How Bad Can the Owner Be Before the Business Pays?
redactedNvidia has public filings, earnings calls and an army of people watching it. The owner of the business your client wants to buy may have a good reputation and fifty Google reviews.
But how much does either of those things actually tell you about the person running the company?
You can reconcile the bank accounts. You can call customers, check public records and ask around town. Maybe the owner is a devoted family man, a regular at church and the person everyone calls when the local hockey team needs a sponsor.
All of that might be true. But it doesn’t tell you what that person does when nobody’s looking, or what might surface tomorrow. That’s an uncomfortable problem when you’re buying a company whose reputation, relationships and revenue may depend on one individual.
Consider what happened at Nvidia.
You already know the story. It’s been covered all over LinkedIn. On Sunday 8 November, Jensen Huang told Nvidia’s employees he had been unfaithful to his wife.[1] He didn’t name anyone. News reports had linked him to an executive at Megaspeed, a Singapore-based Nvidia customer, whom US officials reportedly believed was working for Chinese intelligence.[2]
The same day, the board opened an independent investigation into reports that he had been targeted by a foreign intelligence effort. It said it had no reason to believe national security or company secrets had been compromised, and that it had confidence in his leadership.[3]
On 11 November, The New York Times reported a second relationship, this time with an Nvidia executive who reported to him.[4] The next day, Nvidia announced Huang’s resignation and appointed CFO Colette Kress as interim CEO.[5]
None of that establishes that secrets were compromised or that any law was broken. But the founder who had led one of the world’s most closely watched companies for 33 years was out in four days.
Think about who was making the call. Not an outside buyer with three meetings and a data room. A board with full access, outside counsel and every incentive to get it right. On Sunday it said it had confidence in him. By Thursday he was gone.
Now imagine a privately held business without Nvidia’s management depth, governance or institutional resources. The owner might be the person who wins the contracts, settles the disputes, keeps the best employees around and takes the largest customers out for dinner.
What happens if something about that person becomes public tomorrow?
Do customers leave? Do employees reconsider? Does the bank get nervous? Or does everyone shrug and carry on? Those are very different outcomes, and the difference could be worth a substantial part of the purchase price.
That’s the tricky part. Not whether the owner is a good person, or whether their private life meets someone else’s moral standards.
It’s two things.
You can’t underwrite how a person behaves. Not from references, not from reputation and, as Nvidia shows, not even from the boardroom.
And even when an owner does behave badly, the damage to the business isn’t fixed. It depends on how much runs through that person, and whether the customers have anywhere else to go.
The first you can’t measure. The second you can.
They told everyone. In writing.
Here’s what got me thinking.
Nvidia had already warned investors about its dependence on senior leadership. Its annual 10-K explicitly states that the company is “highly dependent on the services of our longstanding executive team” and warns that poorly managed leadership transitions could undermine its future success.[6]
So the risk was disclosed. There was a board, public filings, quarterly earnings calls and an army of analysts watching the company.
Yet on the first trading day after Huang’s message and the board’s announcement, Nvidia’s shares fell 4.27%. That’s about $194 billion in market value, using the share count in its annual filing. The stock was already under pressure, so not all of that is Huang. But it’s a remarkable move for a risk that had been disclosed in writing.[6][7]
And here’s what makes it more interesting.
There had been earlier reporting. In October 2025, The New York Times reported that US officials were investigating whether Megaspeed was helping Chinese companies circumvent export restrictions. Singapore authorities were investigating too. Nvidia said its compliance team had found no evidence of diversion.[8]
That was more than a year before Huang’s resignation, and it involved the same company at the centre of this month’s story.
Were those warning signs? Of compliance exposure, perhaps. Not of personal conduct, and not of a leadership crisis.
And it’s tempting to connect those dots now.
Psychologist Baruch Fischhoff demonstrated this problem back in 1975: once we know how something ends, we become convinced the outcome was more predictable than it really was.[9] That’s hindsight bias, and investors, advisers and buyers are hardly immune. We can look back and assemble a convincing story from information that was available all along. It doesn’t mean we could have predicted what happened next.
Now consider the privately held business your client wants to buy.
There are no quarterly earnings calls, no team of analysts and possibly no independent board. Just financial statements, a few references and an owner who’s been running the place for twenty years.
If a board with that much access can go from confidence to resignation in four days, how much weight should we put on our read of someone we’ve met three times?
And more importantly, what happens to the business if we get that read wrong?
Take away the public record
Here’s a business I made up.
A heating and air conditioning company in a town of 30,000 people. The owner started it twenty-two years ago. Their name is on the vans. They coach the under-12 soccer team, they’re at church most Sundays and half the town has their number.
