The Art of the Start: The Time-Tested, Battle-Hardened Guide for Anyone Starting Anything

By Guy Kawasaki

10 min read

Why does it matter? Because most fraud is committed by people who genuinely believe they're honest.

Most people assume dishonesty is a minority sport — the con artist, the fraudster, the one rotten apple. Dan Ariely ran the numbers, and the picture is far stranger. The Kennedy Center's gift shops were hemorrhaging $150,000 a year, not to a scheming employee, but to three hundred elderly volunteers who loved theater and music and quietly helped themselves to loose change. The standard model of theft — rational actors weighing punishment against reward — couldn't account for them at all. What Ariely found, across dozens of experiments on three continents, is that dishonesty isn't a character flaw concentrated in a few bad people. It's a predictable output of psychological machinery that runs in nearly everyone. Understanding it changes how you see fraud, institutions, and the specific moments when your honest self-image quietly makes room for something else.

The Real Fraudsters Were 300 Elderly Art-Lovers

Dan Weiss spent weeks convinced he was hunting one thief. He was the young manager of the gift shops at the John F. Kennedy Center for the Performing Arts in Washington D.C. — a network of small stands staffed by about 300 retired volunteers who loved theater and music. The shops did over $400,000 in business each year. They also hemorrhaged $150,000 annually, cash and merchandise disappearing with no explanation.

Weiss set a trap. He arranged a sting with a National Park Service detective, seeded the cash box with marked bills, and waited in the bushes outside. When a suspected employee left for the night with marked bills in his pocket, it looked like case closed. Except the employee had taken $60. And even after he was fired, the money kept disappearing at the same rate.

The real culprits were the volunteers. The retirees — well-meaning, culturally engaged, the kind of people who donate their weekends to support the arts — were quietly pocketing small amounts and helping themselves to merchandise. Not one person orchestrating a heist. Three hundred people taking a little, regularly, invisibly. When Weiss introduced a simple paper inventory system (write down what you sold, write down what you received), the theft stopped entirely.

This is the part that should disturb you. Because those 300 volunteers almost certainly didn't think of themselves as thieves. There were no cash registers, no written records, just a loose cash box and an atmosphere of informal trust. In that environment, taking a small amount probably didn't feel like stealing. It felt like a rounding error, a minor perk, something the organization could absorb. The moral calculation, if they made one at all, was soft enough to disappear.

The dominant theory of dishonesty says this shouldn't happen. The theory most anti-fraud policy is built on says a person decides whether to cheat by weighing the potential gain against the probability of getting caught, multiplied by the severity of the punishment. Crime is arithmetic. Raise the odds of detection or stiffen the penalties, and you reduce dishonesty. The cheaters are out there doing the math; change the math, change the behavior.

But the Kennedy Center story doesn't fit. The volunteers weren't calculating. They weren't risk-managing. They were good people in an unstructured environment, rationalizing small amounts, and their collective take dwarfed whatever a single calculated criminal might have managed. The one employee caught in the sting took $60. The collective low-level dishonesty of 300 art-lovers cost $150,000 a year.

The danger isn't the scheming minority willing to risk big punishment for big gains. It's the rest of us, given sufficient ambiguity and the right environment to quietly tell ourselves it doesn't really count. This is the case that opens Dan Ariely's investigation into why honest people cheat.

People Steal Coke From a Communal Fridge but Leave the Dollar Bills on the Plate

One reason ambiguity works is simpler than you'd expect: it puts distance between you and the thing that feels like money.

Imagine two things in a communal fridge: a can of Coke and a dollar bill. Same value. You're pretty sure no one's watching. Which one are you more likely to take?

Most people already know their answer — and it surprises them a little.

Ariely seeded refrigerators in MIT dormitories with either six-packs of Coke or plates of six one-dollar bills. Within 72 hours, every Coke had vanished. The cash went untouched, even though a student could have taken a single dollar, walked to the nearby vending machine, and come back with a Coke and change. Economically identical moves. But one felt like taking something, and the other felt like stealing money.

The distance is everything. When what you're taking doesn't look like money (a drink, a pen, a token), the internal signal that reads "this is theft" just doesn't arrive with the same force. A follow-up experiment shows exactly how much that distance matters.

Participants shredded their worksheets after a math task and claimed their extra earnings one of two ways: cash directly, or plastic chips immediately redeemable for cash at a table twelve feet away. Same room, same dollar amount, a transaction that took seconds. People claiming chips cheated roughly twice as much as those claiming cash. One psychological step between you and money, just the step of naming it something other than money, nearly doubled dishonesty.

