Service vs Scalable

This is the reading version of the Service vs Scalable module. The module is fast and interactive. This is the slow, deep version, for when you want to sit with the idea.

Two doors, and one of them eats first-timers

You have a skill and roughly no capital. In front of you are two doors. Door A: build a product. Bottle it, app it, manufacture it, dream about a million users. Door B: sell a service. Do the thing for one person, get paid, do it again.

Both doors lead to wealth, but only one quietly bankrupts most first-generation founders before they ever find out whether they were any good, and it is usually the one your instinct points at. The glamorous door and the smart door are not the same door.

Here is the rule I give every first-gen founder, and this essay is the full argument behind it: always start with services. Not because products are bad. Because products are a second move, and almost everyone tries to make them a first move.

The J-curve, or who pays for your education

A product business has a specific financial shape. You sink money in first: inventory, tooling, months of building before anyone has heard of you. The line dips below zero and stays there. Only later, if you survive, does it curve up. That shape is the J-curve, and it carries an unspoken entry fee: capital to live through the dip, the temperament of a seasoned operator, and a backup so failure does not end you.

Read that list again. A first-time founder typically has none of the three. J-curve businesses are won by people who can afford to lose, and that is not you yet. Not an insult, a starting position. Starting positions change.

A service business is the opposite shape. You typically get paid in advance, so the client covers your costs before you lift a finger. There is no dip to survive because there is no upfront bet. The market funds your education, and if you are bad at this, you find out cheaply and quickly, which is exactly what you want when you are starting blind.

One line carries the whole module: products make you pay to find out. Services pay you to find out.

A second-order gift: a product founder gets one big verdict at launch, while a service founder gets marked by reality every week, and those honest grades turn a guesser into an operator.

Now the objections, because they deserve real answers. "A digital product costs almost nothing to make." True in rupees, false in what matters: even a cheap-to-build product eats months of your work before a single rupee arrives, and months of runway are precisely the capital you cannot spare. "Speed to market matters more than the model." Speed matters, but the model decides who carries the risk; speed only decides when you hit the wall. "A product is an asset you own; a service dies when you stop." This one hides the most. A well-run first service builds assets that never appear on a balance sheet: a network of wealthy clients, a reputation inside a niche, hard data on what the market actually pays for. Exactly the ingredients the eventual product bet requires, and none of them can be bought.

The shape: six rich clients, not fifty cheap ones

Once you pick services, you pick a shape, and here the instinct fails again. Same yearly income target, two plans: fifty clients paying small, or five to six clients paying a lot. Almost everyone's gut says fifty, because fifty feels safe.

The gut is wrong, and the value equation explains why. Your offer is the dream outcome multiplied by the perceived likelihood of achieving it, divided by the time, money, and effort it costs. In a service business, your own time is the denominator that breaks first. Fifty humans with demands, revisions, and moods will shred your calendar, and you will spend all of it on management instead of delivery.

Five to six high-ticket clients keeps the denominator low while the numerator stays fat. And there is a blunt capacity truth underneath the math: a first-year founder can genuinely serve about five to six clients well, so build the plan around that instead of fighting it. The target shape: your first twelve transactions guaranteed across the year, each paying you more than a month of runway, from five to six clients you can count on. Serve people wealthy enough that a handful of them nets you real money, so a sale feels like a win and not a headache.

What about concentration risk? Losing one of six clients hurts, true. But the fifty-client plan does not remove that risk, it converts it into time bankruptcy, and time is the one resource you cannot invoice back. The high-ticket plan handles it differently: each transaction pays more than a month of runway, so a departure buys you time, and a sharp niche means the replacement is another buyer of the exact same shape.

And the blend, a base of small clients for safety plus one or two anchor accounts for the money? Sounds balanced, lives like a trap. The small base still eats the hours while the anchors demand your best work, so you end up doing both jobs badly. Pick the shape. Do not hedge it.

The numbers that separate a business from a busy job

A service business without margin discipline is not a business, it is an expensive hobby that keeps you tired. Here are the floors. Charge a 2.2x markup, minimum, or there is nothing left for you after costs. Aim for gross margin above 50 percent, and 60 is lovely. Aim for net profit above 40 percent, and 50 is lovely.

Why so rigid? Because margin is what freedom is made of. It lets you hire help without panic, refuse a bad client without fear, and build the thing so it runs without you lifting a finger, which is the entire point of going high-ticket and low-volume. It is also the war chest for your second move: when you eventually play the J-curve, you play it with saved capital and a completed education. Under those floors, you have simply bought yourself a job with extra steps.

Niche down: my 150-meeting lesson

This part I learned with my own calendar.

