The Business Model Leverage Screen
Pick businesses with scale economies, zero marginal cost and network effects.
- Difficulty
- Advanced
- Time to result
- ~months to results
- Steps
- 5
- Confidence
- 89%
Some businesses have leverage baked into their economics before anyone works a single hour. Three microeconomic properties do the work. Scale economies mean production gets cheaper the more you make, which builds an automatic barrier against commoditisation. Zero marginal cost of reproduction means creating another copy is free, so the thousandth customer costs nothing to serve and revenue compounds even though early per-user revenue is tiny. Network effects are the subtlest and strongest: each additional user adds value to the existing base, and by Metcalfe's law value scales with the square of the node count, so a hundred-node network is a hundred times more valuable than a ten-node one rather than ten. Network effects produce natural monopolies, which is why the screen has a hard condition attached: only enter if you can be number one.
Origin
Extracted from Naval. Nivi asks whether a choice of business model can itself confer leverage, and Naval walks through scale economies, zero marginal cost of reproduction and network effects, citing Bob Metcalfe's law and using language as the oldest network effect.
Core principles
- 01Scale economies mean unit twelve costs less than unit five and unit ten thousand far less again.
- 02Zero marginal cost of reproduction means the next customer is free to serve.
- 03A network effect means each new user adds value to the existing user base.
- 04Metcalfe's law makes network value scale with the square of the nodes.
- 05Network effects create natural monopolies, so being number two is close to worthless.
- 06The three properties tend to arrive together and reinforce each other.
How to run it
- 1
Test for scale economies
Check whether producing more makes each unit cheaper. Rising volume that lowers unit cost builds an automatic barrier to entry and resists commoditisation.
Pro tip Compare widget five, widget twelve and widget ten thousand explicitly rather than assuming the curve bends.
- 2
Test for zero marginal cost of reproduction
Ask what it costs to serve one more customer. Technology and media products approach zero, which is what lets revenue compound over a large base.
Pro tip Anything with zero marginal cost automatically has scale economies, so this test partly subsumes the first.
Watch out These models earn very little per user and take a long time to get going. Judge them on trajectory, not early revenue.
- 3
Find the mechanism by which users add value to each other
State precisely how an additional user makes the product better for existing users. If you cannot articulate the mechanism, you do not have a network effect.
Pro tip Language is the cleanest example: each new English speaker makes English more valuable to every existing English speaker.
Watch out Weak network effects, like video-sharing, can still be decisive; do not dismiss them because they are not absolute.
- 4
Confirm you can be number one
Network effects create natural monopolies, so the runner-up captures a small fraction of the value regardless of product quality.
Pro tip Ride-sharing shows the mechanism plainly: the platform moving more drivers and riders will always have better unit economics.
Watch out A better product with fewer nodes loses. This is the one condition on the screen you cannot execute your way around.
- 5
Engineer the hooks deliberately
Design explicit features that let customers create value for each other, rather than hoping the effect emerges. That is the point at which your customers work while you sleep.
Pro tip Zero marginal cost products are the easiest place to add these hooks, since stamping out another copy costs you nothing.
In the wild
Naval's cleanest illustration. Take a hundred people speaking ten different languages, one each: everyone is translating constantly and it is painful. One extra person learns English, so now eleven speak it, and the next newcomer choosing a language is more likely to pick English for that reason. By the time twenty or twenty-five speak it, the outcome is settled and the other languages get competed out. The same mechanism explains why technical education worldwide runs through English: the books simply have not been translated.
→ A single language captures the entire market not by being better but by crossing a node threshold.
Naval uses podcasting to show zero marginal cost compounding. Joe Rogan is working no harder on his current episode than he was on episode one. Episode one probably lost money. Episode eleven hundred earns something on the order of a million dollars. Nothing about the production effort changed; what changed is that the audience accumulated against a product whose cost of serving one more listener is zero.
→ Identical effort per unit produces a revenue curve that bends sharply upward over time.
Common mistakes
Entering a network-effects market as number two
Natural monopolies mean the second-place network is structurally disadvantaged forever. Naval notes he likes DuckDuckGo for privacy reasons and that it will always be behind for exactly this reason.
Quitting a zero marginal cost business early
These models make very little per user at first and only really add up over time. Judging them on early per-unit revenue kills them before the compounding starts.
Assuming users will connect themselves
Network effects have to be designed as concrete hooks that make each new user valuable to the existing base. Without a stated mechanism it is just a large customer list.
Is it for you?
Best for
Founders and investors choosing which business to start or back, where the model itself can still be designed rather than inherited.
Not ideal for
Local or service businesses where physical delivery genuinely caps replication and users have no way to add value to each other.
From the transcript
“A network effect is when each additional user adds value to the existing user base.”
“the value of a network is proportional to the square of the number of nodes in the network”
“When you're picking a business model, it's really good idea to pick a model where you can benefit from network effects, low marginal costs, and…”
From the episode
How to Get Rich: Every Episode