1,527 joined
✓ You're in
MOAT AI Research · Empirical Study · n = 1,500

At least 245,760 ways your business dies. Most of them are silent.

We studied 1,500 companies across 30 industries using logistic regression, Fisher’s exact test, and Spearman correlation analysis. One finding held across every industry, every revenue range, and every business model: the problem was never the product. It was always upstream.

If you’re lying awake at 3AM staring at the ceiling, wondering why revenue won’t move despite doing everything right — the answer is in this research. And it’s not what you think.

Read the findings ↓  or join the study above ↑

0
Companies studied
0
Industries analysed
Risk multiplier
0%
Model accuracy
0
AUC score
Finding 001 — The Killer Number

A company with a broken customer definition is one hundred and four times more likely to stall.

Not 2×. Not 10×. One hundred and four times more likely to be stuck in a pattern of silent revenue decline — working harder every year, trying every fix, watching nothing change.
Fisher’s exact test · p < 10⁻⁴⁴ · n = 1,500 · 30 industries · 95% CI

For context, the threshold for statistical significance is p = 0.05. This finding is forty-four orders of magnitude past that line. The signal is not subtle. It is the most consequential variable we found — more than funding, team, product quality, or marketing spend.

Finding 002 — The Silent Majority

40–55% of businesses between $2M and $20M are dying right now. Quietly.

They don’t file bankruptcy. They don’t make headlines. They don’t lose 90% of their stock price. They just exist at $3M–$8M for a decade while the founder works harder every year for the same number.

“Most entrepreneurs develop thick skins, become aloof, depressed, or worse — they give up — having given up a lot to achieve so little.”

That’s not a statistic. That’s the 3AM reality for the founders in Pattern One. They tried a new website. Hired an agency. Fired the agency. Tried Facebook ads, then Google, then LinkedIn, then a rebrand. Nothing moved. Because every fix was aimed downstream. And the disease was upstream.

Eighty percent of business owners are trapped in what we call the evil trinity — debt, risk, and overwork — with negative cash flows, no predictability, and deteriorating health. And vague is where most of them live and die.

0%
Estimated true prevalence of Pattern One — “right product, wrong customer definition, wrong message” — in the $2M–$20M revenue range. Cross-referenced with BLS five-year failure data (49.4%), CB Insights post-mortem analysis (43% cite product-market fit), and Notion Capital stall-rate research (60% stall at $3M–$10M).
Bayesian prevalence estimation · sampling bias correction · BLS 2024 · CB Insights (n=431) · Notion Capital
Finding 003 — The Upstream Discovery

The thing that matters most is the thing that’s hardest to observe.

Our logistic regression revealed that product positioning (WHAT) is the strongest predictor you can observe from outside a company — coefficient 0.447. But the odds ratio told a different story: when customer definition (WHO) is identified as broken, the company is 104× more likely to be stalled.

Two different statistical lenses. Two different answers. Both true. Because they measure different things. WHAT is where the disease shows up. WHO is where the disease lives. And upstream of both — the thing no public data can capture — is WHY.

WHY absent
WHO undefined
WHAT commoditised
Plateau

The founder blames execution. Fixes downstream. Nothing changes. The dominoes never fall because the first one was never knocked over.

You cannot fix “nobody in particular” with a better website. You cannot fix a commodity product by spending more on ads. And you cannot define your customer if you haven’t articulated why your business exists beyond “I’m good at this.”

You need a diagnostic instrument to get upstream. That’s what we built.

Finding 004 — The Pattern Distribution

At least 245,760 ways to die. One way to thrive.

Every company in the study was classified into one of three failure patterns — or confirmed as aligned across all four diagnostic pillars.

40–55%
Pattern One
Right product. Wrong WHO.
27.7%
Pattern Two
WHO evolved. Company stayed.
4.4%
Pattern Three
Had everything. Never took off.
60.4%
Aligned
All four pillars intact.
Deeper Findings

What the data revealed

Pillar Correlation

WHO ↔ WHAT: rho = 0.895

When customer definition is broken, product positioning is almost always broken too. They are not independent failures. They are the same failure measured at two different layers.

Spearman rank correlation · very strong
Upstream Driver

WHY ↔ WHO: rho = 0.467

The founder’s reason for building the business moderately predicts customer definition quality. When WHY is absent, WHO defaults to “anyone.”

Spearman rank correlation · moderate
Model Validation

92.5% classification accuracy

Five-fold cross-validation. AUC 0.864. The model discriminates between patterns with high confidence despite class imbalance in the dataset.

5-fold CV · multinomial logistic regression · AUC 0.864
The Graveyard Finding

Every graveyard is filled with great products

In 25 years and 1,500 companies, we have never found an exception. The graveyard is not full of companies with bad products. It is full of companies with great products and no customer knowledge.

