The Rabbit Hole, Part 2: How Brands Learned to Build Worlds
How community-driven brand strategy turned enthusiast obsession into the most profitable model in consumer goods, and the playbook behind world building.
There is a company that sells small plastic models of imaginary soldiers. You buy them unassembled, in a box, and then you glue them together and paint them yourself, badly at first. By its own reporting, Games Workshop earns a return on invested capital of roughly 69 percent, converts about 36 percent of revenue into free cash flow, and carries no debt at all.
Those are numbers you would expect from a software monopoly. They come from plastic.
Nobody arrived at those margins by making a better plastic soldier. Injection molding is not a secret, and the models are not technically hard to copy. What cannot be copied is everything that surrounds the model: the forty years of fiction it lives inside, the rules that make it matter on a table, the vocabulary its owner has learned, and the room full of other people who care.
The five properties are not an accident, they are a business model
Part 1 of this series, Nobody Falls in Love With the Object, described the five properties of a system that swallows people, and it described them from the inside, the way the person falling in feels them. That was the psychology. This is the balance sheet.
Here is the part most brand teams have not internalized. Those five properties are not weather. They do not just happen to certain hobbies because enthusiasts are strange. They are constructible. A company can sit down and build each one on purpose, in a sequence, and when it does, it gets the most durable economics available in consumer goods. That construction project is what community-driven brand strategy actually is, and it has almost nothing to do with running a nice Discord server.
The thesis of this piece is blunt.
“The most durable consumer businesses of the last thirty years did not win on product features. They won by building a world around the product. A customer who has learned your vocabulary, your history, and your status ladder is not a customer any more. They are a resident, and residents do not comparison shop.
”
Comparison shopping requires a comparison. A resident cannot make one, because the rival product does not offer the same thing. It offers an object. The world offers everything the object is attached to, and switching means leaving all of it: the fluency, the standing, the people, the years already invested. Price stops being the axis of the decision, and that is where pricing power comes from.
What a world looks like on a balance sheet
Games Workshop is the cleanest specimen in the wild, so it is worth taking apart properly rather than admiring from a distance.
annual growth over the past decade, per its own reporting
return on invested capital
free cash flow margin
estimated share of the global fantasy and sci-fi miniature market
Roughly 18 percent compounding annual growth over ten years. A return on invested capital near 69 percent. Free cash flow margins around 36 percent, no debt on the balance sheet, and an estimated 70 percent of the global fantasy and science-fiction miniature market. Its investor materials describe a distribution footprint of more than 540 company-owned Warhammer stores, a wholesale network of more than 7,000 independent retailers, and a direct e-commerce channel that keeps growing.
Read those figures cold and you would guess a payments network. This is one of the most profitable consumer businesses in the world, and its product is models you assemble at a kitchen table.
Why the numbers look like that
Three things stack, and the stacking is the whole story.
It owns the intellectual property outright. Not a license. Not a partnership. The setting, the factions, the history, the names, the visual language, all of it is owned. That means every book, every video game, every television or film adaptation, every new model line draws on an asset the company already holds and never has to rent. When you own the world, expanding the catalog is not a new product bet. It is a withdrawal from an account you have been paying into for four decades.
It owns manufacturing and the supply chain. Most consumer brands are marketing organizations bolted onto somebody else’s factory, which caps their margin and hands their unit economics to a supplier. Owning the molds and the plants means the cost curve belongs to the company, and so does the release schedule. Vertical integration is unglamorous and it is doing an enormous amount of the work in those cash flow margins.
The community does the marketing, the onboarding, and the retention for free. This is the part that looks like magic and is actually just structure. New players are recruited by existing players, taught to build and paint by existing players, and kept in the hobby by existing players who want opponents. The company does not have to buy that attention, because the acquisition is peer-to-peer and the training is volunteer-run. A hobby shop with a painting table is a customer acquisition channel that is also a customer support desk that is also a retention mechanism.
Then there is a structural detail that deserves more attention than it usually gets. The games are played in person, between people in a room, across a physical table. Not against a screen. That format is not nostalgia, it is the retention mechanism, and it is one no app can copy. A digital game keeps you by competing for your attention against every other digital thing. A tabletop game keeps you because a person you know is expecting you on Thursday with an army that has to be ready. Social obligation is stickier than any notification, and the loop closes neatly: the more enjoyable the games are, the more miniatures people want, because a better game creates demand for more of the thing the game is played with.
