Travel P&L Open the app

How to figure out who owes who money after a group trip

The trip is over, the group chat has gone quiet, and somewhere in that silence five people are privately wondering whether they came out ahead or behind. One of them paid for the house. Another covered the rental car. Somebody bought every round of groceries and said "we'll sort it out later" each time.

Sorting it out later is a job with a method, and figuring out who owes who money after a group trip fits in five steps. You do not need to remember who handed cash to whom in which parking lot, and you do not need a spreadsheet with matrix formulas. You need each person's total paid, each person's fair share, and the difference between the two. Everything else in this article is what to do with those differences - including how to pay them off in the smallest number of transfers a group of five can mathematically achieve.

If you want the answer before the explanation, our free who-owes-whom calculator does the netting for you when the trip splits equally: type in who paid what, and it prints a short list of payments that settles everyone. This page is for the two things a calculator cannot do - the expenses that only half the group shared, and defending the final number in the group chat.

How do you figure out who owes who money after a group trip?

Add up what each person paid, work out what each person's share of the spending was, and subtract one from the other. That difference is their balance. People with negative balances owe money, people with positive balances are owed money, and the balances always cancel out exactly. Then the debtors pay the creditors until every balance reads zero.

As a checklist:

1. List every expense of the trip: amount, who paid it, and who it was for.
2. Total what each person actually paid out of their own pocket.
3. Total each person's share - their piece of every expense that was for them.
4. Subtract share from paid. Positive means the group owes them; negative means they owe the group.
5. Have the biggest debtor pay the biggest creditor, repeat until everyone is at zero.

One thing the checklist never computes is "how much do I owe Leo, specifically". Person-to-person IOUs are the long way around. The whole trick of settling a group trip is that you owe the group, the group owes some people, and the pairing of who pays whom at the end is arbitrary. That single idea is what turns a shoebox of receipts into four transfers, and the rest of this page unpacks it slowly.

Why does settling up take two weeks without notes?

Because reconstruction is harder than recording, roughly in the way that unbaking a cake is harder than baking one. A group that logs expenses as they happen settles in minutes. A group that saves the accounting for the end inherits a pile of receipts, half of them in a language nobody reads, and a set of card statements that refuse to match those receipts because the bank applied its own exchange rate.

That is a description of a real trip. A financial planner named Kenneth Tucker wrote up his group's twelve days in Greece, five adults, pay-as-you-go, settle at the end. His records ran into every wall in order:

"ALL OF THE RECEIPTS ARE IN A FOREIGN LANGUAGE! Why did I not anticipate this!?" Kenneth Tucker, Tusk Financial Planning, checked 19 September 2026

Then a 41-euro receipt that showed up as $45.76 on the card statement, and finally the summary line - "It took two weeks for our group of five to settle up." One member of the group, wading through the mess, accidentally paid a friend she had meant to send a payment request to.

Arguments follow the mess. In a 2026 Expedia Group survey for Vrbo, roughly one traveler in three said they had argued with friends or family on a group trip, and among Gen Z it was 43%. Vrbo's top tip in the release was boring and effective: agree on the money before and during the trip, while the facts are still fresh.

So the honest first step of "how to figure out who owes who" happens mid-trip: write each expense down the day it happens, with the payer and the list of people it covered. Each entry takes seconds; the reconstruction it replaces took Tucker's group two weeks.

What is a net balance, and why does it always add up to zero?

A net balance is one person's total paid minus their total share, and across the whole group those balances must sum to exactly zero - every dollar somebody overpaid is a dollar somebody else underpaid. That zero is your free error check: if the balances do not cancel, an expense is missing, duplicated, or split against the wrong people.

Worth pausing on why the zero works, because it is the part people take on faith. Every expense enters the math twice: once on the "paid" side, under the person who tapped the card, and once on the "share" side, spread over the people it covered. Same amount both times. Sum everything and the two sides cancel line by line, so the total of all balances has no choice about what to be.

This gives the treasurer a diagnostic no shoebox of receipts offers. Balances sum to -$40? Some expense was recorded as paid but never assigned to anyone, or one person's share got counted twice, and the size of the gap tells you what to search for. A total that reconciles to 0.00 does not prove every entry is right. It does catch the majority of honest bookkeeping mistakes - transposed digits, a forgotten taxi, a dinner split over the wrong four people.

