Startup Business Models and Pricing | Startup School

Foreign [Music] I'm Aaron Epstein. I'm a group partner here at Y Combinator, and in this video, we're going to be talking about business models and pricing. There's three main things that we're going to cover in this video. The first is the nine business models of nearly every billion-dollar company. It turns out there's just a handful of them that build the biggest winners.

Next, we're going to talk about business model lessons from the YC Top 100 companies list. And finally, we're going to cover some startup pricing insights that we've taken from the thousands of companies that have gone through YC.

So first, let's talk about business models that build winners. If you're not familiar, a business model is a fancy term for how you make money. It turns out the business models are important because we see founders that often get frustrated when investors won't fund them, and their business won't grow. Oftentimes, they're not sure why. Usually, this is because they're not using a proven business model, and there are actually only a handful of business models that are responsible for nearly all billion-dollar companies.

Rather than trying to reinvent the wheel, you should actually just copy one of these. Here they are: Nearly every billion-dollar company is one of these nine business models. There's SaaS business models, which is Software as a Service, which is cloud-based subscription software that customers pay either monthly or annually in order to access the software. There's transactional business models that facilitate transactions and take a cut of those transactions. These are often fintech companies.

Then there's marketplaces, which facilitate transactions between buyers and sellers. These are often referred to as two-sided marketplaces. There's also hard tech businesses, usage-based business models, enterprise, advertising, e-commerce, and bio.

In this video, I'm actually not going to get too deep into the specifics of each of these business models. Instead, we're going to have a business model guide that I've put together that's going to be linked in the description down below. This guide is going to cover the metrics that matter most for each business model, key takeaways for each of them, and other similar companies that you can learn from depending on which business model you are using for your company.

In this video, what I want to focus on is things that we can learn from the top 100 YC companies. The top 100 YC companies is pulled from ycombinator.com, top companies, which is a list of the most valuable companies that Y Combinator has ever funded. For the purposes of this video, I've gone through this list and matched each company up with their primary business model to try to see what interesting insights we can get from them.

Now, some later-stage and larger companies actually have multiple business models. However, for your purposes as an early-stage startup, you should just have a single business model that you're focused on. So here they are: These are the top 100 YC companies organized by business model, and there are some interesting things that we see here.

First is that SaaS businesses actually make up 31% of the top 100 YC companies. Transactional businesses make up 22% of the top 100 YC companies, and marketplaces actually make up 14%. So just with these three business models—SaaS, transactional, and marketplaces—it makes up 67% of the top 100 YC companies.

On the flip side, with business models like advertising and e-commerce, they barely register on the top 100 YC companies list. If you're familiar with startup outcomes and venture capital returns, you know that there's a power law effect, which means that the biggest winners far, far outperform all other businesses by orders of magnitude, and this is true for the YC top 100 companies list as well.

Turns out that 50% of the overall value of the top 100 YC companies actually comes from just the top ten. It's interesting to look at what insights we can get from these ten companies too. Here they are: These are the top ten YC companies by value. There's Airbnb, there's Stripe, there's Instacart, there's Coinbase, there's DoorDash, there's Reddit. There's a number of companies here that you're probably very familiar with or use on a regular basis.

What's especially interesting is that five of the YC top ten are actually marketplaces. There’s Airbnb, there’s Instacart, there’s DoorDash, there’s OpenSea, and there’s Fair. The interesting takeaway here is that marketplaces are most likely to build winner-take-all companies. They tend to become so big and dominant in their industry that it doesn't leave much room or market share for other competitors once marketplaces actually get huge.

So marketplaces are 14% of the top 100 companies, but they actually create 30% of the overall value because so many are represented here in the top ten. While marketplaces are really tough to get off the ground, they have a chicken-and-egg problem where you can't just build your product and then sell it to customers; you actually need to solve for both sides of the marketplace, the supply and the demand, at the same time in order to get customers.

However, once they hit the inflection point and they start to work, they get massive network effects, where each new user of the platform increases the value for everybody else. That's what makes them dominant winners. You can think of companies like Airbnb; if you are looking to rent out a place short-term to stay, then chances are you would go to Airbnb because that's where all the inventory is.

Similarly, if you wanted to buy or sell NFTs, you would probably go to OpenSea because that's where everyone is. That's how these become the big winners. It also turns out that three of the YC top ten are transactional businesses too. These are companies like Stripe, Coinbase, and Brex, and the main takeaway here is that transactional businesses far outperform because they're directly in the flow of funds. This means that they are the platform that money flows through, making it very easy for them to just take their cut.

