Nobody Lies to a Founder More Than the Founder: The Four Written Rules That Stop It

Key Insights

• Every founder has someone checking their decisions, but that person usually answers to the founder. The market will eventually show you when you’re wrong; by then it’s too late, too costly, and public. A written rule catches the mistake earlier, cheaper, and behind closed doors.
• In 2025, roughly 98,700 companies left Hong Kong’s register. That figure sits in the same official press release as the record-high incorporation number, and nobody quotes it.
• A one-member company has no board, no dissenting shareholder and no AGM challenge. The founder is the only person deciding whether the warning signs matter.
• This article converts four Stoic virtues into four rules you can write down tonight, plus four clauses Get Started HK recommends putting into a shareholders’ agreement before the company exists.

Founders fail less from bad ideas than from unexamined ones, and in a Hong Kong SME nobody audits the founder’s reasoning — around 76% of newly incorporated companies have exactly one member. The fix is conversion, not motivation: turn each of the four Stoic virtues into one written, enforceable rule — a metric set, a wage ratio, an annual client review, and a spending cap. Then write them into your shareholders’ agreement at incorporation, not after the dispute.
By the Get Started HK editorial team. Get Started HK has supported 46,000+ entrepreneurs through Hong Kong company incorporation and fintech onboarding; the patterns in this article come from that work. Figures are from Hong Kong Government primary sources, listed at the end. This is general information, not legal or tax advice.

How Many Hong Kong Companies Quietly Left the Register in 2025? About 98,700

In 2025, Hong Kong’s Companies Registry recorded 195,343 newly registered local and re-domiciled companies. Over the same year, the total number of companies on the register rose by 96,609, to an all-time high of 1,557,103. Both numbers are in the same press release, dated 16 January 2026.

Do the simple subtraction:

195,343 − 96,609 = 98,734

That means about 98,700 companies disappeared from the register in one year through deregistration, striking-off or winding-up. You probably didn’t hear about them. Nobody posts a deregistration on LinkedIn. The incorporation number gets the headline; the other number sits in the same paragraph, in the same font, and then nobody talks about it.

What does the 98,700 figure actually include?

To be clear, this is Get Started HK’s derived calculation, not an official “business failure” statistic. The Companies Registry does not publish a single failure count, and the net figure bundles together very different exits:

• Old dormant companies finally being cleaned up
• Holding vehicles closed after a restructuring
• Owners who chose to shut down cleanly and solvently
• Companies that ran out of cash
• Companies where the founders fell out
• Companies where people stopped being honest with themselves

So no, not all of these were failures. But almost none of them were publicly called failures either, and that is the point. Failure in a Hong Kong SME is private, and privacy is what lets self-deception run for years.

What Do Four Struggling Founders Have in Common? Nobody Checked Their Thinking

These are composites drawn from patterns Get Started HK sees repeatedly across the founders we onboard. Details are changed; the shape is familiar. They appear in the same order as the four rules later in this article, so each story is the problem and the matching rule is the fix.

1. The profitable company with no cash. HK$2 million revenue, growing, healthy on paper. Customers paid at 60 days; suppliers had to be paid at 30. The P&L looked fine the whole way down. The lie was: “Revenue is up, so we’re fine.” This is the failure Rule 1 is built to catch: the founder had numbers, but not the ones capable of delivering bad news.

2. The 19-month agency. Two friends, 50:50, no vesting, no stated time commitment, no agreement about what happens if one of them stops. In month seven, one took a full-time job and kept half the shares. The company still exists, barely; they can’t sell it, and they can’t easily close it either, because voluntary deregistration in Hong Kong requires every member to agree. The lie was: “We don’t need paperwork, we trust each other.”

There is a second cost here that founders only discover at the fintech onboarding stage. Hong Kong banks and fintech platforms generally require identity verification from every individual holding 25% or more of the company. An absent 50% co-founder who stops answering messages doesn’t just block deregistration; he blocks the business account, the payment gateway and the multi-currency wallet too. This is the failure Rule 2 is built to catch: nobody had written down what each founder was owed, or what they owed in return.

3. The prestige client. A design studio with a famous logo on its website. That client consumed 59% of billable time at the worst margin per hour of any account. For two years, leaving felt impossible. The lie was: “We can’t afford to lose them.” This is the failure Rule 3 is built to catch: the founder knew the right decision and had no date on which to make it.

4. The HK$41,000 a month nobody decided. A trading company with 23 software subscriptions, each added by someone for a good reason, none ever reviewed. Nobody had approved HK$492,000 a year on tools. Everybody had approved HK$1,800 here and HK$3,500 there. The lie was: “It’s only a small monthly charge.” This is the failure Rule 4 is built to catch: no single decision was wrong, and no one was checking the total.