One warning before the example: balances are net figures, so they already contain every "you got the coffees, I got the tolls" micro-deal of the trip. Once you settle by balances, those side arrangements are inside the number. Trying to honor them again on top of the netting is double counting.

Five friends, 18 IOUs, four transfers: the worked example

Eighteen person-to-person IOUs collapse into four transfers on this trip, and the table below is where the collapse happens. Five friends rent a house up the coast for a long weekend. Sam books the house, Leo drives and fills the tank, Ivy does the supermarket run, Max pays the one taxi, and Noor books a wine tour that only she, Sam and Ivy join. Made-up numbers, chosen to divide cleanly; when a real pot refuses to divide - four stray cents across six people - the largest-remainder fix and the tie-break rule are worked out in the six-friend ski trip split.

The ledger, expense by expense:

• House, all five: $1,000, paid by Sam
• Car and fuel, all five: $460, paid by Leo
• Groceries and one big dinner, all five: $340, paid by Ivy
• Airport taxi, all five: $60, paid by Max
• Wine tour, three people: $240, paid by Noor, for Noor, Sam and Ivy

The all-five pot is $1,860, which is $372 a head. The wine tour adds $80 to each of its three participants. Now the three numbers for each person:

Counted as person-to-person IOUs, this trip contains 18 debts: four people owe Sam for the house, four owe Leo for the car, four owe Ivy for groceries, four owe Max for the taxi, and two owe Noor for the tour. Max is simultaneously owed $12 by Sam for the taxi and owes Sam $200 for the house, which is the kind of arithmetic that makes groups give up. The balance column ignores all of it and says something shorter: Sam should receive $548, Leo should receive $88, and Ivy, Noor and Max owe $112, $212 and $312.

Four transfers close the whole weekend: Max sends Sam $312, Noor sends Sam $212, Ivy sends Leo $88 and Ivy sends Sam $24. Every balance hits zero, and Sam gets back exactly the $548 they were out of pocket, which is the check that the split is right.

Keeping that ledger by hand is the tedious part, and it is the part we built Travel P&L for: log each expense with its payer and its people as it happens, and the balance column from the table above just exists, all trip long, in 154 currencies. It does the arithmetic and shows everyone the same numbers, which on most trips is the half that starts arguments; the money itself moves through whatever your group already uses.

How do you settle up in the fewest transfers?

Repeat one move: the person who owes the most pays the person who is owed the most, until balances run out. Each payment brings at least one person to exactly zero, so a group of N people always settles in at most N-1 transfers. Five people, four transfers, no matter how tangled the trip was - and if some balances happen to match up, fewer still.

Walk it through on the example. The biggest debtor is Max at $312, the biggest creditor is Sam at $548, so Max pays Sam $312 and drops out. Noor pays Sam $212 and drops out. Sam is now owed $24. Ivy, the last debtor at $112, pays Leo his $88 - Leo is now the larger creditor - and pays Sam the final $24. Four payments, five zeros.

Why can it never take more than N-1? Every transfer wipes out at least one participant: either the debtor pays off everything they owe, or the creditor receives everything they are due, or both. You cannot wipe out the last person alone, because balances sum to zero - when the second-to-last account closes, the last one closes with it. That argument comes from Tom Verhoeff of Eindhoven University of Technology, whose 2004 paper Settling Multiple Debts Efficiently is the closest thing this little problem has to a classic text.

The same paper prices the alternatives. Paying every debt on its own - Verhoeff calls it the trivial solution - costs up to N times N-1 transfers, which is 20 for a group of five. Cancelling each pair of mutual debts first, his paired solution, still costs up to half of that. Our five friends had 18 separate debts on the books. Netting cleared all 18 in four transfers.

Can you ever settle in fewer than N-1 transfers?

Sometimes, and finding those cases is genuinely hard - hard in the formal, computer science sense. Fewer than N-1 transfers exist only when some subgroup of balances cancels to zero on its own, so the group splits into islands that settle internally. Spotting such subgroups quickly has no known reliable shortcut, which is why apps promise few transfers rather than the absolute minimum.