Ariely's point is that we don't hold ourselves back through rational cost-benefit calculation. We hold ourselves back because stealing feels like stealing. Abstract the money away, even slightly, and that feeling softens.

The same principle extends beyond money itself. In a study using high-end Chloé sunglasses (all actually genuine), female MBA students were told their pair was either authentic or counterfeit. Those told their sunglasses were real cheated 30% of the time on an unrelated math task. Those told they were wearing fakes: 74%. A control group told nothing came in at 42%, closer to the authentic condition than the counterfeit one.

Nothing about the sunglasses changed. What changed was the story the wearers told themselves. Believing you're carrying a fake sends a quiet internal signal: I'm someone who accepts a less-than-honest version of things. Once that self-image shifts, even slightly, it bleeds into whatever comes next — and apparently that includes math.

A Prisoner's Fate May Hinge on Whether the Judge Just Had Lunch

Self-image isn't the only internal resource honesty draws on — there's also willpower, and it has a daily limit. Moral failure isn't a character defect distributed randomly across the population. It's a predictable output of a system that runs out of fuel, and the timing of that failure is more clockwork than conscience.

Israeli researchers studied years of parole board rulings and found a pattern with nothing to do with the prisoners' records. Judges granted parole most often at two specific moments: the first hearings of the morning, and the first hearings after lunch. The default ruling is denial — the safe, effort-free choice. Overriding it requires deliberate reasoning, and deliberate reasoning draws from a limited reserve. As the day's decisions accumulated, judges gradually defaulted to denial. By late morning their reserves were draining; by afternoon they were rubber-stamping denials. A prisoner's odds of freedom shifted dramatically based on where in the day their hearing happened to fall.

Call it depletion: every decision, trivial or consequential, drew from the same cognitive budget. There's no internal switch that reserves capacity for the important ones. The same mechanism that sends someone reaching for pizza at the end of a long day was routing prisoners back to their cells.

Depletion also erodes honesty. Mike Adams tracked "grandmother deaths" reported to professors across multiple semesters. Grandmothers turned out to be ten times more likely to die before midterms, nineteen times before finals, and fifty times more likely when the student was already failing. The students were exhausted people whose capacity to resist the easier story had simply worn thin.

Self-control is a daily budget, and moral failures concentrate exactly where you'd expect: at the end of a long day of spending it.

The Experts Who Advise You Can't See Their Own Corruption — and Disclosure Makes It Worse

Depletion corrupts judgment from the inside. Conflicts of interest do it from the outside — and the person whose judgment is corrupted usually can't tell.

She came in for a checkup. A young law student in Missouri, no pain, no symptoms — just a routine visit. Her dentist had recently bought a $40,000 CAD/CAM machine that mills custom ceramic crowns, and he'd spotted craze lines in her teeth: hairline fractures in enamel that are, in most cases, entirely harmless. He recommended a crown anyway. She trusted him. The crown made the tooth symptomatic. The tooth died. She needed a root canal. That failed. She needed another. That failed too. She graduated, researched her case, discovered she'd never needed the crown, and sued. She won.

The dentist almost certainly wasn't running a cynical calculation about recovering his equipment costs at his patient's expense. He'd spent a fortune on a device he believed in, and that investment quietly recolored his clinical judgment without him noticing. Craze lines that once looked like nothing now looked like a problem worth fixing. He probably felt he was helping her.

A conflict of interest doesn't bribe you. It reshapes your perception so you arrive at the self-serving conclusion through reasoning that feels honest. A pharmaceutical rep who takes a doctor fishing isn't buying a prescription; he's building a friendship, and the doctor who later prescribes his drug experiences it as loyalty, not a transaction. The corruption is invisible from inside.

You might think the fix is disclosure: if advisers announce their bias upfront, clients can adjust. In a jar-of-coins experiment, participants estimated the value of coins in a jar, guided by advisers paid more for higher estimates. Without disclosure, advisers inflated their guesses by about $4 over the neutral baseline. When required to announce the conflict upfront, they inflated by another $4 on top; disclosure appeared to license more bias, perhaps because the ethical obligation felt discharged. Clients did adjust downward, but only by $2, half of what was needed. They ended up worse off than before disclosure existed.

