How To Price For B2B | Startup School

[Music]

Hi there, my name is Tom, and I'm a partner here at Y Combinator. Today, I'm going to be talking about one of the most common questions I get from founders, which is how to price.

So, the founder's been working on outbound sales, contacting people, and they finally had a sales call that went really, really well. The champion, that's the person at the customer who's potentially buying, is really, really interested in the product and asked us for pricing, and we just froze. What number should we pick?

Often, when you haven't worked at a big company, you don't have good calibration about what kind of prices these companies tend to pay for software. You might think of the last time you bought software, you know, a subscription to GitHub or ChatGPT, and you pick a ludicrously low number—$19 a month or $49 a month or something like that—because for founders who have spent the last two or three months building a product, asking for tens or even hundreds of thousands of dollars can feel very uncomfortable. You almost can't say it with a straight face.

So, what I'm going to talk about today is a way to come up with a price and justify that price to your customer. There are three core elements here, and by far, the most important is what I call the value equation.

So, this is the idea that you sit down with your champion, that's the person at the customer that’s really into your product, that perhaps sees it's going to solve one of their biggest problems, and you write down, with this champion, what they expect your product to do for them—what value it's going to deliver to their company. That might be a cost-saving, it might be a time-saving, or an increase in revenue.

You've got to write this down step by step and then get the customer to challenge it, to prove it, and really make sure the assumptions are correct. Because ultimately, it's a tool for that person to take to their boss or their CFO to justify the purchase of this contract.

So, we're going to walk through a quick example here. Say you're selling a, I don't know, a customer service tool to a big company that has 100 customer support agents, and maybe just for the sake of argument, each customer support agent is paid $50,000 a year in salaries, and then there's another $50,000 per employee in additional costs that might be offices, overheads, health insurance, all that stuff.

So, the fully loaded cost of each customer service person is $100,000, and they've got a hundred of them, so that's $10 million of total customer service cost. And say we're going into this customer saying we've got this new AI-powered customer service tool that will eliminate 20% of the queries or 20% of the total time spent by that customer service team—that's $2 million of potential cost saving.

So again, remember you're normally saving time, which is cost, or reducing cost directly or increasing revenue; those are typically the three things that companies care about. Once you've established what value you're delivering, the pricing is pretty simple.

I typically pick somewhere between 25% and 50% of the value you're delivering, so they keep roughly two-thirds, you keep roughly a third. So, our previous example—$2 million of savings—they keep $1.3 million; you charge them maybe $700k, something like that, and it's a great deal for both of you. This person can take it to their CFO and show a really good return on investment.

So, the great thing about this value equation is it also gives you the success metrics that you need to prove during a pilot project. You might go to the customer and say let's try this tool for a month with just a portion of your team; maybe let's get 10 customer service agents to try it out and see if it actually does reduce queries.

Let's measure it, and as long as it reduces queries by 20% or saves the customer service agents at least 20% of the time, we know this value equation holds. If the metrics come back slightly different—maybe it only saves 15%, or it does really well and saves 25%—you can even adjust the pricing based on that, but the value equation tells you the success metrics that you need to prove during a pilot process.

So, that's the first part of pricing, and by far, it's the most important. If you just stop with that value equation, honestly, you'll get 80% or 90% of the pricing spot on. But there are a couple of other elements; it's usually useful to consider. The first of those is cost. What does it cost you to provide this service to the customer?

It's important you never start with cost. Some people like to do a cost-plus-margin pricing, and it just always ends up with you underpricing your software. Cost should only ever be a floor. So, you do your value equation, you take a third of it, that comes out to $700,000. Perhaps your costs are mainly OpenAI fees or something like that and AWS fees, and that comes to something like $200,000. So $700k is your contract value; $200k is your cost—you're golden.

If, however, you've come up with a value equation, and your share of it only comes in at $150k and your cost is $200k, you're in a bad business. You'll have to price below your cost, which is not sustainable, and so you're either going to have to figure out how to demonstrate more value or change what you're building ultimately, or get out of the business entirely.

Really, you should be aiming for software margins of like 80% or 90%. Note on credits—people like AWS, Microsoft, OpenAI giving out tons and tons of credits to startups—you should treat those as a cash cost. Don't assume you'll have unlimited credits forever; it'll totally mess up your margins.

There are occasionally situations where you might want to price at or even below cost, but it's a really, really risky maneuver, typically used by founders who want to grab market share in an arms race. It's a land grab situation; it's very, very dangerous.

