The Better Customer–Startups or Big Enterprise?

I just want to turn my startup into like a real-time strategy game where I can sit at my computer and click on things and watch numbers go up. If I can do that and just sit on the couch and have people bring me food and I click things, we're in good shape.

[Music]

Hello, this is Michael with Harj. Welcome to Inside the Group Partner Lounge. As Y Combinator group partners, we find ourselves often repeating the same advice over and over again during our office hours with the startups we work with. For COVID, we'd often gather together in the Group Partner Lounge at the Y Combinator office to try to figure out why this was the case and how we could help startups figure it out faster. Today, we're going to talk about when it is a good idea to sell software to startups.

Harj, do you want to set this one up?

Yeah, so you know, I think especially amongst YC companies, selling to other startups has been a formula for success. Some of our top companies, like Stripe, started out being just a payments processor company for other YC startups for a long time. PagerDuty, one of the first YC companies to go public, had all of their early customers as other startups. Even when you read their S-1 filing when they went public, they were still referencing lots of those startup customers years later. So, it's clearly something that can work really, really well.

I think the interesting thing, though, is when it doesn't work right. When founders are kind of maybe cargo culting, not understanding why these folks do this. What do you think about that, Brad?

Yeah, so people have seen that those companies started off selling to startups and assume that they should do the same thing too. But it's not always the best move for them. There's a whole lot of reasons why companies think that they should be doing it and it's not the best move for them. I think the first one that I see all the time, and Michael, we've worked with some companies together that did this, is that their product naturally should be sold to bigger companies. They're the only people that want it, but the sales cycle seems really long.

They're worried that if they just try going after those big companies, they won't get any traction and the company will never take off. Therefore, they should try selling this very enterprisey product to startups. This one is so painful because you know we will see a founder who worked in some large company for three, four years, who saw a problem in that large company, who has a ton of insights on how to solve that problem for a large company. It's a problem that tiny companies don't have at all.

One of the core reasons why we funded them was because they have this experience and they understand large companies so well. Then, in the first office hours, it's like, "Oh no, I'm going to sell this enterprise-grade alerting dev tools system that is really only useful when you have over 500 engineers, but I'm gonna try to sell it to startups for $100 a month, $400 a month."

Yeah, yeah, and like three startups will talk to me about it. So, that's evidence that it's solving an important problem for them. I don't know about y'all, but that was always that office hour moment where I'm just like, "Oh no, please. How do I get you to not think this way?"

And this is like classic examples of this, right? If you look at companies that went public even recently, like Snowflake. Snowflake is basically a company you only need that product if you have massive amounts of data, so in warehouses, right? If you tried selling Snowflake to a startup, you'd just waste your time.

I heard a story I remember about Workday, a big HR company. Apparently, all of Workday's early customers were doing six, seven-figure deals with them because you only needed some big comprehensive HR suite if you already had hundreds of employees. Two amazing examples.

The other lie that I see early-stage companies telling themselves is that it's going to be easier to sell to startups, like it's going to be easier. I always try to push back on that and say, "Why would it be easier to sell to a customer who doesn't have the problem than to sell to a customer who has the problem but takes a long time to make decisions?"

I understand it would suck to sell to a customer that takes a long time to make decisions, but it only seems like it would be worse. The only thing worse than that would be to sell to someone who doesn't have the problem at all. Yet a lot of times, founders are not seeing that weird, horrible path they're walking.

It's the same thing with the last lie where just because someone's willing to talk to you at all doesn't mean that that's an easier sell. People confuse that these other startups are willing to talk to me about this problem even if they don't actually have it, therefore I should be selling to them.

I think another version of that is, let's assume that startups do have the problem. I think sort of a lie or mistake founders will sometimes make is they'll assume that the startup's an easier customer than the enterprise customer. That, "Oh, they'll be lower maintenance," or "They won't be so demanding." I think there's a grain of truth to it. Sometimes you have to jump to enterprise companies that make you jump through more hoops, fill out more forms, to get onboarded.

But once you've actually got someone who's using your software, there's not that much difference. The end-user can be just as annoying and just as high maintenance, bugging you every day, whether they're working at Microsoft or they're working at some small startup.