The business does $6 million in revenue. Your client wants to buy it.
What do you actually know about the owner?
There may be no independent board, no quarterly call, no reporters covering the business and no public archive of the owner’s explanations when things went wrong.
A US public reporting company generally has four business days to disclose its CEO’s departure. An ordinary private HVAC business has no equivalent obligation to announce it to the investing public.[10]
That doesn’t mean a private seller has no legal obligations. It means a prospective buyer cannot rely on the same public disclosure machinery.
Instead, you have the broker’s description, the owner’s account of how the company grew, customer reviews, referrals and perhaps a lender who has known them for years. Whatever an employee, a supplier or a former partner noticed may never reach the people assessing the dealredactedAll useful. Each answers a particular question.
A five-star review might tell you the technician arrived on time and fixed the boiler. It does not tell you who can create a supplier and approve its payments. A respected place in the community tells you how someone is regarded. It does not independently verify their explanation for a cash transfer.
The danger is letting an answer to one question stand in for an answer to another.
Public image is not private behaviour
Here’s where it gets uncomfortable. We all build a picture of people from what we can see.
Goes to church. Coaches the team. Family man. Pillar of the community.
None of that tells you what happens when nobody’s looking.
The person who passes the plate on Sunday can be the person who has been moving company cash for years. The family man everyone admires can be living a different life in private.
I’m not saying either is likely. I’m saying the public picture doesn’t tell you.
And the evidence agrees.
Every two years, the Association of Certified Fraud Examiners publishes a global study of fraud cases. Its 2024 edition covers 1,921 cases reported by examiners, rather than a representative sample of all businesses.[11]
Of the people who committed fraud, 87% had never been charged with or convicted of a fraud-related offense, and 85% had never been disciplined by an employer for fraud-related conduct.
When the fraud was committed by an owner or executive, the median loss was $500,000, the highest of any group. In organizations with fewer than 100 employees, the median loss was $141,redacted]
A clean background check is worth getting. It just can’t tell you everything.
The SEC even has a name for fraud that travels through a community: affinity fraud. It “exploits the trust and friendship that exist in groups of people who have something in common.”[12] The SEC was writing about investment scams, but the mechanism is the one a small-town reputation runs on: trust that comes from belonging.
Fraud is the extreme case. The same blind spot covers an affair, a bad debt or a comment that ends up on the town’s Facebook group. The people closest to the owner often don’t see it coming. That isn’t necessarily a failure of diligence. It’s the nature of the risk.
That’s why you can’t underwrite behavior.
So stop trying to predict the person. Ask the second question instead: if the worst does come out, how much does it actually cost the business?
Leaving is leaving
Back in September, I wrote about a study that relates to this piece.
Four economists, Matthew Smith, Danny Yagan, Owen Zidar and Eric Zwick, linked about eleven million US businesses to their owners through tax records. Then they looked at what happened when an owner died early.[13]
Profits fell 82%.
That was the article: “The Owner Died. 82% of the Profit Went With Them.”
But the part I keep thinking about is the other experiment in the same paper.
They also looked at owners who retired. Owners who stepped back and, as far as the tax records show, let a hired manager run things. The paper infers retirement from the owner’s wages stopping.
Profits fell 83%.
Death and retirement. Almost the same number.
The business doesn’t care why the owner is gone. Only that they are.
Two honest caveats. The death study focused on owners who earned more than $1 million in the year before their death, so these aren’t necessarily Main Street businesses. And the deaths occurred between 2005 andredacted]
I haven’t seen a comparable study on smaller businesses. But I can’t think of a reason a business where one person knows every customer by name would automatically be easier to hand over.
That’s the first thing you can measure: how much of the business leaves when the owner does.
There’s a third way an owner can leave, and the study doesn’t measure it.
They can stay.
And their reputation can go.
So how bad can the CEO be?
Say the worst is true. How much does it actually hurt the business?
No study answers this for small businesses. I looked.
But there’s enough research to sketch the shape of it. It comes down to two questions.
How much of the business depends on this person?
And can the customers go somewhere else?
Back to the HVAC company.
Say the owner gets caught up in something ugly, and it’s all over the town’s Facebook group.
It’s July. It’s 100 degrees. Your air conditioning just died.
Are you really going to wait two days for a company forty minutes away?
Probably not.
There’s research that points the same way. In a 2025 study, Felix von Meyerinck, Vesa Pursiainen and Markus Schmid tracked customer visits to 25,155 locations of companies hit by class action lawsuits. After the filing, visits fell about 4%.[14]
But the drop depended on the competition nearby. The less competition, the smaller the drop. In extreme cases with only one competing firm, the researchers found no observable reduction in customer visits.
