How to Build a Moat in Business: The MOATS Framework

The short answer

To build a moat in business, create advantages competitors can't easily copy. The MOATS framework groups them into five factors: Marketing supremacy (cheap acquisition and brand), One (no single point of failure), Audience (low customer concentration, network effects, ongoing usage), Toughness (real competitive advantage, high gross margin, scale) and Switching (high switching costs and embedding). You don't need all five, but the more you have, the wider the moat.

Key takeaways

A moat is what stops a competitor copying your subscription overnight. Subscription plus moat is the ultimate business.
MOATS stands for Marketing supremacy, One, Audience, Toughness and Switching, covering 11 moat factors in all.
No client should be more than 10% of revenue, and your top five no more than 25%.
Low churn during a boom is not a moat. Peloton's monthly churn rose from 0.73% to 2.2%.
AI erodes moats built on content and seats, and strengthens moats built on proprietary data, embedded workflows and trusted relationships.

David Velez was sent to Brazil by his employer, the venture firm Sequoia, to look for deals. Then Sequoia pulled out of the country.

But Velez had noticed something. Brazil had lots of companies that looked exactly like one another, especially banks, and very few with anything different to offer.

In other words, there were many companies but very few moats.

His own experience as a bank customer in São Paulo sealed it. Opening an account felt like the bank was doing him a favor. Security patted him down when his laptop set off the metal detector. And the numbers were insane: a 2018 Harvard Business School study put the average credit card interest rate in Brazil at 272 percent.

There was a problem. In 2013, foreigners couldn't start a bank in Brazil, and Velez didn't even speak Portuguese.

So he created a credit card.

It was purple, the shade you'd see on Barney & Friends. People told him he was mad, that customers wanted a conservative bank color. It had no annual fee, credit limits from about $50, a free app to manage everything, and its own credit score for people the big banks ignored. Twenty percent of Nubank's first customers had never had a credit card.

When I wrote the book, Nubank had been valued at $30 billion and had a $750 million investment from Warren Buffett's Berkshire Hathaway, all in eight years.

It has kept going. In the second quarter of 2026, Nubank reached 139 million customers across Brazil, Mexico and Colombia and passed $1 billion of quarterly net income for the first time. It now has a full banking license in Mexico and a premium offer, Croma, for its best customers.

The credit card is a recurring relationship: customers repay what they borrow, with interest, month after month. The moat is everything that made it different. Once your salary lands in Nu, switching is a hassle. The Nubank case study tells the whole story.

That's what this chapter is about: building a subscription business that competitors can't simply copy.

What is a moat in business?

Picture a castle with a deep, wide body of water around it. That's a moat. Its job is to stop attackers getting in.

In business, a moat is any advantage that stops competitors taking your customers or your margins. In 1995 Warren Buffett told Berkshire Hathaway shareholders: "We're trying to find a business with a wide and long-lasting moat around it."

Subscription revenue alone isn't a moat. If you sell dog biscuits on subscription, I can find a supplier, put up a template website overnight, and we're in competition. That's why the book's "ultimate business" is subscription revenue plus a moat. If you're starting from scratch, read why recurring revenue is the ultimate business first. Short definitions are in our economic moat glossary entry.

MOATS: the five key factors for protecting your subscription castle

The five strategic factors conveniently form the acronym MOATS:

Marketing supremacy
One (no single point of failure)
Audience
Toughness
Switching

You don't need all five to run a successful subscription business. You need some. The more you have, the wider your moat.

Across the five factors there are 11 moat elements. Here they are on one page, with the test I'd use for each.