A small example shows the gap. Say five balances are +$700, +$600, -$600, -$500, -$200. The biggest-debtor-pays-biggest-creditor routine takes four transfers: 600 to the 700, then 500 to the 600, then the 200 splits its payment across two people. Look closer, though, and +$600 and -$600 cancel each other exactly. Pair them off in one transfer, settle the remaining three people in two, and the whole group is done in three.

Verhoeff proved that hunting for the best such split is an NP-hard problem - his phrase for the transfer-minimizing variant is "an elegant NP-hard problem", and he shows it is "at least as difficult as 3-Partition", a known monster. In practice, as he puts it, you can "not do much better than trying out all possibilities" - and for balances that means checking subgroup after subgroup, which explodes long before your friend group does.

For a treasurer this theory collapses into one comforting sentence. The lazy rule - biggest debtor pays biggest creditor - is one short step from perfect: it never needs more than N-1 payments, and squeezing out the rare extra transfer is a job even computers find expensive. Settle in N-1 and go to bed.

Why does the app say I owe someone I never borrowed from?

Debt-simplifying apps reroute money along the shortest path, and the shortest path often runs between two people who never shared a single receipt. The total each person pays or receives stays exactly the same; only the addresses on the envelopes change. Every major app that does this says so, and every group that skips reading the explanation has this argument.

Splitwise, which has run this feature for over a decade, explains the rerouting with three names: "if Anna owes Bob, and Bob owes Cathy, then Anna can just pay back Cathy directly." Their own worked example collapses eight payments into three, and their explainer page - a page that exists precisely so users can send it to confused friends - puts the promise in one line:

"'Simplify debts' does not change anyone's total balance. It just reduces the number of payments, so everyone gets paid back quicker." Splitwise, "What exactly is 'simplify debts'?", checked 19 September 2026

Settle Up says the quiet part in its official tips: "you might end up paying to somebody you don't directly owe." Tricount's FAQ states the same goal - "The goal is to limit the number of transactions between participants."

The trouble is that a rerouted debt no longer tells its own story. A commenter named Nishit put it plainly under Splitwise's announcement post back in 2013:

"After debt simplification there is no way to find out the original amount to check whether debt simplification worked correctly. so it just leaves me confused and with no choice." Nishit, 29 August 2013, in the comments under Splitwise, "Debts Made Simple"

Thirteen years later, that is still the complaint.

Two rules keep the peace. Decide about simplification before anyone sends money, because flipping it mid-settlement rearranges debts people have already half-paid - Splitwise's own help pages warn that turning it off after payments start can mean money has to be re-sent. And when someone balks at paying a near-stranger, show them the balance table from this article: they are paying their own share of their own expenses, just through one envelope instead of five.

One more thing worth knowing before a group storms out of an app over this: a balance is not stored inside the app. It is paid minus share, and it survives the move intact - the work is copying it across without anyone losing a number, which we wrote out step by step in switching from Splitwise without losing your balances.

Should you settle during the trip or once at the end?

Once, at the end, after the last expense is in - including refunds that arrive after you get home. Every payment made mid-trip is a payment made against balances that are still moving, and the piecemeal transfers themselves become entries someone has to track. The exception is a trip so long that carrying the balance genuinely hurts.

Splitwise users have been asking for a guard rail against exactly this since 2015. A feedback thread titled "you should have an option to 'lock' the settle-up button until everyone is done inputting their expenses" is still open and still collecting votes. "So tired of my friends settling up before all expenses have been entered," reads a comment from 2021; another describes paying a host too early, and the host "had to refund until end of weekend and until everyone joined". Deposits come back, the last tank of fuel lands after the flight, and the person who settled on day two now owes a second, smaller, more annoying amount.

There is a decent middle path for long trips, and Settle Up ships it as a feature: "The person with the biggest debt in the middle should pay next", meaning whoever is deepest in the red picks up the next restaurant bill. Balances drift toward zero all trip without a single transfer being made.

If you do settle mid-trip anyway, log the payment in the same ledger as the expenses. A transfer that lives only in two people's memories is exactly the kind of entry that surfaces a year later as a Venmo request nobody can explain.

What about couples, latecomers and expenses not everyone shared?

All three are the same move: change the list of participants on the line, never the arithmetic of the netting. An expense's cost is divided among the people it was for - that may be two, three or all five - and the netting at the end works unchanged no matter how ragged the participant lists are.

