Pricing Page unPacked - LINEAR: Linear's AI at cost, and the ceiling on seat pricing

Rob Litterst (0:14)

Welcome back to Pricing Page unPacked. I'm Rob Litterst, and I am joined by Ulrik Lehrskov-Schmidt, pricing expert and CEO at Willingness to Pay. Each week, we take a real company's pricing page and break it down. The decisions behind it, the trade-offs, and what it tells us about how the company actually wants to grow. No slides, no scripts, just a real conversation between two friends who live and breathe pricing. Let's dive in.

Ulrik Lehrskov-Schmidt (0:47)

How's it going, Ulrik? You're good.

Rob Litterst (0:48)

Awesome. Linear today, right? We do. Yeah, I'm excited to dive into Linear. Those really hot products that I think have like a really good design ethos. Like I think their, CEO actually used to be like a designer or like head of design at Airbnb. I think his name is Karri. I'm probably butchering his name, but Karri Saarinen. And yeah, they have like a great design ethos in general. So while I don't use linear, I feel like I've seen them quite a bit and really appreciate what they're doing from a branding perspective.

Ulrik Lehrskov-Schmidt (1:24)

I must admit, I know nothing about them. So this is going to be really interesting just from like a truly sort of outside-in perspective. Sometimes something happens and you just kind of miss it. And I kind of just missed Linear. So excited to have a look at it.

Rob Litterst (1:38)

Yeah, well, I think it makes sense because, you know, you're running Willingness to Pay, which is a services business, and Linear is really built for, like, software products and kind of like product businesses in general. So the kind of like frame of Linear, are basically the anti-Jira product that is kind of trying to build like the bug tracking and product, you know, product building tool for the agentic future, right? So one of the craziest things that's happening there is agent-created work went from around 3% to 50% of all work on the platform in the last year. And Linear has responded by charging zero markup on AI, which is very interesting. They were founded in 2019. I mentioned their CEO used to be an Airbnb designer. They are three Finnish-born founders who basically built the tool that they wished that they had. As far as positioning goes, they're fast, opinionated, issue-tracking. They've expanded into initiatives, roadmapping, customer request intake, insights, and native AI. Issues, sub-issues, cycles, projects, triage, routing, basically could use for building a product, understanding what's working, what's not, and fixing any issues with that product. They have about 40,000 paying customers as of August of 2026, up from 15,000 in June of 2025. So they're growing pretty fast. Their customers include OpenAI, Salesforce, Figma, Coinbase, Ramp, Cursor, Cognition, Harvey, Scale AI, Perplexity, Baseten. As you're probably thinking, they are the kind of bug tracking, issue tracking solution for this kind of like AI native era. And while they have very simple pricing, which I will get into, their net revenue retention is 177% as of this year, up from 145% in late 2025. This is Linear's own disclosure. And agents are installed in 95% of their paid workspaces. So they are definitely seeing engagement with AI. From a money standpoint, they passed $100 million in ARR in mid-2026, which is amazing, but compared to some of the kind of AI-native products, it's kind of like a sluggish growth story, which is kind of insane. They've been cash flow positive since 2021, and they state that they hold more cash than they have raised. As of August of 2026, they have 160 employees, so a pretty efficient team.

Ulrik Lehrskov-Schmidt (4:10)

Revenue sounds low. 40,000 paying customers, and then they had, what was it like 120 something million, 134? It said 100 million, yeah. Yeah, so that's 3,300 or so on average, right? That's 250 a month. Like, that's the ACV. So when we get to sort of see that we probably have out of 40,000 paying customers, We're probably going to see some version of like an extreme long tail. And then I think the really interesting thing is going to be like, how good are they? Is their pricing model and monetizing like the tip, not in the long tail tip, but as like the peak of the hockey stick, like the really valuable customers. Like, is there a way where I could pay linear like 10 million a year?