Transactional companies are 22% of the top 100 YC companies, but they actually create 29% of the overall value, and this is because they are as close to the transaction as possible. This was advice that I received during my YC batch back in 2010, and that was to get as close to the transaction as possible. If you're a company like Stripe that literally processes money for companies, or Brex that is the corporate card they use to spend money, then you're directly in that flow of funds, and so it's really easy to take your cut.

On the opposite extreme, if you are an affiliate business, multiple things have to happen before you ultimately get paid, which means that you are very far from the transaction, which makes those not as good of a business for transactional businesses. Because they're so close to the transaction, they often become critical infrastructure for other companies that they build on top of, and that usually means that they are solving a top-three problem for them.

So you can imagine if you use Stripe as your primary method to get paid from your customers, the thought of ripping that out sounds terrible. You would never want to do that, and that's why these transactional businesses become so dominant. We also see that SaaS businesses are most likely to make the top 100 list, and this is because they have consistent revenue.

So 31% of the YC top 100 companies are actually SaaS businesses; that's nearly a third. This is because the recurring revenue makes them great businesses. This means that customers keep paying them every single month or every single year until the customer explicitly says to stop. So this has lots of benefits, including the predictable revenue that they get, which allows them to compound and grow their business.

We can also see that very few advertising businesses become big winners, and this may be surprising because we're so familiar with so many companies that have built their business off of an advertising business model. There's Google, there's Facebook, there's Twitter, just to name a few. But really, only 3% of the top 100 YC companies use an advertising business model as their primary way to make money, and that's because advertising businesses need organic virality to win.

They need to catch lightning in a bottle and become the hub where all users go to hang out or to see live streams in the case of Twitch. When that happens, they get really strong network effects, just like marketplaces. So people go to hang out on Reddit and form communities there because that's where everybody else is. People go to Twitch to watch live streams because that's where all the streamers are.

It's really important to remember that you should not use ads as your primary business model unless you expect to be a top-ten site on the internet. Otherwise, it's too hard to monetize and build a huge scale to become a massive company.

So what are some overall lessons that we can take away from this list? First, it's interesting to look at what's not in the top 100 list. There are no services or consulting businesses there, and so it can be a good idea to start doing services or consulting for your customers primarily as a way to learn and make sure that you're building the right product for them. However, consulting businesses suffer from having non-recurring revenue, scaling with people rather than software, and having very low margins as a result.

So that's why these businesses are not venture-scale. Similarly, affiliate businesses tend to be too far away from the transaction; that means that you have to acquire a customer, send them off to another product or service, hope that they actually make a transaction on that other product, or you will get some small commission from that 30 to 90 days later. So it's too hard to make a lot of money at scale doing an affiliate business.

Similarly, hardware businesses require lots of capital to get off the ground to buy physical parts, and they have low margins as a result. So that makes it really difficult to start these businesses and also to scale them. Even if they're working, just requires so much capital. Businesses that are built on other platforms don't show up in this list either; that's because they tend to have a lot of platform risk.

If your business is built on top of another big successful platform and your business starts to work, then it's actually in the interest of that platform to shut you down and capture all of that revenue for themselves. So that's why even if these look like they're working in the early days, they can be turned off at any moment.

We also see that recurring revenue consistently creates winners, and this is because it is highly predictable. Once a customer has committed to pay, they're going to continue paying until they explicitly say that they want to stop paying. They also have higher customer lifetime values versus one-off transactions, and this results in lower customer acquisition costs.

So you don't have to keep reacquiring customers over and over. If you have a one-off transactional business, then you have to invest money in acquiring that customer the first time and then also keep putting more money into trying to get existing customers to spend more with you. That's not the case with recurring revenue businesses.

But recurring revenue only works when you have strong retention. It's not enough for your product to deliver value right up front and then never again. You need to keep delivering value over and over again, otherwise your customers will churn and stop paying, and then you can't scale a leaky bucket. If you have lots of churn, to give you an example of that, if you had 95% monthly retention for your recurring revenue product—so that means that 5% of your customers will churn and stop paying you every single month—and if you started with 100 customers at the beginning of the year, then by the end of the year you would only have 54 customers of your original 100.

That means that you would lose 46 of your customers in just one year and you would need to get 46 new customers just to break even with where you started the year. Let's say, for example, you had 90% monthly retention instead of 95. Just a 5% difference there would actually lead to only 28 customers at the end of that first year. That's a huge difference and a huge hill to climb.

