Paul Graham says that when advising startups, asking "What would make this company more powerful?" can drive far bigger change than simply asking "How can we make more money?" Customer relationships, money flows, network effects, long-term strategy, and ways of capturing data and customers early can grow a startup's power not by a few times but by orders of magnitude. Every strategy, however, can only succeed if it actually makes customers' lives better.

1. Asking About "Power" Rather Than Revenue

The author offers the following as the question he finds most useful in startup office hours.

"What would make this company more powerful?"

"How can we make more money?" is also a useful question, but it tends to lead to answers that improve the current business only slightly. Thinking about how to make the company more powerful, on the other hand, can uncover structural changes that shift the company's value not by a few times but by tens or hundreds of times.

The starting point for that change is making sure the company does not remain a mere component supplier. Rather than only providing technology that goes inside other companies' products, it should, if possible, be the party that has the direct relationship with the customer. Another key question is the following.

"Is there a way to make money flow through this company? It's always good to have money flowing through you."

Having payments or tokens pass through your platform lets you hold the center of the customer relationship and the transaction. At the same time, you should also consider whether there is a risk that model providers will eventually absorb the platform or its customers.

2. Building Platforms and Network Effects

The most powerful form is an app-store-like structure where other companies build things on top of your product. As the value other companies create accumulates, the platform itself grows in value too. If you also secure the customer relationship and the payment flow on top of that, it becomes even more powerful.

If you can't build an app store, proposing a standard for how different companies' products interact is another option. New fields often have no standards yet, and because people want standards, an approach proposed early can take hold regardless of the size of whoever proposed it.

The author tries to look for network effects in almost every startup. Even services that seem far removed from network effects can often, surprisingly, introduce them. As a result, a simple service can even turn into a marketplace.

For example, you could let users see their results compared with other users', if they want to. For an AI product, you could let users voluntarily contribute their interaction data to model training. Some users would be reluctant, but those who opt in get a model improved by more data, which creates a virtuous cycle for the product. He explains that YC, too, turned out to have network effects in an unexpected area, though it wasn't designed that way from the start.

3. Broadening the Idea to Create a Market

Network effects often appear when you generalize the idea. For a startup that helps AI agents pay for goods or services, for example, the author says he would first ask, "Can agents pay each other too?"

If agents can transact with one another, the company can become not just a payment tool but the market itself. Even if it isn't immediately obvious what would be traded, marketplaces are so valuable that it's worth investing a long time in finding that possibility. If necessary, the company can even act as a market maker, becoming the buyer or seller itself to generate the initial trades.

Of course, these hypothetical mutations don't always lead to a good business. But the process of transforming the original idea into other forms itself helps you understand the essence of the business more deeply.

4. The Full-Stack Strategy of Swallowing Your Customer

The author says that manipulating startup ideas sometimes feels similar to the way a programmer works with code. One change that feels especially powerful is the full-stack strategy.

Instead of stopping at selling technology to company X, you use that technology yourself to do what X was doing and compete with X. In other words, you subsume the company that was your customer and switch to a structure where you meet the end customers directly.

Another form is gradually absorbing the customer's role by taking over, one by one, the tasks the customer found hardest. Taken to the extreme, the startup does all the key judgment and intellectual work, and the original customer handles only simple execution. At that stage, the real customer is the original customer's customer, and the initial customer effectively becomes nothing more than a conduit.

The core of this strategy is not simply stealing customers, but serving the end customer directly in a better way.

5. Finding the Real Business in a Peripheral Feature

Startup history is full of cases where a feature that was originally incidental became the main business. The classic example is PayPal: the company initially built security technology for handheld devices, and PayPal was a demo to show off that security software. But eBay sellers began using it as a payment method, and a few months later the founders admitted they were in fact in the payments business they had never intended to enter.

So whenever founders build something peripheral to their main product, the author always asks:

"Could this actually be the real product?"

When users use a product differently from how the founders intended, you shouldn't get angry and treat it as mere misuse. Users are forcing the tool they've been given to solve a problem they desperately want solved, even if it isn't the problem it was originally meant to solve.

"Don't get annoyed when users misuse your product; listen to the message they're sending. There may be something valuable in it."

This behavior is a signal that the market is sending strong demand. A startup has to read the signal hidden in how its product is used.

6. Help Customers Make Money, and Play the Long Game

An important reason PayPal grew so fast is that it helped users make money. Products that help customers generate revenue get adopted quickly, and customers are very willing to pay for them. So user growth and revenue growth accelerate at the same time. The author says many of the most successful companies YC has funded contributed to their users' revenue, and that YC itself is such a case.

Playing the long game is also a source of power for a startup. Many competing startups operate in hopes of being acquired, and executives at large companies often focus on quarterly results on the assumption that they may leave the company in a few years. As a result, choices that pay off big ten years out tend to be undervalued.

This is where the advice comes from to offer very good terms early on to win customers, and, if necessary, to sell at low prices to acquire users. At first, market capture and growth rate can matter more than margins. You should be careful, though, that excessive discounting, as when you grow by selling a product worth $10 for $5, can blur the value customers actually perceive and the demand signal.

7. Give More, Grow Bigger

Quoting Tim O'Reilly, the author stresses that a company should create more value than it captures.

"Create more value than you capture."