Factor	Moat element	The quick test
M Marketing supremacy	1. Low-cost acquisition	Do most new customers come from referrals, organic search, community or partners rather than paid ads?
	2. Strong brand	Do customers ask for you by name without comparison shopping?
O One	3. No single point of failure	Would losing one supplier, channel, staff member or investor seriously hurt you?
A Audience	4. Low customer concentration	Is your largest client under 10% of revenue, and your top five under 25%?
	5. Network effects and shareability	Does your product get more valuable as more people use it?
	6. Ongoing usage	Do customers use or buy from you every week or month?
T Toughness	7. Sustainable competitive advantage	Can you name one thing you do that nobody in your niche does as well?
	8. High gross margin	Is your gross margin high for your industry, and stable?
	9. Marginal profitability and scale	Do costs rise more slowly than revenue as you grow?
S Switching	10. High switching costs	Would leaving you cost a customer real time, money or data?
	11. Embedding	Are you integrated into the customer's systems or daily routine?

Score yourself as you read. The free scorecard below does it properly.

MOATS ScorecardScore your business on the five moat factors.Open the free tool →

Let's go through each factor.

M – Marketing supremacy

Marketing supremacy has two parts: acquiring clients without a big marketing spend, and branding.

1. Acquiring clients without a high marketing spend

Manuel Suarez, founder of Attention Grabbing Media, built Facebook Messenger chatbots for online personalities. When I wrote the book, his business generated about $200,000 a month in recurring revenue while spending barely anything to acquire clients.

How? He used Messenger to turn people who engaged with his content into conversations, asking simple questions: What's your current revenue? Do you have a YouTube channel? How many staff do you have? That told him exactly who to talk to and how to help them.

The tactic will change. Messenger bots were the hot thing then; in 2026 it might be LinkedIn, a newsletter, a community or being recommended by AI assistants. When TechCrunch reported Canva's $4 billion revenue milestone in February 2026, the headline was rising referral traffic from AI chatbots. The principle doesn't change: a business that doesn't need to spend millions a year on marketing is far more attractive to a buyer or investor.

Referral programs, investing in existing customers and knowing exactly who your target customer is all lower your spend. The book's Chapter 8 covers growing on a shoestring, summarized in growing subscribers on a shoestring budget, and my first book, Feed a Starving Crowd, has 227 free or cheap marketing strategies.

2. Branding

What do Booking.com, Google and Apple have in common? Their brands are the first port of call.

We don't "search for it." We "Google it."

Most people who buy an iPhone don't comparison shop. If they did, they'd find phones that do similar things for a few hundred dollars. They buy Apple for the status and the ecosystem.

But here's the honest part. Brand is one of the weaker moat elements on its own. Kodak, Yahoo, Blockbuster and Borders all had powerful brands and still lost to stronger rivals. Branding helps when it's combined with the other factors.

O – One: no single point of failure

There's only one element in this factor. But I tell you, this is a big one.

Listen to enough startup coaches and you'll hear the rule of one: one product, one target customer, one marketing channel, one sales funnel. It makes life simple early on.

Eventually, the rule of one falls apart. What happens when your one thing stops working?

When the one thing stops working

Matt Mickiewicz, co-founder of Flippa, 99designs, Hired and SitePoint, found out in 2000. SitePoint made its money from online ads and sponsorships. Then the ad market dried up, seemingly overnight. SitePoint survived only by pivoting to selling memberships.

James Altucher lost $9 million in a day when his company's largest shareholder turned out to owe the IRS $90 million in back taxes. He lost his one major investor and had another failed company.

I once had lunch with a well-known venture capitalist who said he wouldn't invest in companies that relied too heavily on Facebook or Google advertising. He'd seen too many fall over at the mercy of those platforms.

A 2026 example: when the channel and the product went at once

Chegg built one of the biggest homework-help subscriptions, with most students arriving through Google search. Then generative AI gave away instant answers for free, and Google started answering questions directly with AI Overviews. Chegg's revenue in the second quarter of 2026 was $51.8 million, down 51 percent. One channel and one product, both hit at the same time. Read the Chegg case study.

The supplier who doubled the price overnight

I once had a key supplier double its prices on me overnight, which destroyed our margins. It took six months to find a new supplier at our old rate. If I'd had a second supplier lined up, I could have saved six months of pain. The only thing you need to know about the rule of one is that it's the one rule you should break.