25 years · 300+ client engagements · 1,500-company study
The Invisible Disease

Pattern One companies don’t get written about

They plateau at $3M–$8M for a decade. The founder blames execution. Tries downstream fixes. Nothing moves. Nobody writes the story of a $4M company that’s been $4M for six years.

Sampling bias analysis · correction factor applied
The Transition Trap

“What made you successful at $1–3M will lead directly to your failure at $3–10M”

The founder’s hustle carries the business to $2M–$3M. Beyond that, the business needs a defined WHO, an articulated WHY, and a WHAT that speaks to identity — not features.

Notion Capital · 60% stall at $3M–$10M
The Instrument

The Business MOAT Grader

Four diagnostic pillars. Eight minutes. At least 245,760 diagnostic states. Statistically validated. Empirically weighted. Deceptively simple on the surface. Underneath, it navigates the full topology of how a $2M–$20M business stalls — and names yours.

WHY
Absent
WHO
Broken
WHAT
Partial
HOW
Strong
PATTERN ONE DETECTED

Strong operations aimed at undefined customer. Upstream intervention required.

“If your diagnosis doesn’t show you something about your business you couldn’t see before — I’ll find it with you. Personally.”

— Vivek Mehta, Principal Researcher
The Study Is Live — And Growing

1,500 companies diagnosed. 5,000 is the goal.

We’re expanding the 1,500-Company Diagnosis Study™ to 5,000 companies across 30 industries by July 20, 2026. Every new founder who joins sharpens the instrument. Every diagnosis makes the patterns more precise.

1,527 founders joinedGoal: 5,000
30% completeTarget: July 20, 2026
Early Access — Limited

Join the study and get early access to the Business MOAT Grader™ AI before it opens to the public.

Everyone who joins before we reach 5,000 gets their diagnosis free. After 5,000, the grader goes paid.

This Page Learns

When you scroll, pause, click, or return — we measure what resonates and what doesn’t. Not to track you. To sharpen the instrument. Every interaction teaches us which findings matter most to founders at your stage. That’s how the Business MOAT Grader™ gets smarter before you ever take it.

Your identity is never attached to this behavioural data. We see patterns, not people.

We don’t use this data for retargeting ads. We don’t sell it to advertisers. We don’t share it with platforms. It stays inside MOAT Labs and serves one purpose: making the diagnostic sharper.

Your vulnerability is a trust you place in us. We treat it as sacred.

Encrypted at rest
Anonymised in aggregate
Never sold to third parties
No retargeting ads
Deletable on request

Is your business vulnerable?

Join the study. Get the findings. See if yours is at risk.

Methodology

How the study was built

The 1,500-Company Diagnosis Study™ was designed to answer one question: why do businesses with good products, real skills, and genuine effort fail to grow past a revenue ceiling?

The Dataset

1,500 companies across 30 industries. Each scored on four proprietary diagnostic pillars (WHY, WHO, WHAT, HOW) using a documented codebook. Classified into three failure patterns or confirmed as positively aligned.

n = 1,50030 industries4 pillars3 patterns + positive

Statistical Analysis

Five-point analysis: pattern prevalence estimation with sampling bias correction, pillar correlation matrix (Spearman), multinomial logistic regression, Fisher’s exact test for odds ratios with 95% CI, and five-fold cross-validation.

Logistic regressionFisher’s exact testSpearman rho5-fold CVBayesian estimation

Source Registry

21 research sources including Harvard Business Review, CB Insights, Bain, McKinsey, BLS, and Notion Capital. Cross-referenced with 25 years of practitioner diagnostics across 300+ client engagements and six founded businesses.

HBRCB InsightsBLS 2024McKinseyBainNotion Capital
About MOAT Labs

The research behind the instrument

MOAT Labs is the research division of MOAT AI Technologies™, focused on one question: why do businesses with good products, real skills, and genuine effort fail to grow past a revenue ceiling?

Behind this research is 25 years of accumulated know-how — north of $60M — not just from textbooks or case studies, but directly from the People. From four countries, three continents, and five languages. Founders who built $100M companies and shared how over their own tables. Fortune 500 executives who sat as equals in academic settings with their armour off. Partners at global firms who showed how large systems are interrogated for truth. Public servants who revealed how institutions fail the people they were designed to serve. First-generation millionaires who let you watch them decide in real time. 300+ business owners who opened their books and unknowingly showed the same pattern, over and over, until it was undeniable. And the losses that proved the only asset that survives everything is what you know.

The value of this research is not just in the data. It’s in those rooms, with those humans, that taught one person to see what the data would eventually confirm.

Vivek Mehta
Principal Researcher, MOAT Labs
BEng, MEng, MBA (Rotman)

“Every business graveyard is filled with great companies who once had a great product.”
— From the 1,500-Company Diagnosis Study™
Collapsing Decades Into Days™