This may sound like idealistic advice, but it is actually a practical path to creating greater wealth. Squeezing every last penny out of existing customers can at most roughly double revenue. Creating entirely new value for customers, on the other hand, can lead to 10x or 100x outcomes.

This principle shows clearly in open source. A company gives its product away for free, but in return makes that product a standard and earns users' trust. Once the product spreads widely, the company gains more value even if it takes only a small share of a much larger market.

If full open source feels like too much, another option is to make the product extensible. The app store is an example at one end of the extensibility spectrum. The less you tightly control or charge for extensions, the more likely a larger ecosystem is to emerge in the long run.

Making a product callable through an API is the ultimate form of extensibility. Many companies are reluctant to offer an API because they dislike losing control, but the author thinks this is usually a mistake. In an era when AI agents rather than human users use products, it's hard to predict how they will combine and use your product. For an early-stage startup, which has little to lose, he recommends defaulting to offering an API.

8. Capture Customers and Data at the Earliest Point

Surprisingly, selling to earlier-stage companies is a powerful strategy for startups. Early-stage startups have little money, but if you charge based on usage, your revenue can grow at startup-level rates in step with how fast your customers grow.

Stripe focuses on signing up companies as customers at the earliest possible point. Payment infrastructure is an area companies rarely switch once it works, and companies that install Stripe generally don't churn. Founders of early-stage startups understand products well and make decisions quickly. If your product is the best, you are likely to win.

Conversely, if your product can only be sold once a customer has 500 or more employees, you can easily get stuck in enterprise sales, which takes a long time and involves complex processes. In that market, the best product doesn't necessarily win.

The author says the question "What size of company should we target as customers?" should be reframed like this:

"At what point in the customer's life will you acquire them?"

If you're confident customers won't churn later, why wait until they become large companies? It's more powerful to acquire fast-growing, quick-deciding early customers and grow along with them. To do this, the ideal product should be something that your fellow founders in the same batch can install and use right now.

The same goes for data. Rippling started with onboarding, the point where employee data is first created. Because it aimed from the outset to build an operating system for employee data and a range of applications running on top of it, it built a far better onboarding product than one aimed merely at the onboarding market. As a result, the product spread quickly, and the company captured the upstream end of the data flow.

"Whether it's money, user relationships, customers' stage of growth, or data, being upstream is almost always good."

9. Routing Around Slow Markets and Powerful Incumbents

Because startups generally build better products, they are strongest in markets where everyone competes fairly. Conversely, they are weak in markets dominated by powerful incumbents, such as record labels or PBMs, which the author likens to a "mafia." In those markets, a good product alone rarely wins, and you may only be able to exist to the extent the incumbents allow.

That doesn't mean such forces can't be beaten. But instead of a head-on fight, you have to go around them. In other words, rather than trying to defeat the incumbent directly, you have to win on a different dimension that gradually makes them irrelevant. That way, beating the incumbents becomes not a precondition for success but a side effect that follows from it.

Dealing with customers that make decisions slowly, such as hospitals or school districts, is also like wading through mud. In these markets, you should, if possible, go after faster-deciding customer segments first. For example, rather than selling to the school system itself, you could find a way around it by serving students directly.

10. Breaking Free From Other Companies and Your Own Fear

Finding ways to make a company powerful is often a matter of finding ways to escape the constraints imposed on it by other companies. If you're a component supplier, you can escape a framework built by another company by owning the customer relationship directly. If you have to sell to a large bureaucratic organization, you can sign that company up as a customer back when it was smaller and made decisions faster, or adopt a full-stack strategy of serving the end customer directly.

You can turn this into a question for finding ways to expand the business:

"In what ways is the current idea constrained by other companies?"

Often, though, the biggest thing holding a startup back is not outside companies but the startup itself. The author says a surprising number of sentences in his advice contain the word "just."

"You don't need to do X. Just do Y."

The reason an early startup's idea gets tangled may be that it evolved from earlier ideas and never shed parts it no longer needs. But especially very early on, companies often choose to do something less ambitious than they're actually capable of, out of fear.

In these cases, "just do Y" effectively means:

"Just stand up straight."

When a company sets aside its fear and chooses a more direct and ambitious goal, it can become something far bigger than before. 🌱

11. The Condition for Power: Making Customers Better Off

All the strategies above share a common condition that must be respected. Introducing network effects, taking control of payment flows, or going full-stack isn't something a company can do simply because it wants to. Customers accept it only if the result is better for them.

"All of these strategies have to make things better for the customer."

Startups have almost no power at first. It is precisely because of that weakness that new startups generally have a good effect on the world. They can't force anything on customers, and can only grow by making customers' lives better.

The author says you can also apply this principle in reverse to find business ideas. First, imagine what a perfect world would look like from the customer's point of view. Then, among the pieces needed to build that world, if there is a role a startup could take on by changing itself, that is very likely what the company should become.

12. Closing

A startup's power doesn't come from nudging short-term revenue up a little. It comes from owning customer relationships and transaction flows, building ecosystems and network effects, and capturing the upstream position in customers, data, and market-entry timing.

But the most important criterion never changes. Every structural change made in order to become powerful must actually deliver greater value to customers. The way a startup moves past its initial weakness and gains lasting power is, in the end, the same as the way it helps its customers better.