Action exercise: list your top five key dependencies across sales, marketing, product, operations, customer success, engineering, suppliers, finance and investors. One quota-beating salesperson? One strong engineer? One supplier for a critical part? Write a mitigation next to each.

A – Audience

This factor has three elements: low customer concentration, network effects and shareability, and ongoing usage.

1. Low customer concentration

Imagine you make engines for commercial airliners. You can really only sell to two companies, Airbus and Boeing. That's high customer concentration.

Or you land a big deal with Walmart. You take a thin margin for huge volume, and one day Walmart deletes your product line, leaving you with a warehouse of stock you can't sell.

Don't get me wrong. I love enterprise sales. Just be cautious about the downside.

A 2015 study by researchers from Cornell and Drexel universities found that high customer concentration exposes a business to more risk. Netflix doesn't panic when one subscriber in Tampa cancels. That's the point.

My rule of thumb: no client should be more than 10 percent of your revenue, and your top five no more than 25 percent.

Babylon Health, which the book used as an insurance-model example, shows what happens when you ignore it. Contracts worth nearly half of its 2022 revenue weren't renewed, and its US business filed for liquidation in 2023.

Worked example: the 10/25 concentration test (illustrative)

An IT services firm earns $1.2 million a year from 40 clients. Its five largest pay $180,000, $110,000, $90,000, $60,000 and $50,000.

Largest client: $180,000 ÷ $1,200,000 = 15%. Over the 10% line.
Top five: $490,000 ÷ $1,200,000 = 41%. Well over the 25% line.

To get the top five under 25% at today's sizes, the firm needs total revenue of about $1.96 million ($490,000 ÷ 0.25). That's roughly $760,000 of new revenue from other clients. The fix isn't to fire big clients. It's to make winning more mid-sized clients the sales priority, and to put the largest on a longer contract. (Figures are illustrative.)

2. Network effects and shareability

In 2007 a YouTube user named Tay Zonday uploaded "Chocolate Rain." For three months, almost nobody watched. A handful of users made parodies, YouTube featured the parodies and the original on its homepage, and the video went on to more than 130 million views.

That's network effects: a product becomes more valuable as more people use it.

Udemy instructors live on it. More students mean more reviews, which attract more students. My friend Shani Raja, a former Wall Street Journal editor, had more than 460,000 students on his writing courses when I wrote the book. (Who knew so many people wanted to learn to write?) Udemy itself has since moved hard into subscriptions and, in May 2026, completed a merger with Coursera, which says the combined platform has 290 million learners.

Of all the moat elements, I rank network effects first. They get their own chapter, and their own page: the six types of network effects.

3. Ongoing usage

You don't want to sell something once and never be needed again. A doctor who cures an ear infection has no ongoing usage. The patient moves on.

Netflix is the opposite. Its huge library and relentless recommendations keep you watching. In the first quarter of 2021, its monthly churn was about 2.4 percent, against an average of 7 percent for other premium streaming services.

Netflix has since stopped reporting subscriber numbers altogether (from 2025), and points investors to revenue and engagement instead. That tells you what it thinks matters: usage drives retention, and retention drives revenue.

In software, measure daily and monthly active users. In e-commerce and physical products, measure purchase frequency. Action exercise: brainstorm three ways to get customers coming back more often.

How wide is your moat?

Score your business on the five MOATS factors with the free MOATS Scorecard and see where to dig first.

Open the free tool

T – Toughness

Toughness has three elements: a sustainable competitive advantage, high gross margins, and marginal profitability and scale.

1. Sustainable competitive advantage

Of all the moat elements, this one is second only to network effects. No competitive advantage, no sustainable business.