When I had not narrowed down, I pitched everyone: students, startups, shops, anyone with a pulse. I did over 150 meetings and barely closed one or two clients, and none of them stayed. Every meeting started from zero trust, because I was nobody's specialist. Then I focused on a specific kind of surgeon: low competition, every market criterion met, and an edge I will not disclose, because the edge is the edge. Ten meetings got me five clients. Same me, same skill. The only variable was being narrow enough to be the undeniable option.

The physics of the room changes. A generalist walks in as one of a hundred interchangeable vendors and builds credibility from scratch. A specialist walks in as the only person obsessed with this exact buyer's exact problem, and the meeting starts from trust instead of suspicion. Generalists pitch. Specialists get picked.

The standard objection: surely a tiny niche runs dry before the specialist edge pays off. It almost never does. A good niche feels tiny from the outside and is deep on the inside; surgeons alone could fill my book many times over. And six clients is a full year, so your capacity runs out long before the market does.

Less is more, and get exploited on purpose

Under pressure, your instinct is to add: more services, more verticals, more types of client. It feels like safety and it is the opposite. The move is to do less, for fewer people, better than anyone. You do not need a mind-blowing idea. Take a normal, existing service and point it at a sharply segmented market. Picture the only Jamaican restaurant in all of Bombay: low competition, first-mover edge, and once you have served real customers, you will know how to Indianize the menu better than anyone who never took the leap. Narrow is not the limitation. Narrow is the advantage.

But how do you find the niche? Not between your ears. Your first idea probably sucks; real businesses are forged in the interplay between a founder's vision and their first messy customers. So go get exploited early, on cheap or thankless projects, but choose who exploits you. Be exploited by a famous builder and you walk away with their network, a lucrative market, and insider insight nobody else has. In Naval Ravikant's framing, that is leverage. In your first hundred projects you are earning data more than money: which work you enjoy, which clients pay best, who is actually good to work with. The balance point of those three is your niche. You do not get to skip the exploitation phase. You only get to choose who does it and what you steal on the way out.

The ceiling: why services do not scale, and why that is fine

Your six-client machine is humming, and a friend says the obvious thing: just add forty more clients, that is how you get rich. Here is why that advice destroys good service businesses.

People are complex. That is the whole reason services do not scale, and the whole reason your first business should lean into that limit instead of fighting it. A service is built out of humans: you, your small team, your clients. Every additional client is not a nearly-free marginal unit like an app download; it is a new human relationship, and relationships get more expensive to manage as you add them, not less. Economies of scale belong to products. In services, the complexity caps you around ten clients before quality and margin start bleeding.

"Fine," says the friend, "so hire ahead of demand and systematize delivery." Mohnish Pabrai tried the grown-up version of exactly that. He built an IT services firm to 170 people and millions in revenue, then looked up and realized what his job had become: a full-time source of motivation and HR for 170 humans. What lit him up was strategy. So he sold the firm and moved to value investing, where 95 percent of the work was the strategic thinking he loved, and went on to build a billion-dollar portfolio. Hiring did not remove the complexity of his service business. It converted client complexity into staff complexity, because people are the complexity.

The lesson is not that scaling is bad. It is that scalable is a different game, played by a different kind of operator, and it is the ceiling you graduate into, not the floor you start on. Pabrai earned his way up to it. So will you.

And if you do eventually walk through the product door, smuggle the service mindset in with you. Real estate is technically a product, but its operations are basically sales, so treat it that way. Prefer bulk over single units: a chocolatier should chase corporate gifting, not one-bar sales, because one order means many units and real money per deal. Start as a trader before you touch manufacturing, because most manufacturers do not retail and most retailers do not manufacture; they are different skill sets. And since product overheads are brutal, target 90 percent gross margins. Take all of this lightly though: products are not my strongest area, and I would rather say so than pretend.

How to actually apply this

Design your first service business on one page, four boxes.

1. The slice. Pick a normal skill you already have, then point it at one weirdly specific, wealthy segment. The narrower and stranger, the better. Think "a certain kind of surgeon," think "the only Jamaican restaurant in Bombay."

2. The price. Set it so that five to six of these clients hit your yearly income target, then sanity-check the unit math: 2.2x markup, 50 percent-plus gross margin, 40 percent-plus net. If the numbers only work at fifteen clients, the price is wrong, not the model.

3. The channel. Name the one place these specific people actually are. A network, a referral chain, a single platform. One, not all of them. Six rich clients need one warm room, not mass marketing.

4. The leverage line. Write down who you would happily let exploit you for six months, and exactly what you would walk away owning: the network, the market access, the insider edge.

If you cannot fill all four boxes, good: you found the part of your plan that is still fantasy, on paper instead of with your savings. Six rich clients, one narrow niche, one channel, one source of leverage. If it does not fit on a napkin, it is not ready. And if it does, you are holding a business where the market pays for your education, your time stays yours, and the scalable dream waits for you at the ceiling, where it belongs.

Want the fast, interactive version instead? Run the Service vs Scalable module, or explore the whole codex.