Food Court Splitting: Mall, Airport, and Stadium Multi-Vendor Chaos

You're at the airport food court. Flight boards in 90 minutes. Five friends scatter to five different counters. One gets a $14.50 pizza slice. One waits 12 minutes for a $19.75 burger. Someone grabs $31.25 in coffees for the group. Now figure out who owes what before you reach security.

The multi-vendor splitting problem

Food courts represent the most complex splitting environment in casual dining. Unlike a restaurant with one check and one server, food courts generate 4-7 separate transactions for a typical group of five. Each vendor has its own receipt, its own tipping policy, its own tax calculation. The splitting math doesn’t just multiply—it compounds.

The Airports Council International 2023 benchmarking survey documented what travelers already know: airport terminals now average 12-20 vendors per food court, each with its own payment terminal. Mall food halls have followed the same trajectory. What used to be a cluster of fast-food counters is now a curated collection of independent kitchens—and every additional vendor adds another receipt to reconcile.

 4-7xmore payment transactions at a food court compared to a single restaurant. Each transaction adds variables to track, receipts to collect, and math to reconcile. 

The core problem: when everyone orders from different vendors, there’s no single bill to split. There’s a collection of independent purchases that happened to occur in the same general location. Someone always ends up covering more than their share. Someone always ends up owing money they’ll forget to pay. Uri Gneezy, Ernan Haruvy, and Hadas Yafe’s 2004 field experiment at the University of California, San Diego demonstrated that informal IOUs have a 56% non-repayment rate—and Hermann Ebbinghaus’s foundational 1885 memory research explains why: recall accuracy for incidental details decays exponentially within the first hour.

Sources: Gneezy, Haruvy & Yafe, “The Inefficiency of Splitting the Bill,” The Economic Journal (2004); Ebbinghaus, “Memory: A Contribution to Experimental Psychology” (1885); Airports Council International, “Airport Concessions Benchmarking Survey” (2023).

Three types of food court chaos

Not all food courts are created equal. The splitting challenges vary dramatically based on venue type, and each brings its own pricing pressure, time constraints, and coordination failures.

Mall

Mall Food Courts

The most common type. Usually 8-15 vendors ranging from $8 fast food to $18 fast-casual. Groups tend to be smaller (3-5 people), time pressure is moderate, and the pricing is relatively standard.

Avg. markup10-15%

Time pressureLow

Airport

Airport Terminals

The highest-pressure environment. Captive pricing pushes costs 28-40% above street level according to ACI’s 2023 benchmarking data. Groups often include mixed expense accounts (business vs. personal). Boarding deadlines add urgency that eliminates careful math.

Avg. markup28-40%

Time pressureHigh

Stadium

Stadium Concessions

The highest prices per item. Groups scatter and regroup between innings or quarters. Technomic’s 2024 venue research reports average concession spending per fan has increased 23% since 2019. The $14 craft beer and $12 loaded nachos are standard, not outliers. Splitting often happens retrospectively—after the game.

Avg. markup50-200%

Time pressureEvent-driven

For concert and game day expenses, the multi-venue challenge extends beyond food to tickets, parking, and rideshares. But food courts concentrate the problem into a single sitting where everyone expects to settle up before leaving.

Sources: Airports Council International, “Airport Concessions Benchmarking Survey” (2023); Technomic, “Sports Venue Food and Beverage Trends” (2024).

The asynchronous food problem

At a restaurant, everyone orders at the same time, food arrives roughly together, and the bill comes at the end. Food courts break this rhythm completely. The asynchronous nature of food court dining creates coordination problems that compound the splitting challenge—similar to the multi-vendor confusion at food halls, but with added time pressure.

0:00Group arrives, scatters to different vendors

0:03Alex gets pizza ($14.50, ready immediately)

0:08Jordan’s $19.75 burger is called at the counter

0:10Sam grabs coffees for the table ($31.25, shared expense)

0:14Taylor’s $17 pho finally ready—others are half-finished

0:25Casey arrives late, gets $12.50 sushi, joins conversation

0:35Someone says “should we figure out the split?”

By the time anyone thinks about splitting, 35 minutes have passed. Multiple transactions occurred. Someone bought something for the group. Someone arrived late. The receipts are scattered across the table or already in pockets. Hermann Ebbinghaus’s foundational 1885 research on memory decay established that recall accuracy drops approximately 40% within the first 30 minutes for incidental details—and at a food court, “incidental details” includes everything you need to split fairly: who paid for what, how much the coffee round cost, and whether the churros were shared by four people or five.

The key insight

The person who gets their food last usually organizes the split. But they also have the least information about what happened before they sat down.

They're reconstructing from incomplete data, social pressure, and guesswork—exactly the conditions Ebbinghaus identified as most vulnerable to memory error.

Source: Ebbinghaus, “Memory: A Contribution to Experimental Psychology” (1885).

