. In the authors’ words, the replications “did not provide evidence of an effect of power on perspective taking.” That test covered only one of the paper’s experiments, but it weakens the main experimental evidence that power makes people misjudge what others know.
So treat the research as one possible reason. None of these studies looked at CEOs at lunch. The fix holds whatever the cause, though. From the boss’s seat, lunch arrives, and the card and the total are someone else’s view. Whether the boss is busy, distracted or blind to it, the number has to be put in front of them.
Sources: Galinsky, Magee, Inesi & Gruenfeld, Power and Perspectives Not Taken, Psychological Science 17(12), 2006; Ebersole et al., Many Labs 3: Evaluating participant pool quality across the academic semester via replication, Journal of Experimental Social Psychology 67, 2016, section 4.5.
Who pays when nobody asks?
Someone always pays. When the ask does not happen, the boss’s share lands in one of three places.
The share stays on the personal card. In the letter, that person is the most junior in the office and, as the writer notes, the boss knows what they make.
The writer suspected that colleagues “simply pay for him,” and said the other colleagues, all wealthier than the writer, were “willing to eat the cost to avoid asking him to pay.” The ask is skipped, and the cost moves sideways instead of up.
This is the only outcome where the person who ate the food pays for it. It is also the only one that needs someone to speak up the hierarchy.
The amounts add up because the order repeats. Take the letter’s numbers as an illustration only: a $14 salad, and an order once or twice a week. The letter does not say how often the CEO joins. It says only “on the occasions” he does, and it does not say what he orders each time. So the arithmetic below is an assumption about his weekly share.
The letter names a $14 salad. It does not say if that figure includes tax or fees. If $14 is the menu price, tax and any fees come on top. On splitty’s US-leaning restaurant receipts (an October 2026 snapshot), roughly 7 in 10 bills (about 72%) carry an effective sales-tax rate (the tax line divided by the subtotal) of 8% or higher, and the median bill’s rate falls in the 8–9% band. Those figures come from restaurant receipts, and a delivery order can add fees on top. Either way, the number to post is the share after tax and fees, which is the same gap between menu price and real share that the guide on paying back what you ordered works through.
Why change the setup instead of asking after lunch?
Every fix that happens after lunch needs the same two things: the junior person has to start the conversation, and the boss has to take in a cost they did not see. The research above gives reasons to doubt both steps, though none of it tested repayment. The fixes below remove one step or both. No study has compared them, so treat this as practical reasoning.
Uber Eats and DoorDash both document the strongest version. Uber Eats offers a group-order option called “Guests pay for themselves,” where participants pay for their own order, taxes and fees. Each participant has to be logged in to an Uber Eats account, and an order can be split a maximum of 18 ways. Uber also states that group orders default to “You pay for everybody.” So if nobody changes the setting, the app makes one person the fronter.
DoorDash’s guide for its business accounts describes “Everyone Pays Separately,” where each person pays for their own items, using their preferred payment method or an expensed meal budget where one is available. The guide adds that once the group order is created, “this payment option selection cannot be changed.” The decision has to be made at the moment the order is created, before anyone picks lunch. That is the right moment anyway. For how each app’s split works in detail, see the guides to splitting an Uber Eats order and DoorDash group orders.
Sources: Uber Help, How to Place a Group Order; DoorDash Help Center, Group Orders Guide (DoorDash for Business). Both read October 2026; app settings change, so check the current screen.
How do you get the boss to pay you back?
Move the ask to before the food arrives, and make it the same ask for everyone. In order of strength:
If the app offers it, everyone pays for their own food at checkout, the boss included. Nobody fronts anything, so nobody collects anything.
Before the food arrives, post each person’s share after tax and fees to the whole order thread, the boss’s line included. One message to everyone reads as the order’s receipt, and it puts the number in front of a person who may not look for it.
If the share is still unpaid, ask in private, with the exact amount and the way to pay: “Your lunch today was $15.40. Cash or a transfer both work.” A specific number is a logistics question. A hint is something the boss has to decode, and the boss may not. ($15.40 is an illustrative figure.)
If the office orders together every week, ask for the order to go on a company card or a set meal budget. That is one request, made once, about how the office orders.
If someone has to front it, make it a rotation that includes the senior people. The job of fronting should not go to whoever is least able to say no.
If you are the boss reading this: the fix is on your side too. Ask what you owe when the order is placed, and pay it the same day. The person who fronted your lunch may find it harder to ask you than anyone else in the office. For the other direction, when the boss is at the table and picks up the check, see the guide to splitting the office lunch. For the newest person at a group dinner, see the intern’s guide to the group dinner.
Is a team lunch order a work expense?
Sometimes, and the answer changes what you should do. A lunch your coworkers chose to order together is a social arrangement. A lunch your manager told you to order for a meeting is closer to a work errand, and work errands go through expenses.
California, for one, puts this in law. California Labor Code section 2802 says an employer “shall indemnify his or her employee for all necessary expenditures or losses incurred by the employee in direct consequence of the discharge of his or her duties, or of his or her obedience to the directions of the employer.” Whether a given lunch order counts as a “necessary expenditure” depends on the facts, and rules differ by state. This is not legal advice. The practical point is simpler: if you were told to place the order, ask how to expense it before you place it.
Source: California Labor Code section 2802, read from the public.law mirror of the official code text.
Where splitty fits
splitty helps with option 2: the order you had to front. Scan the itemized receipt. Every item starts split among everyone, so tap to remove anyone who didn’t share an item. Tax and tip are divided in proportion to each person’s share, so each person sees their real share with tax and tip included. Then each person gets a pre-filled payment request in their own payment app. Only you need the app. The others just receive the request.
For an office that orders together every week, save the regulars as a Group so the next split starts with the right people. splitty works one order at a time. It does not keep a running balance across weeks.
Common questions
FAQ
Questions & Answers
01 How do I ask my boss to pay me back for lunch?
Ask once, in private, with the exact amount and how to pay. A specific number reads as logistics. Better still, post every person's share to the whole order thread before the food arrives, so the boss sees the number at the same time as everyone else.
02 Why don't bosses notice they owe for lunch?
There is no settled answer. In two experiments in a 2006 Psychological Science paper, high-power participants were less likely to take into account that others did not share their privileged knowledge. But a 2016 replication of one of those experiments, with 2,969 participants, found no effect. The fix is the same either way: put the number where the boss will see it.
03 Should the junior person always place the office lunch order?
No. The person least able to say no ends up carrying the cost and the ask up the hierarchy. Use an app's pay-separately option, put a recurring order on a company card or meal budget, or rotate the job of placing the order to include senior people.
04 Can everyone pay separately on a delivery group order?
On some apps, yes. Uber Eats has a 'Guests pay for themselves' option for group orders, though Uber says group orders default to 'You pay for everybody'. DoorDash's business guide describes 'Everyone Pays Separately', which cannot be changed after the group order is created. Check the current app screen.
05 Does my employer have to pay me back for a team lunch I bought?
It depends on why you bought it and where you work. California Labor Code section 2802 requires employers to indemnify employees for necessary expenditures incurred in the discharge of their duties or in obeying the employer's directions. Whether a lunch order qualifies depends on the facts. If you were told to order, ask how to expense it first.
06 What should the share include?
The menu price plus that person's share of tax, fees and tip. On splitty's US-leaning restaurant receipts, about 72% of bills carry an effective sales-tax rate of 8% or higher, so the menu price alone undercounts the share.