What's funny is when that end-user is at a big company, they're probably paying you a lot more money. You can afford to support people to reply to them. Startups can be super demanding and not want to pay, which is a great combo.

The other one I was thinking about a lot is we have a lot of YC companies that think, "Oh, I'll sell to an early-stage company and then they won't churn as that company grows." But what's interesting is in some categories, it's almost assured. You mentioned HR; almost every single early HR company gets churned. All their customers churn as they grow.

Another classic example that we talked about before was Heroku, where all the customers churned as the companies grew. Just because you get someone early doesn't mean they'll stick.

Right, not at all. I think that's the number one aspect of this selling-to-a-startup strategy where companies fall down. If you do have a product like Stripe, where you can actually just grow with the companies, they never churn. It's a fantastic business model. You get businesses cheap when they're like two people and they stick with you when they're like thousands of people and doing billions of dollars in revenue.

Or like awesome, but I think you have to be thoughtful about whether that's actually going to happen or you're going to end up more like an HR system or a Heroku-type product where the company hits like 500 employees and is like, "Well, you're no longer working for us; this doesn't work for us. We need something brand new. We need you to custom build something."

Then, your choices are to let them go or become a sort of outsourced dev shop for some big enterprise customers.

Yeah, and what's happening there is the buyer is actually changing. I see this with my startup, Perfect Audience. Early on, we could sell to startups because it was just like the founder or one of the marketers. But later, once it’s a marketing team and once things have changed, they need a totally different product. It's the same with Heroku. Early on, the dev team is small; they have certain requirements and constraints.

But later, when the organization's a lot bigger, those constraints just become too much, and they can't just use something like Heroku anymore. We had the same thing with Triplebyte, where we were building software to help screen engineers, and the first version of it was a highly technical product built for engineers to screen other engineers.

It worked great when our customers were startups because the user was always like a CTO or the technical founder, or another engineer. But every company eventually hires a recruiter or recruiting team. Those teams are completely non-technical, and we basically had to build a brand new product to work for them because it was like completely different end-user.

Maybe what we're getting at is that with Stripe, they don't have to build a brand new product because it's doing something that companies just don't want to mess with, like payments. With this other stuff, you're going to be recruiting in-house, you're going to be doing marketing in-house, you're going to be doing HR in-house.

I was thinking CRM kind of falls in the same category where startups will pick any kind of lightweight CRM, and then the second VP of sales wants to scale a sales team to 500 people, it's Salesforce. It's always Salesforce.

That's such a tricky thing for founders to realize, maybe a lot of these have to do with the fact of when something scales with people versus revenue. HR tools scale with people; CRM scales with people. Whereas, payments doesn't really scale with people in the same way.

I don't know, maybe there's some insight there, but certainly there are certain categories that just religiously churn out early-stage products.

But maybe the last one on this front are lies around sales, right? I think that's the deadly one. What do you think about that one, Brad?

Well, sales is scary. If you're someone that has more of a product background and you find yourself building a software product for businesses, you just might not want to do any selling. The fallacy, I would call it the civilization fallacy, where you think, "I just want to turn my startup into like a real-time strategy game where I can sit at my computer and click on things and watch numbers go up."

If I can do that, like I'm really good at playing Civilization. If I can do that and just sit on the couch and have people bring me food and I click things, we're in good shape.

But the reality is: you have to go sell people and you have to change people's behavior, you have to go influence them, and get them on your team and get them excited about your product. And that's sales.

A lot of times, I think when startups want to sell just to other startups, it's because they are afraid to do the sales and they think that this would get them out of it.

Yeah, I love that one. You know, we’ll talk to an early-stage company, and their whole kind of sales strategy will be "put a self-serve flow and wait.”

It's like, "You know, MVP, self-serve flow, dot dot dot, riches."

Oh yeah, maybe some ads?

Yeah, yeah, maybe some press? Maybe, you know? And I just love the dot dot dot part. It's like clearly that's all you need, right?

We see this happen a lot during batches. It's like, "Hey, you need to go and do sales," and it's like, "Well, I've got all these other founders who are in my batch. I interact with every day anyway, so there's no extra effort for me to just drop in; like, 'Hey, do you want to try out my product?'"