You're really betting that the costs are going to come down dramatically in the future. Although, having said that, with the development of LLMs, companies like OpenAI and Anthropic do keep dramatically reducing costs. So, there's perhaps some argument that you might want to price a little bit lower now because you know your margins are going to improve with time. But honestly, it's a really risky maneuver, and I would advise startups to really try to keep to that 80% or 90% gross margin.

So, the third element in pricing is competition. You've done your value equation; you've calculated about a third of that; you've checked that your costs are way, way below that, so you're sustaining a sort of 80% or 90% margin. You're in good shape, but you have a direct competitor who's just entered the market, and their software is equivalent to yours, and they've decided to underprice you by half. What do you do?

This is really, really tricky. A founder's first response is to often just engage in a price war—to take the price that your competitor's offered and then just undercut them. The problem is they'll do the same, undercut you, and then you'll undercut them again, and it's just a race to the bottom. So, competing solely on pricing really is not a winning maneuver. You don't want to get into a head-to-head bidding war for a commodity product.

Instead, what you want to do is set your product apart based on functionality or value. It can't be an apples-to-apples comparison. Your product needs to be differentiated. If there's extreme competition in an industry for a commodity product, that's a product where they're all basically the same; all the margin gets driven out.

So, you take the airline industry as an example; basically, commodity: taking a seat in an airplane across the country. The airline industry, on average, has a 2.7% net profit margin; it's a brutal, brutal business, and airlines are on the brink of going bust the whole time because they're just struggling to differentiate.

So now we've talked about the three main elements: starting with the value equation, considering your cost, and competition. We're going to talk about other techniques for determining the price or maybe even the pricing structure.

Another question you want to ask your champion is how and what do they pay for other similar software products. For example, are they used to paying a monthly flat fee, or per seat pricing, or usage bands, or credits? I would really explore the industry you're selling into and understand what they're used to paying and the way they're used to structuring pricing. Then pick a pricing strategy they're used to.

People typically are wary of totally uncapped usage-based pricing, so you might want to put a cap on that. Really, if you can mirror the way they are used to paying for other software, the better you'll do. Overwhelmingly, though, when you're choosing pricing, it's important to keep it simple. Overcomplicating pricing will kill a sales process.

In general, committed recurring revenue—that's monthly recurring revenue or annual recurring revenue—is preferable to usage-based pricing, and that's because during an economic downturn or a slowdown, your revenue is protected, at least until the contract is up for renewal, and then you can have a debate with the customer about whether it's worth renewing.

Whereas if it's pure usage-based revenue, there's a real risk that your revenue just falls off a cliff in a bad month, and investors are really worried about that. So, if you can aim for MRR or even ARR, one technique to do this is to start with usage-based pricing for new customers, run the contract for a month or two, and see what usage is, and then offer to move them to a minimum monthly commitment with volume discounts.

So you can see they're using on average $15,000 a month; offer them a $12,000 a month flat fee which includes all their usage if they commit to a 12-month contract. Another technique is to ask your champion what amount they're able to personally sign off without having to get additional approvals from the CFO or the legal team.

So maybe they have signing authority of up to $15,000; that's a good hint that you should keep your pilot pricing to maybe $14,999, something like that, just to keep it moving really, really quickly. Next, I want to talk about whether you should publish your prices on your website or have a contact sales for enterprise pricing.

There are strong feelings about this on the internet, and often software developers say, "I just want to see the price. I just want to click a button and put my credit card in to buy. I hate talking to sales; why do I have to talk to sales?" The problem is the value equation is going to be different for each enterprise customer.

That's why most enterprise plans say contact sales. If you choose a price randomly, and put it on your website for an enterprise contract, you are certainly leaving money on the table. You'll overprice the product for a whole chunk of customers who don't get that much value out of it, so you've lost them entirely. For the people who get much more value out of it, you're underpricing it for that segment dramatically.

So, typically, what companies do is have one or two cheaper plans—perhaps an individual plan and a small team or startup plan—that contains most of the basic functionality but excludes the core functionality that enterprises really, really want. So you can go and look at other SaaS company pricing pages and see what they gate behind enterprise plans.

Often it's things like SOC 2 audit reports, single sign-on, audit logs, compliance reports, or data being kept in certain geographies—things like that—that really individuals and small companies don't really care about, and enterprises find absolutely vital and can't live without. That allows you to price differently for your small customers and your enterprise customers—sometimes up to 10 times more for enterprise customers for these extra compliance, legal, data privacy kinds of features.