Rob Litterst (5:02)

It's kind of, it's funny because it's like we just on the last episode, we're talking about Notion. And I do feel like while they do very different things, I'm sure some people use Notion for some of this stuff, but I do feel like there's kind of like a similar dynamic where their core pricing is very accessible, very affordable. They don't really have the mechanism built in as it stands to scale to like their top 1% customer, right? And I think that's kind of where both companies need to go. Okay, so we are looking at the linear pricing page right now. And as I mentioned earlier, they have a very minimalist design, very clean pricing page. Four plans, free, basic, business, and enterprise. Free is obviously free. Basic is $10 per user per month, billed annually. Business, $16 per user per month, billed annually. Enterprise is custom, and it's annual only. As far as differentiation, they use features, obviously. Then they have this concept of differentiates free and basic and business. And then they have issues, which differentiates free and basic. But they offer their agent platform for everybody and they offer linear agent for everybody. What are your thoughts just looking at the kind of like high-level pricing and differentiation from their plans?

Ulrik Lehrskov-Schmidt (6:25)

So the first thing I noticed was that at business, which is so we have four tiers, right? So it's like zero, 10 per user per month, 16 per user per month, and then enterprises custom. So it's one of these where, so I don't know how many users you would have on business. Let's say it's like really crazy and you have like 100 users, right? So then it would be 1,600 per month, which is like translates to less than 20K a year. And then the only step after that is enterprise. It's one of these where I'm pretty sure that the enterprise plan is like low five figures on average. So just looking at this is not built to monetize well. Like, it just isn't. Which is fine. So they have a lean team. They found a way to sort of, like, build up some momentum. Great for them. Realistically, the question is whether they should just ride this and become, like, the de facto, let's say, system for agentic AI, which would mean, hey, let's just, like, not monetize essentially. But they're going to have an extremely hard time figuring out how to charge like 3, four, 600k, a million, two, three million. Like some of these logos they have, if they're all in on using them as the system, they wouldn't have a problem paying high seven or even eight figures for like the right product. Especially if it was able to handle high volume AI loops around the product, which is I think really what the use case is.

Rob Litterst (7:54)

They're very good at just like surfacing only like the things that you need to see on the top line kind of like pricing header, right? So as you scroll down, you can start to see some of the other ways that they're differentiating their plans. Like they have a storage limit on free, which makes a lot of sense. Then they have kind of like their core product, which the only other real differentiator that we see from their core product is this idea of pipelines, which goes with their future releases. Release. So this is something that they actually launched, I think, somewhat recently and has kind of pushed them from just a pure issues tracker towards more of like a full-blown product lifecycle application. So they're really kind of building more breadth into the product, which is really cool. Then they have AI and agent workflows. So they have their agent platform, which doesn't have any kind of like usage mechanics included in the pricing page. MCP access, the linear agent, coding sessions, let linear agent write code and create pull requests, and that requires AI credits. Okay, so they do have AI credits that we need to dive into. Loops, which is their automated agent workflows, that also requires AI credits. They have code intelligence, triage intelligence, and then they have some other features around kind of like integrations, team management, analytics and reporting, linear asks.

Ulrik Lehrskov-Schmidt (9:16)