Why Peloton's moat turned out shallower than it looked

In the book I used Peloton as the example. At first glance it was just an exercise bike with an app. Yet it had more than 7 million users, more than $4 billion of annual revenue and a minuscule 0.73 percent monthly churn, thanks to its community, a high-quality bike and the sunk cost of spending $2,000 on equipment.

Here's what happened next. By fiscal 2026 (to 30 June 2026):

revenue was $2.446 billion, roughly 40 percent below the figure in the book;
paid connected fitness subscriptions fell 8.8 percent to 2.553 million;
average net monthly churn rose to 2.2 percent in the fourth quarter, about three times the book's figure;
on the bright side, Peloton posted its first full profitable year, after deep cost cuts and a price rise to $49.99 a month.

The lesson I'd add now: low churn measured during a boom is not a moat. Sunk-cost loyalty wears off as the equipment ages and the novelty fades. Community helped. It wasn't enough on its own once demand fell.

A competitive advantage is simply something you can do that nobody else in your niche can do as well. Coca-Cola and Pepsi have recipes nobody can copy exactly. McDonald's and Walmart win on cost leadership, though I'm not keen on that for most businesses: it needs huge capital, and a better-funded rival can undercut you.

One more thing. Don't obsess over competitors. Understand what they're doing, check in occasionally, then ignore them and run your own race.

Action exercise: draw a features matrix of you against your competitors. You don't have to win on every feature, only the ones customers value.

	Southwest	Other airlines
Lowest fares	✅	❌
Friendliest service	✅	❌
Seat selection	❌	✅

Southwest was the cheapest airline with the best service, and it didn't offer seat allocation. Mailchimp was the first email platform with a free plan and chose ease of use over endless customization. Pick your battles.

2. High gross margin

Strong financials make a huge difference to your moat. A high gross margin gets the attention of buyers and investors, and generally supports a higher price-to-revenue multiple.

Gross profit margin = (Total revenue − Cost of goods sold) ÷ Total revenue × 100

The book's example was WiseTech, the Australian company whose CargoWise software runs global logistics firms. CargoWise had a 92 percent profit margin and about 1 percent annual churn for nine years. If WiseTech started the year with 100 clients and made no sales at all, it would still have 99 at the end.

Action exercise: calculate your gross margin and look for anything you could take out. Close.com, a CRM for startups, originally included unlimited free phone calls in its plans. It later charged per call on top of the monthly plan.

Our guide to gross margin for subscription businesses has benchmarks.

AI features can quietly destroy your gross margin

Traditional software runs at roughly 75 to 80 percent gross margin. SaaStr puts AI-native companies at 50 to 60 percent, because every AI request costs money to serve. A gross margin of minus 23 percent was reported for the AI coding company Cursor in the quarter to January 2026. If you add AI to your product, price it with included allowances, credits or fair-use caps before you scale it.

3. Marginal profitability and scale

Marginal profitability shows how well a business scales. Compare two periods: the change in revenue against the change in costs. If revenue is growing faster than costs, you're scaling.

Microsoft spends heavily to build Office, but each extra cloud subscription costs it next to nothing to deliver. Almost every extra dollar is gross profit.

Coaches can do the same. One-to-one coaching caps you at 30 to 40 clients if you want to work business hours. Pre-recorded lessons plus a weekly group call let you serve far more people without your costs rising.

Investors notice when this goes wrong. When Google posted negative marginal profitability in its first quarter of 2011, its share price fell $48 the next day.

Get it right and scale feeds itself. More users bring stronger network effects and better supplier prices, which let you lower prices, convert more customers and grow further. Amazon is the textbook example.

Action exercise: work out how your costs scale as sales grow, and list ways to serve customers more efficiently as you get bigger.

S – Switching

How easy is it for customers to leave you? For a strong moat, it should be as hard as possible. Not through traps. Through value.

The element here is high switching costs, also called embedding: integrating so deeply into the customer's systems or life that they can't replace you on a whim.