Licensing sits on top of all this and is the most capital-efficient move available. When the intellectual property is extended into books, video games, television, film, and licensed formats, partners fund the development. The company earns income for granting access to something it already owns and has already paid for. Every one of those extensions also does a second job for free: it deepens the canon, which makes the world larger, which makes residency more valuable, which sells more of the thing that started it. The licensing is not a side business. It is world maintenance that happens to be profitable.
The world-building playbook behind community-driven brand strategy
Strip Games Workshop of its specifics and the same six moves appear in every consumer business that has achieved this kind of durability. They can be run deliberately, and increasingly they are.
Give the product a system behind it. The object alone is a dead end, because an object can be finished. A system cannot. Lego did not become an adult business by making bricks nicer, it did it by treating building as an open-ended discipline with techniques, scales, and standards, then aiming sets at adults who wanted the harder version. Specialty coffee did the same thing to a drink: grind size, water temperature, extraction ratio, a new machine every time your palate outgrows the last one. The product becomes the entry point to a practice, and practices have no finish line.
Publish the canon. A world needs recorded history that is bigger than any one person can consume. This is the function of Warhammer’s decades of novels and codices, of Porsche’s motorsport record, of the story every mechanical watch brand tells about the decade a particular movement was designed. Canon does two things at once. It gives a newcomer something to study, which converts curiosity into hours, and it gives the veteran something to be authoritative about, which is the raw material of status.
Create a vocabulary. Every deep world runs on words outsiders do not have. Grail dial. Nine-bar extraction. Kilter. Tactile versus linear switches. Language is the cheapest and most powerful membership test ever invented, because learning it is voluntary, unpaid, and irreversible. Once you can speak it, you are inside, and the effort you spent learning it becomes a reason to stay.
Build the ladder. People need to see where they stand and what the next rung looks like. Peloton put the ladder on a public leaderboard and attached instructors to it. Rolex built one out of scarcity, where a waiting list turns availability itself into a rank. Sneaker culture uses drops, mechanical keyboards use group buys, and specialty coffee uses equipment that visibly announces how far along you are. The mechanism does not matter. Visible tiers of competence and status do, because they convert a purchase into progress.
Host the gathering place. Somewhere the residents can find each other, physically or otherwise. Those 540 stores are not primarily retail, they are venues, and a venue with a painting table and a schedule of game nights produces something no e-commerce funnel can. Apple learned the same lesson with its stores before it turned the ecosystem itself into the gathering place, where your devices, your files, and your messages all live in one country you would have to emigrate from.
Let the community own the culture. This is the move most companies flinch at, and the one that separates a real world from a marketing campaign wearing one. The company owns the intellectual property. The community owns the meaning: the house rules, the memes, the conversion projects, the arguments about which era was best. Patagonia handing customers a repair ethos, rather than a message about repair, is the same move in a different category. Control the canon, release the culture. A world nobody is allowed to modify is a museum, and nobody moves into a museum.
The unit economics of a world
The playbook is interesting. What it does to the arithmetic of a business is more interesting, because it bends all three numbers that decide whether a consumer company is good or extraordinary.
A normal brand pays to acquire each customer, sells them a product, and then pays again to get the next one. A world does not work that way. Members bring in members, which pushes acquisition cost down. The ladder never ends, which pushes lifetime value up. And leaving means giving up more than the product, which is where pricing power comes from. Three of the most stubborn levers in consumer economics move at once, and they move because of structure, not because of a campaign.
Layer 2 · Mechanismhow it actually works
Take them one at a time.
Acquisition cost falls because members recruit. In a world, the strongest sales channel is a person explaining their hobby to a friend with slightly embarrassing enthusiasm. That conversion rate is far above anything paid media produces, because it arrives with trust attached and a free onboarding service bundled in. The recruiter also teaches the newcomer, which removes the cost of getting someone from purchase to competence, the stage where most consumer products lose people permanently.