The captive pricing effect

Airport food courts don’t just create splitting challenges—they amplify the stakes. When a sandwich costs $16 instead of $11, the unfairness of equal splitting becomes more acute. The person who ordered a $9 salad is now subsidizing a $25 steak sandwich at inflated captive prices.

28%Minimum airport markup over street prices (ACI, 2023)

40%Common markup at major hubs like JFK, LAX, ORD

23%Increase in per-fan stadium spending since 2019 (Technomic, 2024)

Richard Thaler, the University of Chicago economist who won the 2017 Nobel Prize for his work on behavioral economics, explains why captive pricing warps splitting decisions. In his 1999 paper “Mental Accounting Matters” published in the Journal of Behavioral Decision Making, Thaler demonstrated that people don’t evaluate purchases in absolute terms—they evaluate them against reference prices. A $16 airport sandwich feels like a bad deal because you know it costs $11 elsewhere. But you’re buying it anyway because you’re hungry and captive.

When groups split equally in these environments, the person who ordered modestly (a $9 salad at the airport) subsidizes the person who ordered expansively (an $18 sandwich plus a $7 drink). The subsidy is larger in absolute terms because the prices are inflated. What might be a $3 difference at a normal food court becomes a $7 difference at an airport.

"

Mental accounting is the set of cognitive operations used by individuals and households to organize, evaluate, and keep track of financial activities.
Richard Thaler, Nobel Prize Lecture, 2017

Captive pricing also creates mixed expense account situations. At an airport, some travelers are on business (expensing meals) while others are paying personally. Drazen Prelec and George Loewenstein’s 1998 research at MIT and Carnegie Mellon on the “pain of paying” found that spending feels fundamentally different when the cost comes from a personal budget versus an institutional one. The business traveler ordering a $24 steak salad has different incentives than the personal traveler watching their budget. Equal splitting feels even more unfair when the person ordering expensively isn’t even paying their own money.

Sources: Thaler, “Mental Accounting Matters,” Journal of Behavioral Decision Making (1999); Prelec & Loewenstein, “The Red and the Black,” Marketing Science (1998).

Why your brain can’t track this

George A. Miller’s landmark 1956 paper in Psychological Review established the limits of human working memory: 7 plus or minus 2 items. Try to hold more than that, and information starts falling out. At a food court with five people, the variables easily exceed 20.

5Individual orders

5Order totals (with tax)

5Payment records (who paid what)

2Shared purchases (coffee round, snacks)

5Shares of shared items

22+Total variables to track

That’s 22 variables minimum—over three times Miller’s upper limit. And we haven’t counted tips at vendors that prompted for them, tax variations between vendors, or the late arrival who changed the share calculations. This is the same mental math overload that plagues restaurant splitting—but multiplied by the number of vendors.

John Sweller’s cognitive load theory, published in Cognitive Science in 1988, formalized what happens when task demands exceed working memory capacity: performance degrades. People make errors. They simplify. They take shortcuts. In bill splitting, the shortcut is always the same: “Let’s just split it evenly.”

Ivan Steiner’s 1972 research on group process losses at the University of Texas showed that coordination costs grow non-linearly with group size. Adding one more person to a food court outing doesn’t add 20% more complexity—it adds 40% or more, because every person creates new interaction points, new shared items, and new timing variations.

The real cost: Gneezy, Haruvy, and Yafe’s 2004 field experiment showed that equal splitting causes people to order 37% more than when paying individually. At a food court, you’re not even splitting equally on purpose—you’re defaulting to it because the actual math exceeds cognitive capacity. The modest orderer subsidizes everyone else by accident.

Sources: Miller, “The Magical Number Seven,” Psychological Review (1956); Sweller, “Cognitive Load During Problem Solving,” Cognitive Science (1988); Steiner, “Process Losses and Process Gains in Groups” (1972); Gneezy et al., The Economic Journal (2004).

How groups actually handle it

Without a system, food court groups default to one of four approaches. Each has predictable failure modes documented in the behavioral economics literature.

Most common

”Everyone pays their own”

Each person pays at their vendor. Simple. But what about the coffee round? The shared snacks? Someone’s out $15 for communal purchases that never get reconciled.

Fast and feels fair in the moment

Ignores shared expenses entirely

Common

”One person covers everything”

Someone puts it all on their card. The group promises to Venmo later. Gneezy, Haruvy, and Yafe’s 2004 data indicates 44% of those promises become IOU debt that decays over time.

Fast in the moment

Creates debt with 44% non-payment rate

Rare

”Rough math at the table”

Someone adds up receipts, divides by people, rounds generously. The math is usually wrong—Sweller’s cognitive load research predicts error rates of 15-20% when mental arithmetic exceeds working memory.

Attempts fairness

Mental math errors average 15-20%

Accurate

Systematic tracking

Log every purchase, note who paid, assign shared items, calculate net positions. Accurate—but requires a system nobody wants to operate manually.