None of these founders were foolish. Not one of them had anybody properly checking their thinking.

Why Does a One-Member Company Have No One to Say No?

Hong Kong has around 357,077 SMEs employing about 1,167,744 people, which is more than 98% of all business establishments (Trade and Industry Department, March 2026). In the Companies Registry’s 2024-25 financial year, 154,220 new companies were incorporated, and around 76% of them were one-member companies.

A one-member company has one shareholder. In practice, that usually also means:

• One director
• One decision-maker
• One person setting the strategy
• One person deciding whether the business is doing well
• One person deciding whether the warning signs matter

There is no board. No one challenged the founder, and no minority shareholder asking awkward questions at the AGM. It means running a company without an opposing voice at exactly the moment you need one most.

Why Does Nobody Sell Governance to a One-Person Company?

Here is the gap in the market. A large company gets six mechanisms as standard, and a founder gets none of them:

• A board to challenge the person in charge. You have nobody.
• A finance team to force honest numbers. You have yourself, on a Sunday.
• Salary benchmarking to price senior labour. You have a guess.
• A capital committee to kill bad projects. You have your feelings about money already spent.
• Lawyers who’ve seen founder disputes. You have an old WhatsApp thread.
• A strategy cycle to decide when to stop. You have exhaustion.

Every one of those is written down. Every one exists because large organisations eventually learned they cannot rely on one person’s judgement, including a good person’s. Yet the incorporation industry in Hong Kong is built around speed: file the NNC1, collect the certificate, move on. Governance is treated as something a company grows into later, and later is usually after the argument.

Why Does a Value Only Count Once It’s Written Down?

Stoicism was founded in Athens around 300 BC and later spread through the Roman Empire. Its central practical move is separating what you control from what you don’t, which is useful for founders because it forces you to see a situation before reacting to it. Founders spend their lives in exactly the material the Stoics wrote about: money, power, pressure, contracts, staff, difficult partners.

The Stoics organised the practice around four virtues:

• Practical Wisdom — seeing what is actually there
• Justice — giving people what they are owed
• Courage — doing the right thing when it costs you
• Temperance — declining what you can afford but don’t need

Here is the problem with virtues as motivation. “Be honest with yourself” has never once stopped a founder from lying to themselves, because lying doesn’t feel like lying. It feels like optimism, or loyalty, or a busy quarter. So convert each virtue into a rule with a number and a date: a rule can be checked, but a value can only be believed.

Virtue	The rule	The number	The date
Practical Wisdom	Four numbers, one page	Pay/hour, cash days, top-client %, one leading indicator	Every Monday
Justice	Pay yourself a market wage before calling anything profit	Benchmark for your role	Once a year, before pricing
Courage	Release the lowest-paying client	Bottom of the pay-per-hour ranking	A fixed calendar date
Temperance	Any new recurring cost waits	~1% of revenue threshold	90 days

Rule 1 — What Four Numbers Should a Founder Check Every Monday?

Practical Wisdom means seeing what’s actually in front of you. In business, your numbers must be capable of telling you something you don’t want to hear. Most founder dashboards can’t: revenue, enquiries, followers, headcount, none of them tells you what to stop doing. Once you’re past survival, stopping is the only decision left, because your constraint isn’t money any more. It’s your hours.

Track four:

1. Pay per hour on your most limited resource. Usually you. Rule 2 tells you what to compare it against.

2. Cash days on hand: how long you survive if income stops tomorrow. In Hong Kong this number has a specific trap in it, because profits tax is charged on a provisional basis. You prepay the current year based on last year’s assessable profits, so a company that grew last year pays tax on that growth while simultaneously funding this year’s receivables. Story 1 wasn’t just a receivables problem; it was a receivables problem with a provisional tax demand sitting on top, sized by the good year that caused it.

3. Biggest client as a percentage of revenue. Anything above roughly a third means one phone call can end the company.

4. One leading indicator, whatever predicts next quarter for your business. If you’re near HK$2 million of assessable profits, forecast that line specifically. Under Hong Kong’s two-tiered profits tax regime, a corporation pays 8.25% on the first HK$2 million and 16.5% above it (7.5% and 15% for unincorporated businesses). It is the only genuine cliff edge in the Hong Kong system, and the lower band is worth up to HK$165,000 a year to a corporation. Crossing it in December because of one rushed invoice is an expensive way to find out you weren’t tracking it.

The test: can your numbers tell you which work to turn down? If not, they’re decorations.