Lifetime value rises because the ladder never ends. A product has a natural end state. You own it, so you stop buying. A system has no such state. There is always a harder technique, a deeper part of the canon, a better piece of kit, a next tier. The customer is never finished, so the revenue is never finished, and the spend tends to increase with tenure rather than decay, which inverts the usual curve.
Pricing power appears because the alternative is exile. This is the one that surprises finance teams. In a commodity market, a price rise sends buyers to a substitute. Inside a world, the substitute does not exist, because the competitor sells the object and not the residency. Leaving costs you fluency, standing, the people, and every hour already invested. That switching cost is not contractual, it is social and psychological, which makes it far more durable than any lock-in a contract can write.
Layer 3 · Math & where it breaksgo deeper
The structural reason all three move together is that the community is doing work that would otherwise sit in the cost line.
Marketing, onboarding, and retention are three separate budgets in a normal consumer business, and they are three of the largest. In a world, all three are performed by members, for free, as a byproduct of enjoying themselves. The spend does not get optimized. It gets structurally removed, which is why a world-shaped business can carry retail rents, owned factories, and heavy content production and still show margins that look like software.
Where it breaks is worth naming. This model has a hard floor: the community must genuinely want to do that work. The moment recruiting feels like unpaid labour for a company rather than an invitation into something good, every one of those three cost lines comes back onto the income statement at once, and the fixed costs that the model was carrying comfortably become a serious problem. The economics are excellent and they are conditional.
You can stop after Layer 1 and still be correct about world economics, just less complete.
The dark side, and the tell
None of this is automatically benign, and pretending otherwise would make this a brochure. The same mechanisms that make a world worth living in can be turned on the residents, and four failure modes show up repeatedly.
Manufactured depth. Complexity added to sell more, rather than to reward mastery. Real depth is generative: learning it makes you better and opens possibilities. Manufactured depth is arbitrary: a rules revision that obsoletes what you own, a specification bump that changes nothing you can feel, a tier added so there is something above the tier you just reached. Residents tell the difference almost immediately, because real depth feels like a door and fake depth feels like a toll.
Engineered scarcity and status extraction. Scarcity that emerges from craft is honest, since some things genuinely take time to make well. Scarcity engineered purely to create a waiting list is a tax on wanting to belong. Once a company discovers it can charge for rank rather than for product, the temptation to keep raising that price is enormous, and the resident who cannot pay it discovers their standing was for sale the whole time.
The ladder becomes an obligation. A hobby is something you return to. An obligation is something you owe. When a world runs on streaks, seasons, and expiring status, the ladder stops rewarding progress and starts punishing absence. That flip is usually invisible in the metrics, because engagement looks identical whether it comes from delight or from anxiety, right up until people leave and do not come back.
The moment a community realizes it is being farmed. Every community can feel the direction of intent. Not the pricing, the intent. And when it turns, the reaction is not a slow decline in satisfaction scores, it is a sudden collective loss of goodwill, and goodwill was the asset doing all the work.
That lag is the trap. Extraction looks like success for a while. Raise prices inside a world and revenue goes up immediately, because the switching cost is real and people pay it once, then twice, and quietly decide they are done. By the time it appears in the numbers, the thing that produced the numbers has already left.
What to take from this if you are building something
The lesson is not that you should go and manufacture an obsession. It is that the choice is available to you and it is more concrete than it looks.
Ask what your product is the entry point to. If the honest answer is nothing, that is the actual problem, and no amount of community management will fix a product with no system behind it. If there is something, a practice, a craft, a body of knowledge, a way of doing things better, then you can build the rest deliberately: write the canon, name the concepts so people can speak about them precisely, make competence visible so progress is legible, give the residents a place to find each other, and then have the discipline to let them own what it means.
Then keep one instrument on the dashboard that finance will not give you: whether the world still exists for the residents or whether it has started existing for the quarter. The answer to that question decides whether you get thirty years of pricing power or three good ones followed by an exodus nobody predicted.
One more thing, and it sets up where this series goes next. Every example here was built when finding your world was hard. You had to stumble into a shop, know somebody, or read a magazine that happened to reach your town. Part 3, The Internet Poured Rocket Fuel on It, is about what happened when that friction vanished and every world on earth became reachable from a phone in about four seconds. The playbook did not change. The physics of who finds it, and how fast, changed completely.