I think one of the other things that I noticed sort of as you just like went over the feature set here is that I think they've done a relatively good job at like differentiating features across four plans. I think you can have people in a product like this thing, oh, like the product is the same for everyone. Yeah, until you decide what actually differentiates like the more sophisticated or demanding customers from everyone else. So one is essentially security and compliance, like advanced roles and permissions, different workspaces, like who gets to say yes to what, like all these things. I think, especially in an agentic infrastructure, having checks and balances for like what agent can override what, you know, core code, I think is a really like key thing. And so allowing this in a way more granular way, I think is a great thing. So they're doing that and then they have some like pipelines, like production speed, like how many highways can I push out? And I think they're doing this. I'm also noticing that actually a lot of the different types of features that would normally like promote consumption, like Slack integration access and all these things, they've also like pushed to like the end of their feature stack. So I think if I were to just like after all of 20 seconds of looking at this, figuring out how to actually create more growth, I would enable more interaction with the system. So email intake, Slack intake, web forms, all these things, I would just like give it to everyone. And then the control of that, like controlling the mess, then becomes paid. Yes. So I think they have some like adoption friction that is unnecessary right now, which I would definitely like pull forward in the product. Just like, hey, let everybody get in with all of their whatever tickets and box and in any way, shape and form possible. And then when we're now cleaning it up and we're triaging it and making like sense of it and delegating and making sure that not everyone has access to everything and can do everything, that's then what gets monetized. I think there's some inherent product philosophy and growth philosophy where I think they have a little bit of a too simplistic view of the form, hey, this is an advanced feature that is valuable, therefore people should pay for it. As opposed to thinking, this is an advanced feature and it is valuable, would it create faster adoption for us And then downstream create something else that we could then monetize for. Yeah. And I don't think that they sort of traded that needle yet.

Rob Litterst (11:48)

Yeah, I think you're right. And I feel like a lot of SaaS companies kind of like are stuck in that mindset of like, I don't know exactly what the timeframe is, but you know, like premium integrations being a feature that you kind of like charge more for or differentiate a higher tier with. And I think like a lot of that goes back to this idea of kind of like license-based pricing where there isn't really a usage lever, right, for pricing where you're not metering any of the work. And I think like once you introduce a usage meter, anything that can increase usage or is kind of like a usage enabler should be pushed up to your point, right? Like all of those things should be pushed into lower tiers to try to drive and engagement as possible. And then to your point, you charge on kind of like the outputs or the outcomes or the usage and the ability to kind of like wrangle that chaos. I like where your head's at there.

Ulrik Lehrskov-Schmidt (12:46)

It's just like, you know, restricting how many people can go into your store. Yeah. Doesn't make sense, right? Just open the doors.

Rob Litterst (12:56)

What's interesting is like from Linear's pricing page, you would have no idea that like you have to pay for credits unless you like follow through on some of the asterisks and really start analyzing it. There's no mention of credits. It's a fine print. Yeah, it's fine. It's very much fine print. And this happens a lot. A lot of companies are doing We're looking at basically this support doc right now of how AI credits are used. And they have two different features they're using credits. One is coding sessions. So this is available on Basic and Up, so all the paid plans. And it's billed per session. Cost includes model tokens at provider-published rates with no markup. In sandbox runtime at $0.25 per 20-minute block. Feature is loops, which is available again on all the paid plans. And this is built per loop run, typically runs without a coding session cost, $0.07 to $0.20. Costs vary based on the work the loop performs. Runs that start a coding session also incur the applicable coding session costs. And just as a reminder, coding sessions when linear agent writes code and creates pull requests, agent loops when a loop runs agent work in the background. Any immediate thoughts on how they're charging for these features?

Ulrik Lehrskov-Schmidt (14:13)

It is a pure flow-through, right? Yeah. They have a very small markup. For a loop run, they're saying, hey, we're going to charge you $0.07 to $0.20, typically. That sounds like we're going to measure how much it costs, and then we're going to not lose money on this. We might add like a little bit, or maybe even not. Like maybe it's just like pure pass-through. The same thing actually with the coding sessions, which is like cost includes model tokens at provider-published rates, so that's like no markup. And then sandbox runtime, which is like, oh, that's actually like a second piece of infrastructure. That's not an LLM, that's like a sandbox infrastructure at this rate. So it's almost like you're saying, hey, we have these like two parts of the pricing architecture. We have the core functionality, which is like whatever, the product ticketing system that we have here. That's priced, value priced, margin profitability priced at whatever, 10, 16 bucks per user. Right. And then we have this massive infrastructure layer underneath where the things just flow through. And we are betting that will increase if we don't try to like skim from the top. So I think they're doing much as they can to say You that it's not going to be cheaper if you do it yourself. This is like the bare bones, actual direct cost just flowing through the system. And I think if you just want to have adoption and footprint, it's not a bad way to do it. The even more aggressive way would be to say this is totally free. Is going to cost us money, but like that's on us and then go and do it. That would even create this scenario where developers would say, well, then it's not going to stay free forever. And you're venture-backed, I'm not going to trust you with, you know, all of my infrastructure and habits and like lean on you as a provider. So you can almost say that this, what they have right now is for certain types of user groups, like the most like adoption focused approach where they say exactly the like the back-end cost or what you're going to pay another dime more and then we're going to take this like very small fee over here and that's the distance because it is explainable but it's also like long-term sustainable so if I'm a developer and I'm I don't like to pay for anything which many developers don't like I can sort of make a case that this is how it's going to be also in like five and ten years from now I'm just looking through how they actually manage credits.