The worked example already did this once: the wine tour carried three names, so only three shares of $80 exist, and Leo and Max never see the line. Groceries for everyone, cocktails for four, museum for two - each line gets its own list, and nobody subsidizes an outing they skipped. The extreme case is a person the group deliberately leaves off the paying side of every line - the groom, the birthday guest, the friend being treated - whose share gets absorbed by everyone else. That is still only a participant-list change, and the netting never notices; the full arithmetic of it, cents and all, is in who pays for the bachelor party.

Couples who run one household budget come down to the participant list too. Treat the pair as one wallet with a double-weight share, and transfers inside the couple stop cluttering the settlement. A commenter on Splitwise's forum framed the confusion as a question back in 2016: what happens when "one person of the couple pays and the other does not but you are still splitting evenly and they need to get money back"? Under the one-wallet rule the question dissolves - the pair has one balance, and it is either positive or negative. Splitting the pair's own balance between the two of them afterwards is a separate question, and it is the one how couples split travel expenses answers - including the version where incomes differ and 50/50 stops being the fair number.

Latecomers divide opinion sharply enough to plan for: in the same 2026 Expedia Group survey for Vrbo, half of respondents said a guest who arrives a day late or leaves a day early should pay a smaller share of the lodging. Half is not a consensus - it is a coin flip on what your particular group already believes.

Pick the rule out loud before the trip. Ours is person-nights, and the full version - the arithmetic, the premium for the good bedroom, the short weekend where it is not worth the spreadsheet - is in splitting an Airbnb with friends. It drops into the balance method as just another uneven participant list.

Does "I paid the hotel, you paid the gas" make you even?

Only if the two amounts per person happen to match, which they almost never do. "You got this, I'll get that" feels fair because both people paid something. Whether it is fair depends on the numbers: your half of the gas and their half of the hotel are two different figures, and the difference between them is a real debt that compounds politely in the background.

Check it with balances in ten seconds. Two people, hotel $380 paid by you, gas $90 paid by them. Your share of everything is $235, and so is theirs. You paid $380, so your balance is +$145; they paid $90, so theirs is -$145. "Even", in this trip, was $145 away from even - and the gas-buyer usually senses it, which is why they got the coffees and the parking and still felt vaguely in debt all week.

Scaled up to five people making these side deals all weekend, pairwise evening-out is how a trip ends up with 18 open IOUs and a two-week reconciliation. The balance method eats all of it: every "I'll get this one" is just another line with a payer and a participant list, and one pass of netting at the end settles the lot.

Can a spreadsheet handle who owes who?

Yes for a simple trip, and it degrades fast as the trip stops being simple. Equal splits among a fixed group fit a spreadsheet fine: one column of amounts, divide by N, subtract - and if that is the whole trip, our free browser calculator does that job without a sheet at all: names, amounts, and the transfer list, no sign-up. The pain arrives with reality - subset expenses, a latecomer, a currency, a changed plan - because every one of those edits formulas by hand.

A veteran of the do-it-yourself approach on the Tiller money-tracking forum described maintaining exactly the sheet this article describes, with a matrix working out each person's cut of each expense. Their verdict on the who-owes-who part:

"it was so complex to change when the group membership changed and was so error-prone I just got rid of it" a poster in the "Multi-dimensional Google Sheet calculations for group vacation expense tracking" thread on Tiller Community

The formulas could not be dragged, names had to be re-edited for every trip, and one wrong cell silently shifted everyone's balance.

That silence is the real argument against the sheet. The zero-sum check from earlier still works, and a spreadsheet will happily display balances that sum to zero while a mistyped participant list sits inside them. Purpose-built trackers - Splitwise, Tricount, Settle Up, and yes, our own Travel P&L - all do the same three things a sheet fights you on: participant lists per expense, currencies converted as you go, and the transfer list generated instead of hand-derived. The honest comparison is that a sheet costs nothing and handles the easy 80%, and the remaining 20% is precisely the part that was going to cause the argument.

Where the apps differ is in emphasis - receipt scanning here, budgets there. What none of them do, ours included, is move the actual money or chase the actual friend. The arithmetic layer and the payment layer stay separate: figure out who owes who in one place, then pay through whatever your group already uses.