Rule 2 — Are You Making a Profit, or Just Paying Yourself Badly?

Justice is giving people what they’re owed. Founders are good at this: they pay staff on time and over-deliver for clients. Then they leave one person off the list.

A consultant bills HK$1,000,000. Costs are HK$400,000. She pays herself HK$200,000 and calls the remaining HK$400,000 profit.

Now ask what she’d have to pay someone else to do her job properly. You don’t have to guess. The Census and Statistics Department runs the Annual Earnings and Hours Survey (AEHS) every year, sampling around 11,000 business undertakings, and publishes hourly and monthly wage percentiles by occupational group and industry section. In the 2025 edition, the median monthly wage was HK$21,200, the median hourly wage HK$85.7, and median weekly hours 43.2 (up from HK$20,500, HK$82.9 and 43.4 in the 2024 edition). The managers-and-professionals rows are free salary benchmarking, published by the government, that most founders have never opened.

Say the survey puts her role at HK$600,000. Redo the arithmetic:

Line	HK$
Revenue	1,000,000
Real market wage for her role	(600,000)
Other costs	(400,000)
Profit	0

Everything she took home was wages. There was never any profit.

How does the Hong Kong tax code enforce the illusion?

If she trades as a sole proprietor, the HK$200,000 she paid herself is not an allowable deduction. Under the Inland Revenue Ordinance, a proprietor’s own drawings are an appropriation of profit, not an expense, so they are added back when computing assessable profits. Her assessable profits are therefore HK$600,000, taxed at 7.5% within the first HK$2 million: roughly HK$45,000 of profits tax on HK$600,000 that was never profit. She is paying tax on “profit” for the privilege of earning a wage.

Had the same work run through a company, a market-rate director’s fee would be deductible remuneration for the company and assessed under salaries tax in her own name, where personal allowances apply. Identical economics, different label, materially different outcome. And the label is a structural decision made at incorporation, not something you fix in March. This is one of the first questions Get Started HK asks a founder before filing anything, because it cannot be answered by the filing itself.

General information, not tax advice; the treatment depends on your facts and a Hong Kong tax adviser should confirm it.

What does the zero actually mean?

She doesn’t own a business. She owns a job that she also does the admin for. That distinction isn’t semantic: a job pays you while you work, and a business is worth something when you stop. Once the real wage is visible, the options are obvious — charge more, change who you work with, or sell something that isn’t your hours. None of them involve working harder.

The rule: do this once a year, in writing, against the published AEHS figure, before you set next year’s prices.

Rule 3 — Which Client Should You Release This Year?

Courage in business is rarely a bold pivot. It’s usually doing the thing you already know is right, on a day you chose in advance. Pick a date now; on that date, rank clients by pay per hour and release the bottom one.

It has to be a date, because releasing a bad client is always right in principle and never convenient in practice. There is always a reason to wait one more quarter. “When it feels right” never arrives. A date arrives.

When one client eats most of your capacity, you’re not running a business, you’re an outsourced department with none of the job security. Rule 1 gives you an objective floor for the ranking, so the decision rests on a number rather than a feeling. Story 3’s studio didn’t need more courage. It needed a date in the diary and a spreadsheet column it had never built.

Rule 4 — Why Does a HK$1,800 Subscription Cost More in Hong Kong Than in London?

Temperance isn’t giving things up. It’s declining things you can afford but don’t need. Any new recurring cost waits 90 days: write it down, date it, look again.

Recurring is the word that matters. One-off costs are decisions. Recurring costs are tenants: they move in, contribute nothing, and stay until you evict them. A monthly tool is really its annual cost multiplied by however many years you forget about it, and Story 4’s HK$41,000 a month was never one decision but 23 small ones nobody re-examined.

Here is the Hong Kong twist, and it runs against intuition. Hong Kong has no GST, no VAT and no sales tax, and founders treat that purely as good news. For recurring costs it isn’t: a London company reclaims 20% VAT on that subscription as input tax, while a Hong Kong company has nothing to reclaim, so the sticker price is very close to the real price. Your only relief is the profits tax deduction, worth just 8.25 cents in the dollar to a corporation below HK$2 million of assessable profits, and 16.5 cents above it.

In other words, recurring costs are more expensive after tax for small Hong Kong companies than for large ones. The discipline is hardest to justify at exactly the size where it pays most. The rule: set the waiting-period threshold as a percentage of revenue — around 1% is a sensible start — so it grows with you.

How Do You Turn Four Rules Into Four Shareholders’ Agreement Clauses?