Rob Litterst (16:52)

So you can add credits ad hoc. You can top up in a minimum of $10. There's the ability to automatically reload, which is something that you can manage in your credit settings. Funds you add expire 12 months after the purchase date. Unused balance rolls over across subscription billing cycles. It is independent of your linear subscription period.

Ulrik Lehrskov-Schmidt (17:12)

Very interesting, yeah?

Rob Litterst (17:13)

Yeah, so another thing is like they show your credits in dollar value. And they are saying that any funds that you add expire 12 months after the purchase date. Promotional balances will have their own expiration dates. So they're probably doing some kind of like giveaways and stuff like that to drive adoption. Then, as we just discussed, unused balance rolls over across subscription billing cycles. It is independent of your linear subscription period. So credits are kind of handled as this totally separate thing.

Ulrik Lehrskov-Schmidt (17:42)

So you can have dollar denominated credits in scenarios like this where you are saying we are just going to, we're never going to offer your discount it's not like you can buy you're like a million credits for million like half a dollars so you're just like essentially you're just depositing money into the account, and then they're going to be drawn down based on whatever the actual lms cost in and the background then the only thing that makes this not actual a bank where you hold customers money is that they say you can never get it back so that means that I'm not holding value I'm not holding dollar I'm not holding dollar value in that sense. Right. And I can't transfer it to other users. And so there are a few rules, like if they were starting to say, hey, you can give it to your friends, you can pull it out again, they would definitely be subject to SEC regulation. But they're sort of very carefully saying all the right things to avoid that. And then if they're not doing any markup, and then that also means that they have a very defensible position to not offer any discounts ever. So they're saying, like, if I offer you a discount, I am losing money. Why would I do that? And people are like, okay, sure, like, I get it. And then discussion is now over, right? If they don't have to offer any discounts, they also don't get into like, you can get weird revenue recognition issues when you sitting in your account you only collected $80 from your customer because you gave a discount. But since they don't do any of that, they don't have any of those issues. And then they essentially just say, hey, we just want to have the adoption and the stickiness of the product, so we're not going to let them expire, at least as long as you stay a customer, right? Which is also interesting. Right now, it's not really a monetization model. It's more a, let's cover the cost while we do the adoption footprint expand model, which is fine. So I think the interesting thing is if they have a clear idea about what the event horizon is, like at what point else, if ever?

Rob Litterst (19:40)

No, that makes a lot of sense. And I mean, I think you're right. I think it's Like, let's bring, let's try to bring more AI usage into linear and as little friction as possible to doing that.

Ulrik Lehrskov-Schmidt (19:56)

Guys, new book, The AI Pricing Roadmap, How to Make AI the Most Profitable Thing You've Ever Sold comes out in a few weeks on December 1st and it's available on Amazon and everywhere else. In the book, I go over how you can design AI pricing that makes the customer want to pay for the value of your product, how you can launch new AI with maximum adoption and minimal risk, and how you can control costs and engineer your margins. Because this is what we have been doing for all of our clients for the last couple of years. So, go buy The AI Pricing Roadmap on Amazon. Hope you like it.

