Keeping It

Why Your Most Trusting Clients Are the Ones You Can Lose Overnight

How people react to good news and bad news about you depends almost entirely on one thing: how much they trusted you to begin with.

Yang, F., Rosenblat, T., & Weng, W. (2025). Managing trust under dynamic information. Humanities and Social Sciences Communications, 12, 1964. Open access: doi:10.1057/s41599-025-06248-y

Picture two prospects walking into your office.

The first arrives skeptical. Maybe he's been burned before. A broker churned his account, or a planner vanished after the paperwork was signed. He's polite, but his arms are crossed.

The second arrives warm. He was referred by his brother-in-law, he went to the same college you did, and he's already halfway to signing.

Conventional wisdom says the second client is the easy one. New research says: not so fast. The skeptic may be the more durable relationship. The warm referral may be the one you can lose with a single stumble.

The Experiment: Trust with the Dial Turned Up and Down

Economists Fanzheng Yang, Tanya Rosenblat, and Weiwei Weng wanted to answer a question that most trust research skips entirely. We know a lot about whether people trust strangers at any given moment. But trust isn't a snapshot. It's a moving picture. How does it change as new information rolls in, when that information is imperfect the way real life always is?

To find out, they ran a classic "trust game" in the lab with 168 university students, with a twist. Participants first stated how likely they thought it was that an anonymous partner would treat them fairly. Then, over four rounds, they received noisy signals about that partner's true character. The signals were accurate only two-thirds of the time, like a customer review that's probably legitimate, or a rumor that's probably true. After each signal, participants revised their beliefs.

The researchers got to watch trust update itself, in real time, under the kind of murky information we all live with.

The Finding: Three Kinds of Trusters

At first glance, the results looked boring. On average, people updated their beliefs like textbook rational actors: good news nudged trust up, bad news nudged it down, in tidy symmetrical fashion.

Split participants by their starting trust level, though, and three distinct characters appear:

The Skeptics (low initial trust) responded strongly to good news but largely shrugged off bad news. They'd already priced in disappointment. What moved them was pleasant surprise.

The Moderates (middle trust) updated evenly in both directions, the closest thing to walking calculators the study found.

The Believers (high initial trust) were the mirror image of the skeptics. They reacted sharply to bad news but barely budged on good news. Positive signals just confirmed what they already assumed. Negative signals hit like a betrayal.

The researchers argue this lopsidedness isn't irrational. It's self-correcting. If you started out over-trusting, being hair-trigger sensitive to bad news protects you from exploitation. If you started out under-trusting, being extra receptive to good news keeps you from missing out on genuinely good partners. Each bias contains its own antidote.

Then They Added Tribe

In a second experiment, the researchers revealed each partner's university affiliation, turning anonymous strangers into "one of us" or "one of them."

Actual trustworthiness didn't change a bit. Students from both schools behaved almost identically as partners. But perceived trustworthiness swung hard. Participants matched with someone from their own university started with significantly inflated trust: a genuine "in-group bonus" of unearned confidence. Those matched with outsiders started lower.

One finding should make an advisor sit up: the in-group bonus came with a hidden cost. Because same-school partners started in the "Believer" zone, they became hypersensitive to negative signals. The unearned trust was withdrawn faster than trust that had been built from scratch. Meanwhile, the out-group condition, where everyone started skeptical, was the one where positive feedback did the heaviest lifting.

Shared identity doesn't change how people process information about you. It just changes where they start. Where you start determines what can hurt you.

What This Could Mean at Your Desk

Your referral clients could be your fragile clients. The prospect who arrives pre-sold through a referral, a shared alma mater, or a country club connection starts as a Believer. That's wonderful for closing. But Believers underreact to your wins and overreact to your slips. One missed callback, one confusing statement, one surprise fee can do disproportionate damage. The relationship that came easiest is the one that needs the most careful maintenance.

Your skeptics are winnable, and sticky. A doubtful prospect isn't a lost cause; she's someone whose updating machinery is tuned to reward you. Every promise kept, every proactive call, every moment of transparency registers strongly. And once won, her trust rests on evidence rather than affinity, which makes it far harder to shake.

Don't confuse the honeymoon with the marriage. In-group warmth is real, and it opens doors. But this research suggests it's best understood as a loan, not a gift. It gets repaid, or called in, based on what you actually do.

Manage the negative signal, not just the positive one. For your high-trust clients, the priority isn't more good news; they discount it anyway. The priority is preventing, softening, and rapidly explaining anything that could read as a negative signal. Get ahead of the confusing statement. Call before the market headline does.