Fair outcome for everyone

Too complex without tools

The fourth approach is the only one that produces fair outcomes. It’s also the one nobody does manually, because tracking 22+ variables while your flight boards or the game resumes is not realistic human behavior.

The systematic approach

Here’s how to actually split a food court outing fairly, whether you’re at a mall, airport, or stadium.

Collect receipts as you go

Don't wait until the end. As each person gets their food, snap a photo or collect the receipt. This prevents the "where did my receipt go" problem that derails most splits.

Flag shared purchases immediately

When someone buys coffee for the table, note it right then. "Sam grabbed coffees—$31.25 for five people." Don't trust your memory to reconstruct this 30 minutes later—Ebbinghaus's 1885 forgetting curve predicts you'll have already lost 40% accuracy.

Calculate before anyone leaves

The moment someone says "I need to head to my gate" is too late. Settle up while everyone's still at the table, before the group fragments.

Use net positions, not individual payments

Calculate what each person consumed (individual order + share of communal items). Compare to what they actually paid. The difference is their net position—what they owe or are owed.

Minimize transactions

Instead of 5 people sending money to 3 different payees, calculate who owes whom and optimize for fewest transfers. One person who owes $12 pays one person who's owed $12. Done.

This process is accurate. It’s also tedious. For a group of five at an airport food court with time pressure, doing this manually takes 8-12 minutes—time you may not have.

A worked example: airport food court

Five colleagues at O’Hare, 70 minutes until boarding. Everyone scatters, then regroups.

Individual orders

Alex: Pizza slice + soda (Sbarro)$14.50

Jordan: Burger + fries (Shake Shack)$19.75

Sam: Poke bowl (Aloha Poke)$17.00

Taylor: Caesar salad (Chopt)$13.25

Casey: Chicken sandwich (Chick-fil-A)$12.50

Individual subtotal$77.00

Shared purchases

Jordan grabbed: 5 coffees (Starbucks)$31.25

Alex grabbed: shared churros (4 people, not Casey)$12.00

Shared subtotal$43.25

PersonPaidConsumedNet

Alex$26.50$23.75Owed $2.75

Jordan$51.00$29.00Owed $22.00

Sam$17.00$26.25Owes $9.25

Taylor$13.25$22.50Owes $9.25

Casey$12.50$18.75Owes $6.25

The settlement: Sam and Taylor each send Jordan $9.25. Casey sends Jordan $6.25. Alex sends Jordan their remaining $2.75 (or Jordan calls it even—they’re colleagues). Total: 3-4 Venmo transactions instead of chaos.

Without systematic tracking, this group would have done one of two things: Jordan would be out $22 that never gets repaid, or everyone would “split it evenly” and Taylor (who had a $13.25 salad) would pay $24.05. Neither outcome is fair.

How research shaped the design

Every finding about cognitive load, memory decay, and group coordination maps to a specific design decision in splitty.

Miller (1956): Working memory maxes out at 7 items→Scan receipts to offload data entry—no typing, no remembering

Sweller (1988): Multiple transactions multiply cognitive load→Add multiple receipts to one session—splitty combines them automatically

Ebbinghaus (1885): Memory decays 40% in 30 minutes→Capture receipts immediately—don’t wait until the end

Gneezy (2004): Shared items require explicit allocation→Tap to mark which items were shared and by whom

Prelec & Loewenstein (1998): Net position math is error-prone→Calculate who owes whom automatically with optimized transactions

Steiner (1972): Time pressure causes shortcut decisions→Complete the split in under 30 seconds—before boarding or the game starts

Food court splitting questions

01 How do you split a food court bill with shared purchases?

Calculate each person's individual order total, then divide shared purchases (like a coffee round) among everyone who participated. Compare what each person consumed versus what they actually paid. The difference is what they owe or are owed. splitty does this automatically when you scan each receipt and assign shared items.

02 Should you split food court costs evenly?

Usually not. Unlike a restaurant where everyone orders from the same menu, food courts produce wildly different individual totals — from a $9 salad to a $25 combo. Uri Gneezy's 2004 research showed equal splitting causes people to order 37% more, knowing costs will be shared. Track what each person ordered and paid for a fair outcome.

03 How do you handle when one person buys food for the group at a food court?

Flag it immediately. When someone grabs coffees or snacks for the table, note the total and who it was for right away. Memory decays 40% within 30 minutes (Ebbinghaus, 1885), so waiting until the end means inaccurate reconstruction. Scan the receipt and mark those items as shared among the relevant people.

04 What's the best way to split airport food court expenses?

Airport food courts add captive pricing (28-40% markup) and time pressure. Collect receipts as you go, flag shared purchases immediately, and settle up before anyone heads to their gate. Use net positions — what each person consumed minus what they paid — to calculate the fewest possible payment transfers.

Five vendors. One split. Done before boarding.

Scan every receipt, assign shared items, settle up with one tap—while everyone's still at the table.