How TechnologyOne made itself too embedded to leave

TechnologyOne sells enterprise software to government, health and education. In 2020 its churn was under 1 percent. Its software is so embedded that customers don't make snap decisions to leave. They talk to TechnologyOne first, which gives it the chance to fix the problem.

It has kept compounding. In fiscal 2025, TechnologyOne's annual recurring revenue reached A$554.6 million, up 18 percent, with net revenue retention of 115 percent. It passed A$500 million of ARR 18 months ahead of plan and set a target of more than A$1 billion by fiscal 2030.

Other kinds of lock-in:

Consumable lock-in: Gillette razors only take Gillette heads.
The learning trap: moving off Salesforce means migrating data and retraining everyone.
Contract penalties: Verizon and AT&T use them, and customers hate them. Use this one cautiously.

That last warning is louder in 2026. In March 2026 the US Department of Justice announced that Adobe had agreed to a $150 million settlement over allegations it hid an early termination fee and made cancelling hard. From 1 July 2027, Australia's new unfair trading practices law is expected to target unreasonable barriers to cancelling (general information as at October 2026). A moat built on friction is now a legal risk. Build switching costs from data, integration and value instead.

Action exercise: list ways to raise switching costs by becoming more valuable: solve more of the customer's problems, integrate with their systems, or offer a minimum contract term with clear, fair terms.

Tell-tale signs of a company with a moat

For a public company, check five to ten years of financials. If it has consistent earnings, share price growth, sales growth and a steadily high gross margin, there's a moat at work. Competitors haven't been able to stop it.

How Stripe started from scratch and built a powerful moat

Brothers Patrick and John Collison founded Stripe in 2010 to make accepting payments online easier. Their answer was a few lines of code a developer could paste into a website.

Since then Stripe has added capital, corporate cards, custom reporting and Radar fraud tools, all with higher margins than payments.

When I wrote the book, I walked through Stripe's moat factor by factor:

Competitive advantage: a superior product for smaller users who value ease over price.
Low customer concentration: big clients such as Shopify, Salesforce, Google, Amazon and Zoom, but millions of small ones too.
Switching costs: larger customers get lower rates, often in exchange for a contract, so moving would cost them more.
Gross margin: Stripe doesn't publish it, but it claims industry-leading margins. Its rival Adyen's were around 55 percent in a 2021 CB Insights report, so I estimated Stripe's were above that.

And its weaknesses: switching processors isn't that hard for a savvy small company, and at the time Stripe let businesses in only about 40 countries accept payments.

What has happened to Stripe since

The moat got wider. Businesses on Stripe processed about $1.0 trillion in 2023, $1.4 trillion in 2024 and $1.9 trillion in 2025, up 34 percent and about 1.6 percent of global GDP. A tender offer in February 2026 valued it at $159 billion. It keeps adding products on the same rails (billing, tax and, in January 2026, usage-billing company Metronome), and every product it adds makes leaving harder. See how Stripe's moat got wider.

Stripe doesn't have every moat factor. Neither does your business need them all. But having several gives you a solid foundation to build on.

2026 update: moats in the AI era

Since the book came out, AI has become the biggest test of moats in decades. Some moats are draining fast. Others are getting deeper.

Moat type	What AI is doing to it	2026 evidence
Content libraries and answers	Eroding. Free AI can produce similar content instantly	Chegg's revenue fell 51% in Q2 2026
Per-seat software	Eroding. Customers need fewer people, so fewer seats	WiseTech removed seat fees in December 2025
Commodity utilities	Eroding. Big platforms bundle them for free	Dropbox's paying users have been flat at about 18 million for two years
Proprietary data and rules	Strengthening. AI needs trusted inputs	WiseTech says its embedded regulatory rules "cannot be replicated by prompting an AI model"
Embedded workflows	Strengthening. AI agents need a system of record to act on	Google now markets Looker's semantic layer as the foundation for AI agents
Primary relationships	Strengthening. Trust is scarce	Nubank's primary-bank relationships and 139 million customers
Community and network effects	Holding. People still want each other	Strava passed 200 million users; Duolingo's streak revival brought back 15.4 million learners

Three practical questions for your own business:

If AI did your core deliverable free tomorrow, what would customers still pay you for? Build more of that now.
What do you know that a model can't learn from the internet? Client history, local rules, proprietary data, your methodology. Put it at the center of your offer.
Are you priced on people or on work? If your clients get more productive with AI, make sure your revenue grows with the work, not shrinks with their headcount.