So just that 5% difference in monthly retention can actually be the difference between life and death for a startup. We can also see that some of the biggest winners are built with moats. There are network effects that many marketplaces have, right? Where each new user increases the value, and they become the dominant player in the market.

There's also lock-in and high switching costs; we see this with transactional businesses like Stripe. If you're the primary way that people actually accept money and process payments, then chances are they're not going to switch off of you. In SaaS businesses, you get the recurring revenue where customers keep paying over and over again until they say to stop.

You can also get lock-in by having customer data on your platform, and once they stop paying, all that customer data goes away. Then enterprise businesses, while they're often difficult and have long sales cycles to be able to sell into large companies, usually once you've sold into the company, the churn is a lot lower.

Technical innovation is another way to build really strong moats, and we see this often in hard tech and bio companies especially. You can think of companies like Cruise building self-driving cars and Boom, which is building supersonic jets. It takes a really long time to even get to a working product for these types of businesses, and so for anybody to compete with them, it takes years of difficult technical development just to catch up.

We also see that higher margins and better unit economics can build moats. In the example of companies like DoorDash and Instacart, they've reached economies of scale where they're so large now that they've been able to drive their costs further down at this scale and improve their margins, which new entrants are not going to be able to compete with.

Finally, if you get organic distribution for your product through virality or word of mouth, you can dominate your market through that as well. If you are able to get users for free because other users of your product tell new users to come join, and you're competing with a company that has to pay to acquire their customers, then you are going to grow much faster and capture way more of the market.

So to recap, the best businesses generate recurring revenue, have high retention, build defensible moats, are as close to the transaction as possible, they scale with software not people, and they're proven and use business models that are familiar to customers. It's important that you focus on innovating on your product; that's what should be new, and copying your business model from one of these proven winners.

All right, now let's talk about pricing. It's important to think of pricing as a tool to help you learn faster. It can help teach you who wants your product, how much they want it, how much value your product provides to your users, and which channels you can afford to use to acquire your customers.

To help you get started, I've compiled five pricing insights from the top YC companies. The first is you should charge. This is actually the most common mistake that we see founders make. Often founders are afraid to charge for a number of reasons. They're often afraid that their customers are going to tell them no, they're afraid that their customers are going to walk away and never come back, and they're afraid that their customers are actually going to go and use their competitor's product.

But it turns out that charging is actually one of the most effective ways to learn a lot of really important things about your business. The first is: Are your users even willing to pay or not? This is often binary, where either they're willing to open their wallet or they don't even see enough value in your product to overcome that hurdle.

It can also teach you which users are most willing to pay. If you're trying to decide whether you should go after customer segment A or customer segment B, trying to charge and figuring out which one is most excited to pay can give you really good signal on who wants your product more.

It can also teach you how much they're willing to pay. By setting higher prices, you can try to figure out how much value they see in your product. Even if everyone refuses to pay, that's still valuable information for you to get because it teaches you that you haven't built enough value into your product yet, or you're talking to the wrong customer segment.

Stripe is a great example of this. In the early days, Stripe wanted to test the amount of value that they were building in their product. So while most of their competitors were actually charging around 3% per transaction, Stripe decided to set their price at 5% per transaction—nearly double what their competitors were charging.

The reason that they did this is because they wanted to test how much value their customers saw in things like one-click sign up and being able to get started quickly, and really in-depth detailed developer API documentation that would help developers get started faster. Rather than trying to undercut the competition in order to win customers, they did the exact opposite and set a really high bar for themselves to prove that they had built enough value into their product.

So where should you begin? The first thing that I recommend is that you don't overthink it. If you look online, there are tons of charts, graphs, and formulas, and all these different complicated ways to maximize your pricing and figure out the right price to charge. But really, when you're just getting started, the important thing is to just find the right order of magnitude for your pricing.

What I mean by that is if you're charging ten dollars for your product and your customers are willing to pay a hundred, you should probably change your price—you're off by an order of magnitude. However, if you're charging ten dollars and your customers are willing to pay 15 or 20, don't worry about it; you're in the right ballpark, which is the really important thing.

Pricing isn't permanent—this is really important. It often takes years to iterate and capture the full value of the product that you've built from your customers. So don't worry about capturing that full value early on; you'll have plenty of time to maximize that.

The next insight is that you should price on value, not on cost. There's three important components here: the cost, which is what it costs you to be able to serve your customer; the price, which is what you're charging; and then there's the perceived value that your customers see in your product.