Rob Litterst (20:36)

One question that this raises for me. So they're charging no markup on the models, which is great. Do you have any sense, Ulrik, for like how easy it is to implement like bring your own key? Like in that situation, would it just make sense for them to let people bring their own model and pipe it in so that they can draw from their own subscriptions?

Ulrik Lehrskov-Schmidt (20:55)

So the question is if they are now or planning to later. So let's do sort of a little bit of a weird example. So they have OpenAI as a customer, right? Okay. So is OpenAI paying OpenAI public rates to use OpenAI inside of Linear? Yeah, that's a good question. So you have this sort of weird circle, because OpenAI obviously has a better deal with themselves than everybody else, right? But just to use that as a frame of reference, that might be what they're doing with someone else. They might get a better price than the publicly listed rates. And then markup comes from the, not the kickback, but the reduced rates that they actually have to pay on the back end. This was a very, was, is a very common like cloud reseller motion. I buy cloud from Asia, the AWS or Google, and I resell it to whatever large, you know, manufacturing or public enterprises, whatever it is, at list prices or like slightly reduced. And then I get a much better deal at the back. Like larger banks, you would still have to pay like something of like 90% of like list prices for like AWS and Google Cloud and all the others. But at the back, like the very largest customers of these like super scalers or hyperscalers would maybe get 80% discount. So you could actually sell cloud infrastructure at like a three, 400% markup to the right customers if you were one of the sort of the largest buyers, right? And that has powered a lot of different business models. Maybe that's what they're aiming for? Because everybody still is paying less prices. Nobody has really figured out how to get OpenAI or Claude to really give them a discount because they have so much demand that they don't have to. So everybody's kind of just whatever paying the pricing page is saying they should pay and also changes so much all the time that it's hard to keep track of. But maybe they have a line in to say, hey, we get it at whatever, 80 cents on the dollar, we sell it on for a dollar, which is what everybody else has to pay anyways, and then we make the spread. However, if that's the case, why are they not making more money? That's a good question. And if that's the case, why are they not giving it away for free? So I think it's one of these where it might be a business model, it might be a very small part of the model, it might be a plan they have, it might be something they haven't considered. Don't know. But that model could be there to be picked up if they chose to.

Rob Litterst (23:31)

Ulrik, there's one more angle that I want to touch on before we close this out, which is kind of like the competitive model for Linear versus Jira. We talked about how, like, I think they used to be a lot more aggressive about their positioning towards Jira. Interestingly, like what I think is interesting about Linear is I think more they're actually expensive than Jira for a lot of customers. It's hard to tell with Jira because they have, it depends how many users you have. And I think Jira also can be pretty aggressive with discounting. But I was looking at a little comparison on reddit of jira versus Linear. And one of the things that I found is that people were talking about how Linear is a little bit more expensive. So it sounds to me, yeah, so here's one. This guy who made this actual post. The good, UI is clean, fast, no bloat. The team actually uses it without complaining. The reality check, missing events, reporting that we relied on from Jira. Integrations aren't as mature. Cost jumped 40%. Worth it. Maybe the team velocity feels better, but I'm struggling to prove it with the data. So this guy basically just talking about whether or not it was actually worth it to move from Jira to Linear. Because I feel like in most situations like this where there's an incumbent and you're the challenger, you kind of like enter the market at a lower price, right? And like that's kind of like a big part of the challenger mentality and like the disruption, like the kind of Clay Christensen idea. I'm curious, like what, like if this reminds you of anything else or kind of like what you think about this dynamic in general. To me, like it's kind of like a premium based on like a quality experience. Maybe that's kind of what they're trying to evoke, but I'd love to get your take on that.