The Bottom Line

Trust between strangers isn't as mysterious, or as fragile, as the old saying suggests. It's fragile when it starts high and surprisingly resilient when it starts low. People aren't fickle; they're running a quiet, sensible program that corrects for their own starting biases.

For advisors, the takeaway is almost paradoxical: cherish your skeptics, and handle your believers with care.

The Magic Words That Let You Change Your Mind Without Losing the Client

New research shows there's a way to abandon a failing strategy and come out looking like a person of integrity, but you have to say the words before things go wrong.

Kristal, A. S., & Dorison, C. A. (2025). Precommitment can allow decision makers to maintain trust when de-escalating commitment. Journal of Applied Psychology, 110(10), 1395-1411. doi:10.1037/apl0001243

Every advisor knows the trap.

You recommended the fund. You championed the strategy. You put your name on it. And now, two years in, the evidence says it's time to get out. But a voice in your head whispers: If I reverse course now, what will the client think of me?

That voice isn't paranoid. It's right, mostly.

Decades of research confirm that people who abandon a course of action are trusted less than people who stubbornly stay the course, even when quitting is obviously the smarter move. We reward consistency and punish flip-flopping, and decision makers know it. That's part of why history is littered with monuments to not-quitting: the Concorde, the Vietnam War, and the Shoreham Nuclear Power Plant, which ballooned from a $75 million estimate to over $4 billion, and then never opened.

But a pair of researchers, Ariella Kristal of Columbia and Charles Dorison of Georgetown, found a specific communication move that lets a decision maker pull the plug and keep their reputation for integrity intact. Across nine preregistered experiments with nearly 7,800 participants, the move worked again and again.

It's called precommitment. And it's simpler than it sounds.

The Experiments: Shark Tank, Governors, and Real Money

The researchers started by confirming the problem. In one study, 1,600 people watched real Shark Tank pitches, then learned that a competitor had launched a better, cheaper version of the product after 90% of the entrepreneur's funding was spent. Half read that the entrepreneur kept investing anyway; half read that she shut the project down.

The verdict was clear: the quitter was rated as having significantly less integrity than the entrepreneur pouring good money after bad. Doing the financially rational thing made her look untrustworthy.

Then came the twist. For some participants, one sentence was added: the entrepreneur had publicly pledged in advance to stop the project if market conditions turned unfavorable. With that single sentence, the penalty didn't just disappear. It reversed. The de-escalating entrepreneur was now rated as more trustworthy than the one who kept spending.

Trust, it turns out, isn't about staying the course. It's about word-deed alignment. Someone who says "we're building this" and then quits looks like they broke their word. Someone who says "we're building this, and here's exactly when we'd stop" and then stops at exactly that point has kept their word by quitting.

This wasn't just talk. In a follow-up study using an economic trust game with real money on the line, participants chose which stranger to entrust their cash to: one who had escalated a doomed project or one who had de-escalated. Without precommitment, only 28% trusted the quitter. With precommitment, that jumped to about 41%: a 46% relative increase in people putting real money on the de-escalator's integrity.

The Recipe: Two Ingredients, No Substitutions

Not just any hedge works. The team tested watered-down versions and found that an effective precommitment requires exactly two components:

It must be conditional, tied to a concrete, observable future state of the world. "We'll revisit this at some point" doesn't work. "If a faster, more economical option becomes available, we'll stop spending" does.

It must be specific, naming the actual action you'll take. A vague pledge to "review the project" if conditions change earned no trust protection. The pledge has to say what you'll do: stop, exit, switch.

Strip out either ingredient and the benefit largely evaporates. This also rules out a cynical interpretation: it's not that audiences forgive anyone with a plausible excuse. They reward one thing: a promise made in advance and visibly kept.

A supplemental study found the protection even held when the specific condition named didn't come to pass but other legitimate adverse conditions did. Setting exit criteria up front signals a certain kind of character, and observers give credit for it.

The Catch: Confidence Is a Currency Too

Before you rewrite every client agreement, one warning. In a lab study with business students working for a "manager," the researchers found that announcing a precommitment at the start of a project came at a cost: workers perceived their manager as less confident in the project and reported feeling less motivated. (Their measured effort didn't drop, but the perception shift was real.)

Exit criteria whispered at the kickoff can sound like doubt. The hedge that protects you later can deflate the room today.

What This Means at Your Desk

Build the exit ramp into the recommendation. When you propose a strategy, fund, or manager, state the concrete conditions under which you'd move on: "If this underperforms its benchmark for six consecutive quarters, we exit." That sentence costs you nothing today and buys you credibility on the day you need to change course.