But it's like, that's a perfectly fine place to start. But it's like, there's so much of sales. It's like, how do I get anyone to know who I am? How do I get someone to reply to an email? What's the right messaging?

You just lose all of that when you just ping people that are accessible.

Yeah, this is a pretty basic thing, but it's just important to state that what you build is just as important as what you build to get customers. A lot of startups don't understand that, and just building a great self-service interface is not necessarily a strategy for getting customers, and you've got to have that figured out, I think, in order to build a big company.

Now let's talk about the flip side here because, as you said in the beginning, Harj, selling to startups is one of the core strategies that YC companies have employed to create billions of dollars of value. I would argue it's one of YC's superpowers. You join YC, you got so many, literally thousands of potential startup customers you can sell to.

So, let's talk about how selling to startups can be a powerful strategy. Are there any kinds of examples here?

Yeah, I mean one that springs to mind is I think of Gusto, the payroll company that YC funded. They're doing tremendously well, right? I think they started with startups and they became like the de facto payroll provider for YC companies. They used that actually as sort of a bridge into the SMB market, just the broader small business market—a huge market in America.

That's where they found the sweet spot. My impression from the way they operated is that although traditional SMBs are different businesses from startups, the product, the software, is very similar. They could basically train themselves, build a really great product, build a really great user experience with startups, and then branch out into the SMB market with a product that was already really great.

I think another example here is that when selling to startups, it creates a path to the large enterprises. When you think about when AWS first came out, Amazon didn't have the kinds of relationships with large companies that Microsoft did.

If they wanted to eventually become the product that those large companies use to bring their compute to the cloud, how do they penetrate those markets? It turns out selling to a bunch of smart engineers at startups—those things they loved—was a great strategy.

Once one startup started getting growing on it, those engineers might go from a startup to a larger company and try to bring them in. They were just training this whole group of people on how to use cloud software. That was a significant avenue into the enterprise that they had to invent because they couldn't walk the Microsoft path of just calling up everyone who uses Office and saying, “Do you want to use my cloud computing thing too?”

I think there's a really important point about that, which is that has become known as the bottoms-up sales strategy. It's become quite popular. It's like, "Hey, you get your developers, you get people to be evangelists, and then they help you make the sale."

That is true, but a point that people forget is those companies, like AWS or Stripe, who’s another popular example for that strategy—they still end up building real enterprise sales teams. This is supposed to be something that gives you a booster; it helps with those efforts. It's a really smart strategy.

But eventually, you still have to have someone that's making the enterprise sale. It’s just like hopefully now they can say, "Look, your whole development team is already using it. They're already advocates. You really should get this sale done."

It's not like Walmart is just going to walk into your self-serve flow and be like, "Yeah." They don’t just sort of magically one day switch themselves on because some of their engineers said that they should, right? You're going to have to make a sales deck.

So, let's wrap this up. What is the big takeaway here on when it's a good idea to sell software to startups and when maybe it's not?

The biggest picture way I could frame this is that you have to know which game you’re playing. It's absolutely reasonable to have startups as your first customers if you're, "Hey, my game is I'm playing the Stripe game," which is I'm going to get these companies when they're like two people and an idea, and the product's going to stick with them all the way through to IPO and billions of revenues.

Or you're like some of the other things we said, hey, like, "I'm going to be like Gusto. I'm going to use startups to build a really awesome product, and I'm going to take that product and launch it in different markets."

But the key thing is you have to know what game you're playing and how selling to startups fits into that, right? The game of, "I'm going to sell to big companies by first selling to people that have nothing in common with big companies," that's not a good game.

I think what's funny about this, and my kind of final takeaway after talking to so many founders, is these games are out there. You don't have to invent this stuff. You can actually try to understand the difference between Gusto, AWS, Stripe, and you can copy their strategies, which is very different from product where you really do have to do a lot more innovation.

So, oftentimes my best advice to the YC company is, "All right, which of these paths do you want to go? Who is the company that's nailed that in the last 10 years? Go study them. Learn everything about them. You don’t have to learn all the hard mistakes that they did; you can learn from them."

Awesome! Well, Brad, Harj, great to see you both. Thank you.

Thanks, guys! Bye.

[Music]