When there’s nowhere else to go, people don’t go anywhere else.
Who got hurt matters too. John Armour, Colin Mayer and Andrea Polo studied UK enforcement cases against listed companies. The hit to a company’s reputation, measured in its share price, was nearly nine times the size of the fine. But only when the misconduct harmed the company’s own customers or investors. When the harm fell on third parties, they didn’t find the same effect.[15]
The market punishes harm to a company’s own customers and investors far more than harm to outsiders.
Now flip it.
Pizza has a substitute on every corner.
In 2018, Papa John’s North American comparable sales fell 7.3%, after an essentially flatredactedIts own annual report said negative publicity from statements by the company’s founder “contributed” to those results. Sales were already slowing, so “contributed” is the right word. The founder gave up the CEO job and then the chair, but his name was still on every box. And nobody had to drive forty minutes to try a different pizza.[16]
Can one person dent a high-end brand? After the controversy around its founder, Wynn Resorts told investors in its 2018 annual report that the matters “could significantly harm our business.” Nevada regulators fined the company $20 million.[17]
Then there’s Nvidia.
Its 10-K says two direct customers made up 22% and 14% of fiscal 2026 revenue. Its software platform, CUDA, has been around since 2006, with millions of developers building on it. Switching is expensive. Customers aren’t necessarily walking away over a headline.
But investors repriced it in a day anyway. In a private deal, the investor doing the repricing is your client.
And the same 10-K lists large cloud companies, including Alphabet, Amazon and Microsoft, among its competitors because they design their own chips. It says some customers “can use or develop their own solutions to replace those we are providing.”[6]
Lock-in delays the bill. It doesn’t cancel it.redactedPut it together and you get four boxes.
The owner leaves, and customers have options. The profit can follow the owner out, and nothing holds the customers in place while you rebuild.
The owner leaves, and the business is the only option. Customers may stay for now. But the licence, the know-how and the relationships still walk out the door.
The owner stays and their reputation goes, and the business is the only option. Customers grumble and keep calling. The staff, the lender and the next buyer may not be so forgiving.
The owner’s reputation goes, and customers have options. That’s Papa John’s: the founder left the top job, but his name stayed on the box.
Is there a right answer? No.
It depends on which box the business is in today. And the box can change. Nvidia moved from one row to the other in four days. A new competitor opens up. A big customer builds its own solution. A key technician leaves and takes the know-how with them.
I tried to prove myself wrong
I don’t want to oversell this.
Most of the research here is about public companies and very high earners.
Brandon Cline, Ralph Walkling and Adam Yore found that executive indiscretions came with “significant wealth deterioration, reduced operating margins, and lost business partners,” and that they happened more often at poorly governed firms. Those are public companies.[18]
Morten Bennedsen and his co-authors found that when a CEO was hospitalized, the company’s profits suffered, more so in family-controlled and people-heavy firms. When other senior executives were hospitalized, they found no similar effect. That’s Denmark, and it’s illness, not behaviour.[19]
Nobody has measured what an owner’s misconduct costs a $6 million business. The four boxes are my framework, not a finding. The class action study measures foot traffic, not revenue or profit. And none of the share-price moves here prove that a single event caused the whole change.
So here’s the claim I’m comfortable making.
One person can matter enormously in exactly the kind of business advisers work on every day: owner-run, people-heavy and known by reputation.
You can’t predict what that person will do. You can measure how much rides on them, and whether the customers have anywhere else to go.
You can’t underwrite behaviour. You can underwrite exposure.
A quality of earnings report ties the numbers. That’s the foundation, and nothing here replaces it.
This is a different question that sits beside it. Who made the decisions behind those numbers? And what happens if that person goes, or stays and becomes the story?
On the operating side, I’ve watched what happens when the decisions sit with one or two people. The numbers can look clean on paper while the organization is quietly breaking underneath.
Here’s what you can actually check.
How much depends on the owner?
• Compare the decision log to the org chart. Who actually approves payments, prices and hires?
• Run the holiday test. The last time the owner was away for a week, what waited until they got back?
• Price the missing manager. If the owner does two jobs, the second one needs a salary, and that salary affects the valuation at the applicable multiple.
Can the customers go anywhere else? For each top customer:
• Who holds the relationship, the owner or the company?
• Is there a contract? When does it renew?
• How long have they been a customer?
• What would it cost them to switch?
• Where is the nearest real alternative?
How far does the owner’s name reach?
• Whose name is on the vans, the reviews and the referrals?
• How much new business comes through the owner personally?
What doesn’t run through one person?