Make it conditional and specific, in writing. Vague review language ("we'll monitor this closely") provides zero protection. Name the trigger, name the action. An Investment Policy Statement is one of the most underrated trust technologies in the profession.

When you de-escalate, quote yourself. The trust benefit comes from the audience seeing the word-deed alignment. Reference the original pledge explicitly: "When we adopted this position, we agreed we'd exit if X occurred. X has occurred."

Pair the hedge with conviction. To avoid the confidence penalty, frame exit criteria as discipline, not doubt: "I believe strongly in this strategy. Part of why you can trust that belief is that I've defined exactly what would change my mind."

The Bottom Line

Clients don't punish you for changing your mind. They punish you for appearing to break your word. Define in advance what would make you change course, and the day you finally do, you won't look like a flip-flopper.

You'll look like what you promised to be.

The Case for Saying Sorry

Two big experiments settle an old debate: when your firm is caught in a gray area, an apology beats a defense, every time researchers checked.

Hornsey, M. J., Chapman, C. M., La Macchia, S., & Loakes, J. (2024). Corporate apologies are effective because reform signals are weighted more heavily than culpability signals. Journal of Business Research, 114620. Open access: doi:10.1016/j.jbusres.2024.114620

Something has gone wrong on your watch: not a clear-cut fraud, but one of those murky situations where reasonable people could argue about how much blame you deserve. Your lawyer whispers one thing: Never admit fault. Your conscience whispers another: Say you're sorry.

For twenty years, the research seemed to side with the lawyer. Influential studies warned that apologizing for an ethics lapse "seals in" your guilt. A denial, the thinking went, at least preserves the benefit of the doubt. One well-cited study even found that denials outperformed apologies when a company was accused of bribery, despite strong evidence the company did it.

A team of Australian researchers led by Matthew Hornsey suspected something was off. Most of those cautionary studies, they noticed, weren't actually about companies apologizing to the public. They were about individuals: a job applicant accused of fudging a tax return, an employee blaming a boss. And they usually forced participants into a stark choice: the accused was either innocent or lying. Real corporate scandals rarely work that way.

So the team built a more lifelike test.

The experiment

More than 1,400 Americans read what they believed was real news: an online fashion retailer called GlobalDress had been sourcing from a Bangladeshi factory with abusive conditions and structural flaws that eventually killed workers. The scenario was modeled on the actual Rana Plaza collapse. The company's guilt was deliberately left ambiguous. The question wasn't whether the factory was terrible, but whether GlobalDress should have known.

Participants then saw one of three company responses: a full apology (fault acknowledged, remorse expressed, a promise to make sure it never happened again), a polished non-apology touting the company's "commitment to ethical sourcing," or silence.

Then everyone rated the same things: How guilty is this company? How likely is it to change? And the questions that matter to any business: do you trust them, and would you buy from them?

What they found

The lawyers were half right. Apologizing did make people see the company as more culpable. That downside is real.

But that's not what made a difference. When the researchers traced what drove trust and buying intentions, the guilt signal barely moved the needle. What moved it was the reform signal: the sense that this company had learned its lesson and wouldn't do it again. And only the apology delivered that signal.

The result: the apologizing company earned significantly more trust and consumer support than the company that deflected, and far more than the one that said nothing.

A second experiment added a wrinkle. Maybe apologies only win when the evidence is airtight anyway? The researchers varied whether the allegations were "clearly shown" by documents or merely "unconfirmed reports." It made no difference. The apology won either way, even when a denial would have worked.

One more finding deserves a place on every crisis-response whiteboard: the company that stayed silent was judged just about as guilty as the company that apologized. In the absence of a response, the public defaults to guilt. Which means the supposed "cost" of admitting fault is an illusion. You're already paying it.

Why it matters for advisors

Trust in financial services lives and dies on these gray-area moments: the fee that wasn't explained clearly, the product that underperformed its pitch, the operational slip that cost a client money. The instinct to lawyer up and minimize is strong. This research says the instinct is backwards.

Clients aren't running a courtroom in their heads, tallying your guilt. They're running a forecast: Will this happen to me again? A genuine apology, one that names the failure and commits to a fix, answers the question they're actually asking. A defensive statement answers a question nobody asked, and leaves the forecast dark.

The takeaway

The authors put it best: this is one of those pleasing cases where honor and profit point the same direction. When responsibility is murky, saying sorry isn't just the decent move. It's the smart one.

One caution: the studies covered ambiguous, gray-area transgressions. If accusations against you are flatly false, deny them. Anything else is dishonest. And if the fault is fully yours, apologize because it's right, not because it polls well.