Founders often start with something called cost-plus pricing. I would not recommend this. What this usually looks like is looking at how much it costs you to serve a customer and then adding an amount on top of that—say, ten dollars—and that's your price. But this actually ignores the full value of what your customers see in your product.

The difference between your cost to serve your customer and the price that you charge— that's your margin. That's how much you make on each transaction. The difference between the price and the value that your customers see in your product? That's your opportunity to be able to raise your prices to be able to capture more of that perceived value.

Now, if your cost is higher than your price, well, that means that you're going to have negative margins, and you cannot scale a business with negative margins. Similarly, if your price is higher than the value that your customers see in your product, that means they're just not going to buy from you.

So how do you find your value? Well, there are a couple interesting ways to be able to do this. The first is to talk to your users. You can ask them about the problem that you solve and get them to articulate the value to you. What this often looks like is if you reach out to a customer, and you get them on a call, you can ask them, “What is the problem that you are hoping our product could solve for you?”

They'll often tell you. Similarly, if you have a user that's signed up for your product but is not actually paying you, you can reach out to them and talk to them and ask them the question, “What problem were you hoping our product could solve for you?” Their response is usually going to be one of four interesting things.

The first is they're probably going to tell you that they were hoping you could help them make more money. This is something every company wants. Or they might tell you that they were hoping that you could help reduce costs—maybe your product saves them time or money. They might also say that they were hoping that your product could help them move faster.

Maybe they have something they were looking to launch in six months, and with your product, they can actually get it launched in a couple of weeks; that sounds really valuable. Or they might say that your product could help them avoid risk if you help with compliance or offloading something, a headache that they don't want to deal with.

Another way to find your value is to keep incrementally raising prices until you get pushback from users. When you keep incrementally raising your prices, you will ultimately find the ideal price, which is when customers complain, but they still pay. This is actually a good thing, right? It overcomes that fear of charging a high price and customers walking away.

The ideal scenario is when you tell the customer a price, they say they have to think about it, they go back, and then they come back to you a week later, and they say, “All right, that seems good. You're the best solution; we're willing to pay.” On the other side, if you were to actually charge a lower price, and they say, “Yeah, that sounds great,” and accept immediately, well, that probably means that you're pricing too low, and you're leaving a lot of money on the table.

Which brings me to my third insight: Most startups are actually undercharging. You almost certainly are. Lower prices are not a sustainable advantage. Sometimes we talk to founders, and they say, “Well, our product is just like our large competitor, except ours is cheaper.” That actually does not sound like a good idea; that's not a way to build a winner.

All that means is that your large competitor can underprice you even way lower than your cost because they have way more money, and they're substantially larger than you until they put you out of business. So I do not recommend having price as your only differentiator.

It also turns out that when you charge more, you get higher margins, and you're able to build a bigger moat. This means if you have higher margins than your competitors, you can pay more to acquire a customer, which means you can acquire all of the customers before they do.

It's also important to remember that pricing implies value. When customers are evaluating your product, they typically don't have a lot of signals on how valuable your product is, but the price that you're charging is actually one of the primary ones. So if your price is lower than your competitors, then your customers might assume that your product is less valuable than theirs.

Similarly, if you charge a higher price, then your customers might assume that your product is even more valuable than your competitors. So that can work really well. It turns out that raising prices is actually the easiest way to grow revenue. If you have a thousand customers and you want to double your revenue, well, it sounds pretty difficult to spend all the time, energy, and money to go get a thousand more customers.

However, if you're able to just double your price—just changing a number on the website or changing the price that you're quoting to customers in a sales call—and your product supports that higher value, well, you've just doubled your revenue with almost no work at all.

But what if users won't pay more? This usually means one of two things. It either means that you need to build more value into your product, right? Maybe the price that you have raised it to is now higher than the value that your customers see, or it could mean that you need to solve a bigger problem.

Maybe the problem that you're solving for customers is just a nice-to-have that they would never be willing to spend a lot of money for. In this instance, it usually means you want to move to a more important top-three problem that they have.

There's a third option too, which is you could give a lower price in exchange for one of four key things. One, you could give a lower price in exchange for your first user if you're just looking for initial feedback and getting somebody on the platform. Totally reasonable to give a lower price for that.

Or, if you're talking to a valuable customer that has a recognizable logo, that can be another good scenario where you would give a lower price. You can use this logo that you get as social proof to get other customers onto your platform at your regular price.