A rule on a page binds you. A rule in a shareholders’ agreement binds everyone. If there is more than one of you, translate the four rules into four clauses at incorporation, while you still like each other. These are the plain-English versions Get Started HK walks founders through; a Hong Kong lawyer should draft the operative wording.

Rule	Clause type	What it says in plain English
Practical Wisdom	Information rights	Management accounts containing the four numbers are circulated to all shareholders by the 10th of each month.
Justice	Founder remuneration and vesting	Founders draw a salary benchmarked annually against the AEHS figure for their role; shares vest over time and lapse if a founder stops working in the business (the clause Story 2 never had).
Courage	Annual review and exit mechanics	A fixed date each year to review client concentration and, separately, defined leaver and deadlock provisions so nobody is trapped in a 50:50 stalemate.
Temperance	Reserved matters / spending cap	Any recurring commitment above X% of trailing revenue requires written approval from all shareholders after a 90-day notice period.

The cost of adding these at incorporation is small. The cost of adding them after the dispute is usually the company.

Who Should Use These Rules, and When Do They Not Apply?

These rules are for the 76%: the one-member company where the founder is also the board, and for any two- or three-founder company that has not yet had its first real disagreement. They apply most to service businesses selling hours, because the wage illusion in Rule 2 is most severe when the founder is the product. They apply to trading companies through Rules 1 and 4, where cash days and recurring costs decide survival.

A system you’re not allowed to question is a belief, not a system. Four limits:

• Don’t fire a client if you have fewer than four. With two, that’s not courage, it’s arithmetic. Below four, the rule inverts: win clients until you can afford to choose.
• Don’t apply the wage rule before you have revenue. For the first six to nine months you are the investment, and underpaying yourself is legitimate funding. The discipline is writing down the date it ends.
• Raise your 90-day threshold as you grow. A fixed cap on an HK$8 million business isn’t temperance, it’s timidity.
• Four numbers can blind you. They’re a decision filter, not a picture of the business. Once a quarter, look deliberately outside them; that’s usually where the thing that kills you is waiting.

What Should You Write Down Tonight?

The 98,700 companies that left the register last year mostly didn’t lose to smarter competitors. They lost to a version of themselves that nobody ever contradicted, for 19 months, or 36, or nine years. You can’t make yourself objective; nobody can. What you can do is move the judgement out of your head, where it’s negotiable, and into a number and a date, where it isn’t.

So write the smallest version tonight. One page. Your real market wage, your cash floor, and a date next year when the bottom client goes. If there’s more than one of you, add the four clauses while you still like each other.

Get Started HK has helped 46,000+ entrepreneurs incorporate in Hong Kong and get onboarded to the traditional banks and neobanks they need to trade globally, and the companies that last tend to have made these decisions before the certificate of incorporation arrived, not after the first argument. That is the role we’ve chosen in the startup ecosystem: not the fastest filing in Hong Kong, but the operator who has seen both the incorporation and what happened twelve months later, and who tells you on day one.

A value you believe is a hope; a rule you’ve written down is a board.

FAQ for Startup Founders

How many companies close in Hong Kong each year?

The Companies Registry does not publish a single closure figure. Subtracting the 2025 net increase in the register (96,609) from 2025 new registrations (195,343) implies roughly 98,700 companies left the register through deregistration, striking-off or winding-up. Get Started HK derived this figure from published data; not all of those exits were business failures.

What percentage of new Hong Kong companies have only one shareholder?

Around 76% of the 154,220 companies incorporated in the Companies Registry’s 2024-25 financial year were one-member companies. In practice this usually means one director and one decision-maker, with no board or minority shareholder to challenge the founder. That is why Get Started HK recommends written rules even for solo founders.

What is the two-tiered profits tax rate in Hong Kong?

Corporations pay 8.25% on the first HK$2 million of assessable profits and 16.5% above that threshold. Connected entities can nominate only one entity to benefit from the lower band, so group structures need to plan for it.

Where can a founder benchmark their own salary in Hong Kong?

The Census and Statistics Department’s Annual Earnings and Hours Survey publishes wage percentiles by occupational group and industry section, free of charge. The 2025 edition reports a median monthly wage of HK$21,200 and a median hourly wage of HK$85.7. Founders can use the managers-and-professionals rows to test whether their business makes a profit after paying them a market wage.

Does a co-founder who leaves affect business bank or fintech account opening?

Yes. Hong Kong banks and fintech platforms generally require identity verification from every individual holding 25% or more of the company. If a departed co-founder still holds 50% and won’t cooperate, the application usually stalls, which is why Get Started HK maps controller KYC at incorporation rather than at onboarding.