The next thing we should talk about is understanding that your pricing strategy dictates your sales channels. But differently at this pricing level, is there enough money in each contract to compensate a sales team or account executives?

So a good rule of thumb is about a 5:1 ratio between new signed ARR and total compensation for a salesperson, including commission. For example, if you pay a salesperson $100,000 a year in annual compensation, including their base salary plus any sales commission—so $100,000 total compensation—you might reasonably expect that salesperson to close $500,000 of new ARR every year. But $500,000 of new ARR can be split down in so many different ways.

Is that $500,000 contracts, in which case each account executive is really hunting whales? They're trying to close a contract every couple of months, something like that, and they might only be working on four, five, or six contracts at a time? Or is it 20 times $25,000 contracts?

So, a $25,000 annual contract is about $22,000 a month, and your account executives would need to close about two of those per month, just under two, to meet 20 per year to get to that $500,000 contract. Still doable? Or are you tasking them with closing $500,000 annual contracts every year?

So, that last example of $1,000 annually is about $83 a month. Your sales rep has to close 42 deals every month; that's almost two every single working day. That's not really a true account executive or outbound sales team. At best, you might have a call center of inside sales, so they're basically picking up the phone when someone wants to buy and, you know, typing stuff into a computer system and answering questions.

They're not out hunting whales anymore; they're really sort of harvesting a field of wheat or something. The next subject I want to talk about is whether you should offer free trials or pilots. In general, offering a really long free trial or a pilot is counterproductive.

The customer has not actually bought into using the product, and so what you want to do is keep these pilots or these proof of concepts really, really short—maybe a couple of weeks, maybe four weeks, with really, really clear success criteria from what we talked about earlier with the value equation.

A better technique, if you're really confident, is to push your customers to sign up for an annual contract from the very start, but with a 30-day or 60-day money-back guarantee and opt-out at the start—so if it doesn't deliver what they want, they can get their money back, no questions asked. But by default, it becomes a recurring contract, and you can count this as recurring revenue pretty much straight away.

Another question I get is, we're just a two or three-person startup; should we put up more people on our website or sign more people up to our LinkedIn company account to act like we're a much, much bigger company to try and close these customers? In general, no, that's not a good idea.

What you should do instead is play to your strengths as a startup. Say to your customers, "You can have the phone number of the founders, and we're on call 24/7 to come and fix your problems." You're certainly not going to get that from Salesforce or Oracle or something like that, so play into your strengths as a startup.

Okay, in conclusion, if you really, really don't know how to price, you've tried the value equation, and it's just not worked; you're too uncertain somehow, honestly, just pick a number that's similar to other kinds of software that your customers buy. Then what I would do is increase that number by 50% for every time you pitch a new customer.

So you start at $10,000 and they say yes, and the next customer, you go to $15,000. And the next customer, maybe you get a $22,000. When you start to lose more than 25% of potential deals based solely on price, you're now probably in the right ballpark. You don't need to win every single deal, and if every single deal is closing straight away, you're almost certainly underpricing.

Just remember, if your company is successful, these initial 5 to 10 customers you sign up at the start are going to be a tiny, tiny fraction of the revenue you make over the next five years. So it's more important to start signing deals and getting into the flow of it. You can always increase prices as the product improves; you add new modules and put them behind a paywall, or upsell customers.

So, over-optimizing for pricing early is a big mistake. Pick a number, try and sell it, and just experiment. As you move on, it gets easier and easier to close customers and increase prices as the company grows because you presumably are getting better at sales over time.

You're getting validation logos on your homepage of all of the happy customers who are using your product, and unless you're doing something really wrong, your product should be improving dramatically over time. The first two or three sales are normally the absolute hardest you'll ever have to do, so just get them closed.

So, to recap, there are three really, really important parts to any pricing: first is the value equation—write it down, get your champion to challenge it, and then price at about a third of the value you're delivering so the customer keeps two-thirds of the value, you keep one-third. Second, consider cost—make sure you're not pricing at or below cost unless you've got a really, really good plan to dramatically reduce costs in the short to medium term.

Then thirdly, if there's competition in the space, you could get into a pricing war, and really no one wins from that. So it's better rather than engaging head-on with a competitor to try to differentiate your product. Pick a niche, focus on certain integrations or certain industries, and show how your product is dramatically better and really in a league of its own compared to competitors.

And that's it; that's how to price. I'd love any questions in the comments below. Thanks for listening.

[Music]