Also, if your product builds lock-in, say by getting customer data on your platform that they would lose if they leave, that can be another reason to offer a lower price. If you're able to renew after the first year and bump your customers up to that higher price, that can be a good reason to get people in at the lower price because you know you can capture more value from them further down the road.

It's also really important to remember that pricing isn't permanent. This is another common fear that we see from founders, where they're afraid that they have to nail their pricing the first time or they're going to lose their customer and never have a chance to get them again. Or sometimes founders are afraid that the set of customers that they're talking to are the only ones they're ever going to get, and so they have to close them all.

If that's the case, you should probably work on a different business. But it acts as relatively painless to be able to increase prices on customers over time too. There are a couple different ways to do this. You can exclude existing customers by letting them keep their current pricing and only raising prices for all new customers—that's one way to do it.

Or you could give advanced notice that you plan to raise prices, and as long as you build in enough value into your product to cover that price increase, you shouldn't see much churn. Most people will probably be willing to pay it if you have a sticky product. Netflix is a great example of this. This chart actually shows price increases that Netflix has made over the last seven or so years, and it's really interesting to see that they are not shy about raising prices on their customers.

Now, Netflix has 221 million paid subscribers, and they've been able to figure out how to raise prices because that is the easiest way for them to grow revenue rather than continuing to try to scale subscriber growth at the same rate. If Netflix is able to find a way to increase their prices on 221 million customers, you should be able to figure out how to do it on your handful of early customers as well.

The fifth insight is to keep it simple. This is an example of a pricing page for Quicken, and as you can see, it's very complex. There's five different buy buttons; there's prices, there's 349, 399, 599, 899; there's crossed out prices; there's one dollar off with a symbol right next to it. It's really complicated to figure out even if you want to be a Quicken customer which plan you should go with, and so this likely results in decreased conversion rates.

It's important to remember that when you're creating your pricing, you don't want it to create friction that prevents customers from signing up and paying you. On the flip side, here's a great example from GitLab. They have three very clear simple plans with clear pricing, and so their pricing and their pricing page is not going to be the thing that adds more friction and prevents customers from signing up and paying them.

I'll leave you with this story of Segment, which helps companies capture and use their customer data. When they started out, they were a couple of engineers that were not used to paying for products themselves, and so they thought they had to give their product away for free in order to get anybody to use it.

Then they wanted to raise money from investors, so they decided maybe we should actually charge our customers money so we can show revenue growth. Tail between their legs, they reached out to all their free customers and sheepishly told them that they were actually going to start charging them ten dollars per month, which was a hundred and twenty dollars per year. They were really nervous about telling their customers this, but surprisingly their customers started responding to them with messages like, “I hope you would charge me more than that; otherwise, I'm worried about keeping my customer data with you.”

The low price was signaling to their customers that maybe their product was invaluable, or it couldn't be trusted in the long term. To grow even more, they hired a sales advisor, and that sales advisor told them, “You should not be charging a hundred twenty dollars a year; instead, you should be charging a hundred twenty thousand dollars per year.” This is an enterprise product, and this scared them to hear this. It was unfathomable to them that anybody would ever pay a hundred twenty thousand dollars a year for their product.

So, when they were going into their first sales meeting with their sales advisor, the advisor told them, “If you don't tell this customer that your price is a hundred twenty thousand dollars, then I quit as your sales advisor.” They went into the meeting, and at the end, when it came time to talk price and the customer said, "So how much is it?" The CEO got really red, he got nervous, and he said, "A hundred twenty thousand dollars." The customer responded, "How about twelve thousand dollars?" They ultimately ended up agreeing on eighteen thousand dollars as their price.

While they didn't actually get the thousand X price increase, they were able to increase their price 150 times from 120 dollars a year all the way up to eighteen thousand dollars a year. It wouldn't have happened if they didn't ask for the higher price. They used this philosophy to continue growing their deal sizes all the way up to six figures and beyond and ultimately led to their acquisition by Twilio for more than three billion dollars.

So the story of Segment hopefully is instructive to you. They started out giving away their product for free and ultimately ended up selling to huge enterprises and building a business worth over three billion dollars.

To wrap up, the five key pricing insights are: The first is you should charge. Next is you should price on value, not on cost. The third is most startups are undercharging, and you probably are too. The fourth is that pricing isn't permanent—don't have fear that you need to get it right the first time; you can change it over time as you learn more and build more value into your product.

Finally, keep it simple. Don't add complexity which adds friction to customers giving you their money. Thank you. [Music]