• Who else can approve a payment, get into the bank account or call a key supplier?
What the fraud examiners see most. In their 2024 study, the most common warning signs were living beyond one’s means (39% of cases), financial difficulties (27%) and an unusually close relationship with a vendor or customer (20%). Treat them as questions, not verdicts. The study can’t tell you how many honest owners show the same signs.[11]
What the deal does with the answers—the structure, the transition, the terms—is the deal team’s call.
But they can only make it if someone asked.
The risk no spreadsheet will ever capture
Nvidia had the disclosures, the board and the analysts. The board backed its founder on a Sunday. By Thursday he was gone.
Your seller has none of that scrutiny.
You won’t be able to predict every decision a person makes. No model will. No spreadsheet will ever capture all of it.
What you can see is how much rides on them, and whether the customers have anywhere else to go.
How bad can the CEO be before the business pays?
You’ll never know how bad the owner can be. You can know how much the business would pay.
Find that out first.
________________________________________
Sources
[1] NVIDIA, A Message to Our Company (Jensen Huang email to employees), NVIDIA Newsroom, 8 November 2026.
[2] Nvidia opens probe after report of CEO’s relationship with alleged Chinese spy, Honolulu Star-Advertiser, 8 November 2026; Nvidia CEO Jensen Huang admits to affair; board orders investigation, South China Morning Post, 8 NovemberredactedThe intelligence allegations have not been publicly confirmed, and China has denied them.
[3] NVIDIA, Statement from the Board of Directors, 8 November 2026.
[4] The New York Times, Jensen Huang Said to Have Had Relationship With Nvidia Executive Who Reports to Him, 11 November 2026.
[5] NVIDIA, CEO transition announcement, 12 November 2026; Form 8-K. The board accepted Huang’s resignation on 11 November and announced it on 12 November.
[6] NVIDIA Form 10-K, fiscal year ended 25 JanuaryredactedItem 1A (executive dependence, competition and customer substitution), Item 1 (business and technology), revenue concentration disclosures and cover-page share count. The filing reported 24.3 billion shares outstanding as of 20 February 2026.
[7] NVDA historical closing prices, Yahoo Finance. Closes of $redactedon Friday 6 November and $redactedon Monday 9 November 2026, the first trading day after the 8 November announcements. The approximate market-value change is calculated using the $7.97 price difference and the 24.3 billion shares disclosed in [6]; it is not a causal estimate of the news’s impact.
[8] Ana Swanson, Tripp Mickle, Paul Mozur and Mara Hvistendahl, A Mystery C.E.O. and Billions in Sales: Is China Buying Banned Nvidia Chips?, The New York Times, 9 October 2025.
[9] Baruch Fischhoff, Hindsight ≠ Foresight: The Effect of Outcome Knowledge on Judgment Under Uncertainty, Journal of Experimental Psychology: Human Perception and Performance, 1975.
[10] US Securities and Exchange Commission, Form 8-K guidance, Item 5.02; Investor Bulletin: How to Read an 8-K.
[11] Association of Certified Fraud Examiners, Occupational Fraud 2024: A Report to the Nations. Figures on criminal and employment histories, perpetrator position, organization size and behavioural red flags.
[12] SEC Office of Investor Education and Advocacy, Investor Alert: Affinity Fraud, 18 June 2014.
[13] Matthew Smith, Danny Yagan, Owen Zidar and Eric Zwick, Capitalists in the Twenty-First Century, Quarterly Journal of Economics, 2019. See also the University of Chicago research summary for the 82% and 83% estimates.
[14] Felix von Meyerinck, Vesa Pursiainen and Markus Schmid, Competition and the Reputational Costs of Litigation, Journal of Financial and Quantitative Analysis, 2025.
[15] John Armour, Colin Mayer and Andrea Polo, Regulatory Sanctions and Reputational Damage in Financial Markets, Journal of Financial and Quantitative Analysis, 2017.
[16] Papa John’s Form 10-K for 2018; 26 February 2019 results announcement.
[17] Wynn Resorts Form 10-K for 2018, risk factors and regulatory disclosures.
[18] Brandon N. Cline, Ralph A. Walkling and Adam S. Yore, The Consequences of Managerial Indiscretions: Sex, Lies, and Firm Value, Journal of Financial Economics, 2018.
[19] Morten Bennedsen, Francisco Pérez-González and Daniel Wolfenzon, Do CEOs Matter? Evidence from Hospitalization Events, Journal of Finance, 2020.
The HVAC company is fictional. The four-box framework is the author’s interpretation, not an empirical finding. Associations are not necessarily causal: no share-price, revenue or profit figure in this article establishes that a single reported event caused the entire measured change.