What if the math is right but nobody pays?

Correct math that nobody acts on is not an arithmetic problem, and no formula on this page will fix it. What helps is making the number explainable, agreeing the deadline out loud, and sending the request while the trip is still warm.

The silence is the default, not the exception: an October 2021 LendingTree survey found that 40% of Americans owed money by a loved one had never even asked for repayment, whether out of awkwardness or to give the borrower more time. Nothing on this page fixes that. What it can do is remove the second excuse - the one where the amount itself is unclear, disputed, or arrived at by somebody else's arithmetic.

The etiquette layer matters more than treasurers like to admit. One essayist's much-shared rule - "a Venmo request down to the cent is an absolute freak move" - is really a rule about context: a bare number with no story attached reads as an accusation. A request that arrives with the ledger attached, the day after the trip, while everyone remembers the wine tour, is a different message entirely. That is half the case for keeping the shared ledger somewhere everyone can see it: the request explains itself before it is sent.

For the harder cases - the friend who "forgot" three times, the debt that ages past embarrassment into silence - we wrote a separate playbook with message templates: how to ask friends to pay you back. The short version stands here: fair math, shown early, beats perfect math shown late.

Sources

Every link below was opened and every quote re-checked on 19 September 2026.

• Tom Verhoeff, "Settling Multiple Debts Efficiently: An Invitation to Computing Science", Informatics in Education, Vol. 3, No. 1 (2004), pp. 105-126, Eindhoven University of Technology - the N-1 bound, the trivial and paired solutions, the zero-sum subgroup condition for doing better, and the NP-hardness of the true minimum.
• Splitwise Blog, "Debts Made Simple", 14 September 2012 - the Anna-Bob-Cathy rerouting example, and the comment thread below it, including Nishit's complaint of 29 August 2013 quoted above.
• Splitwise, "What exactly is 'simplify debts'?" - the eight-payments-to-three example and the one-line promise quoted above.
• Splitwise Help Center, "What is Simplify Debts?" - "It never changes anyone's total balance - it only optimizes the payment paths"; the warning that turning the feature off after payments begin can mean people re-send funds; the note that simplification works within each currency separately and does not convert across them.
• Settle Up, official tips - "you might end up paying to somebody you don't directly owe" and "The person with the biggest debt in the middle should pay next".
• Tricount FAQ - overpaid and underpaid participants, "The goal is to limit the number of transactions between participants", and 1-cent accuracy.
• Splitwise feedback forum, "you should have an option to 'lock' the settle-up button until everyone is done inputting their expenses", opened 2015, still open with 28 votes - the premature-settling comments quoted above, dated 2017 and 2021.
• Splitwise feedback forum, "Groups to include couples" - the one-wallet couple question, asked 14 September 2016, in a user's own words.
• Kenneth Tucker, "Splitting expenses with friends while travelling overseas... is not optimal", Tusk Financial Planning - the Greece trip, five people over twelve days: foreign-language receipts, a 41-euro receipt landing as $45.76, two weeks to settle.
• Vrbo / Expedia Group group-travel survey, March 2026, conducted by OnePoll among 2,000 US adults who had stayed in a vacation rental with two or more people in the past three years - one in three has fought on a group vacation, 43% among Gen Z; half say a late arrival should pay a smaller lodging share.
• LendingTree, money-owed survey, fielded 15-20 October 2021 among 2,051 US consumers - 31% are owed money by a loved one; 40% have not asked for repayment. The oldest figure on this page, and labelled as such.
• Casey Gomez Walker, "we need to talk about venmo", 24 April 2024 - the cent-precision request rule quoted above.
• Tiller Community, "Multi-dimensional Google Sheet calculations for group vacation expense tracking" - a power user's account of why their who-owes-who sheet died.

So who owes who?

Whoever's balance is negative - they owe exactly that number, to whoever's balance is positive. And now you can produce that answer for any trip in five steps and defend every dollar of it. The math has been settled since 2004. The only part left to get right is writing things down while they happen.

Travel P&L keeps that ledger for you: every expense with its payer and its people, 154 currencies with a rate you can override, and the who-owes-who list ready before the airport goodbye. Everything is currently free, friends join by one link, and the friend who fronted the house gets one transfer instead of a shoebox.