The combination is perverse: conflicts corrupt judgment without the corrupted person knowing it, and the standard remedy, transparency, licenses more corruption while leaving clients worse off than if no disclosure requirement had existed at all. The problem isn't bad people hiding bias. It's that bias hides itself, and announcing it doesn't fix that.

Watching Someone Cheat and Walk Away Doesn't Make You Angry — It Changes What Feels Normal

What do you think happens when you watch someone cheat openly and walk away with the money? The obvious guess is that you feel cheated too, indignant, perhaps more vigilant. The actual answer is that you cheat more. And how much more depends on whether that person is one of yours.

At Carnegie Mellon, researchers ran their standard math task: participants had five minutes to solve as many pattern-matching matrices as possible. But they added an actor named David to the room. About sixty seconds in, David stood up and announced he'd finished everything. Impossible in that time. Everyone knew it. He walked to the shredder, destroyed the evidence, pocketed the full cash envelope, and waved goodbye.

In the normal version of the experiment, people who could shred their own worksheets claimed to have solved around twelve matrices. After watching David, who everyone knew had fabricated his answers in full view of the room, participants claimed fifteen. Three extra matrices conjured from nowhere, purely from watching a brazen cheat succeed.

But here's where it gets more interesting. The researchers ran the same scenario with one change: David wore a University of Pittsburgh sweatshirt. The study was at Carnegie Mellon, whose students have a long-running rivalry with UPitt. Suddenly David wasn't one of them; he was an outsider. The result flipped. Participants claimed only nine matrices, barely above the no-cheating baseline. Observing the same egregious behavior, with the same absence of consequences, made CMU students slightly more honest — apparently out of a desire to distinguish themselves from whoever that guy was.

What changed wasn't the information. The participants learned, in both conditions, that cheating in this room carried zero consequences. What changed was the social signal. When the cheater was plausibly one of us, his behavior quietly revised what "people like us" do. When he was visibly not one of us, the effect inverted: he became a benchmark to distance yourself from.

The cumulative math is relentless. Major financial frauds are rare. But these social conditions run constantly and everywhere, quietly convincing millions of ordinary people that small cheats are normal. Each person who sees a colleague round up their billable hours, hears a friend inflate an insurance claim, watches a banker engineer a hidden fee and get a bonus — nudges the shared norm a fraction further. They were reading the room.

Moving a Signature Line Twelve Inches Up the Page Prevents More Fraud Than a Prison Sentence

The cheapest fraud-prevention tool ever tested costs nothing and takes about thirty seconds to implement: move the signature line from the bottom of a form to the top.

Princeton University has run its honor code since 1893. Incoming students receive a written constitution, mandatory lectures, small-group discussions, and a song performed by a campus theater group: a full week of ethics saturation before classes start. For the rest of their time there, they sign pledges on every paper and every exam, and receive reminder emails twice a year. Ariely waited two weeks after freshman induction, then gave Princeton students the same math task, and the same opportunity to cheat, that he'd given students at MIT and Yale, where no honor code existed and no one had received any ethics training. When Princeton students weren't asked to sign anything at that moment, they cheated just as much as everyone else. The week of moral immersion left no trace. What worked — at all three schools — was a single signature collected before any answers went on paper.

The Princeton finding should reframe how you think about fraud prevention. Ethics training, disclosure requirements, cultural campaigns — by the time any of that applies, the small rationalizations have already settled in. A signature at the bottom verifies a story that's already been told.

An insurance company gave Ariely's team 20,000 real odometer-reporting forms — the kind customers fill out so the company can calculate annual mileage and set premiums. Half had the pledge line at the bottom; half had it moved to the top. Customers who signed first reported driving an average of 26,100 miles. Customers who signed at the end: 23,700. A gap of 2,400 miles per customer, produced by shifting twelve inches of text. The company reviewed the findings and declined to change their standard forms.

The Author Caught Himself in the Acknowledgments

Here's what Ariely writes in his acknowledgments: he promises his wife he'll move some boxes to the attic tonight — then revises to tomorrow, then to the weekend, then adds "I promise." Three sentences. A live descent through optimism bias, temporal discounting, and rationalization, happening in print, to the man who spent a career studying exactly this.

Understanding why you deceive yourself doesn't stop you from doing it — it gives you the language to catch yourself mid-excuse. And it points toward the only intervention that actually works: not willpower, not ethics training, not stiffer penalties, but design. A signature at the top of a form. A cash register that creates a record. An environment where the honest path and the easy path are, by construction, the same path. You can't think your way out of a bad system. Build a better one.