The full analysis is in moats in the AI era, and you'll find 30 more real-world examples in competitive moat examples.

Does your business fit the MOATS framework?

Nubank, TechnologyOne and Stripe all have several moat factors, not all of them. Your business doesn't need every factor either. But you do need some, and you should know which ones you're missing.

Action exercise: score yourself out of five on each MOATS factor.

Factor	Your score (0–5)	Weakest element	One action this quarter
Marketing supremacy
One
Audience
Toughness
Switching

For any factor below three, pick one element from the table near the top of this page and work on it this quarter. Start with "One," because removing a single point of failure lowers your risk immediately. The MOATS framework guide has a deeper version of each test.

What to do next

Run the MOATS scorecard and note your two weakest factors.
Do the concentration test on your top five clients.
List your single points of failure and line up a backup for the worst one.
Calculate your gross margin, and if you sell AI features, check your margin on your heaviest users.
Pick one switching-cost project that adds value: an integration, a data history, a deeper service.

A moat is what turns recurring revenue into a business someone will pay a premium to own. It's what my first business taught me, and it's what the rest of the book helps you build.

Read the whole book free

The Subscription Playbook as a free ebook and audiobook. Australian business owners can claim a free hardcover, shipped free.

Claim the free hardcover Get the ebook

Frequently asked questions

What is an economic moat?

An economic moat is a durable advantage that protects a business's profits from competitors, the way a moat protects a castle. Warren Buffett popularized the term in 1995 when he said Berkshire Hathaway looks for businesses with 'a wide and long-lasting moat.' Common moats include network effects, switching costs, cost advantages, brands and proprietary data.

Can a small business have a moat?

Yes. Small businesses rarely win on scale, but they can build switching costs (being embedded in a client's systems), a loyal community, a niche brand, proprietary know-how, exclusive supplier relationships and low customer concentration. A local accounting firm whose clients' books, payroll and tax history all run through it has a real switching-cost moat.

How do investors tell if a company has a moat?

They look at five to ten years of financials. Consistent earnings, sales growth and steady high gross margins suggest competitors haven't been able to dent the business. They also check churn, net revenue retention, customer concentration and how much the company spends to acquire each customer.

Which moat is the strongest?

In the book I rank network effects first and a sustainable competitive advantage second. Switching costs are close behind for business software, because embedded systems are painful to rip out. Brand is one of the weaker moats on its own: Kodak, Yahoo, Blockbuster and Borders all had famous brands and still lost.

Do moats increase the value of a business?

Generally, yes. Buyers pay more for revenue they believe will last, so lower churn, higher gross margins, diversified customers and switching costs tend to support higher valuation multiples. A business that depends on one client, one channel or the founder usually sells at a discount, if it sells at all.

How long does it take to build a moat?

Some elements are quick: adding a second supplier, diversifying traffic sources or adding a minimum contract term can happen in months. Network effects, brand and deep embedding usually take years. Start with the 'One' factor, removing single points of failure, because it reduces risk immediately.

Get both books free

The Subscription Playbook (ebook + audiobook) and Feed a Starving Crowd, instantly. In Australia? Claim the free hardcover.

Robert Coorey, MBA

Robert is the author of The Subscription Playbook and Feed a Starving Crowd and a Director at BPAG Accountants in Sydney. More about Robert.

Sources

Facts and prices checked 27 October 2026. Spotted something out of date? Tell us.