Ulrik Lehrskov-Schmidt (25:14)

That's probably it, right? So I think it's a bit like Superhuman coming in and charging money for an email system where it was free before. But because the interface is so much better and people are in email all day, it just makes a ton of sense to pay for it. I think it's a little bit the same with developers are probably in Jira a very small premium for like a better UI, like even developers need UI. So why not pay? I think Jira, because of their wider relationship with the rest of the ecosystem they're part on, have more ways to monetize. I think Jira could also say, hey, we can be sort of a low-cost entry point into an ecosystem of other products where Linear only has their own thing. So it's like they've got to make money on that one thing, right? I think developers are developer tool. We've done a lot of developer tool pricing. And it's sort of a weird category because it's one of the most vocal ones when it comes to pricing. And I think it's a little bit like developers know how the software is made. So like, why would I pay whatever money for software? I can just make it myself. Like this kind of mentality where when then people start, like everybody else that can't make software, we're like, oh, this is great. Love the tool. I get value out of it. The developers have this kind of inside information on what it is, which means that they of always have this sort of build-or-buy approach to most software. And then, so when they buy, they do so reluctantly because they're like, all right, we'll just build it, right? Which makes developer tools some of the harder ones to price. So you kind of have this scale where at the beginning of the packaging structure, you price for the developer, for the user, like zero, free, 10 bucks a month, forever, whatever. Like you have to be super cost-conscious. And also, they talk a lot. They're on communities and like you spread through that. So marketing works a little bit different. And then at the other end of the packaging, at the enterprise level, now you're selling to a buyer that runs a tech organization and has to like do like multiple teams through like a hierarchy. And they work differently. Like they don't care, right? So for them, it's like, oh, I'm paying this much for infrastructure. Other thing. You kind of like the art of doing developer tool pricing is to know when and how to make that switch. Like when do you have to go from like bare bones, like adoption-based community pricing to this is the buyer, they have a budget, they understand the value of the problem, this is where we monetize. And some tools never do that.

Rob Litterst (27:56)

They just stay on the low end forever. And others, they manage to bridge that gap.

Ulrik Lehrskov-Schmidt (28:05)

All right, Ulrik.

Rob Litterst (28:06)

Every episode we do kind of our Wall Street ranking system, buy, sell, hold on a company's pricing strategy. Where do you land with linear? And just to be clear, this is not like for the company. This is just for their pricing strategy. Buy meaning we love their pricing strategy. Hold meaning we like the pricing strategy. Don't like their pricing strategy. Where would you land here with Linear?

Ulrik Lehrskov-Schmidt (28:33)

I think it's a hold. So it's one of these where given the product they have and given the position they have, I think they have a lot of traction. It's hard to argue with 177 on NRR. So something is working. I think it could probably also work a lot better. It is something that they have to do now, and then they have to figure out an endgame of how to really print money at a later stage. It's going to be difficult for them to go into like a billion plus of revenue. They have 40,000 accounts now. Could they have 400,000 accounts? Yeah, but not like that's a lot harder to do. So you kind of need to sort of figure out, okay, how do you really turn on my decision? So I think they're in like an expand phase where they're trying to get the customer base. The way they're doing that works really well. We have sort of yet to see them really sort of figure out the next phase.

Rob Litterst (29:23)

It's like on one side, they are kind of like pricing themselves as a premium product on the license side, referencing that kind of like Reddit thread. They're coming in higher than Jira in some cases. But then on the other side, with AI, they're being extremely generous. They're not taxing anybody's AI usage. So to your point, they do seem like they're kind of in this like middle ground right now where they're still trying to kind of like figure out what to do with AI pricing. And I think once they do that, the growth story could get very different and they could definitely start to monetize a lot better.

Ulrik Lehrskov-Schmidt (29:56)

But I think from what I could tell from the founder story, the three original founders probably still have control and a relatively keen cap table. So yeah, I think they get you right this one out. It's just like, remember that at some point in time, you just want to get free cash flow, right? And a lot of it. Don't postpone that for too long. Like get, if you're bootstrapped or like semi-bootstrapped, don't always think that there's going to be like two, three, five years until you can. Sometimes that day is just today. That is it for this episode of Pricing Page unPacked.

Rob Litterst (30:34)

If it was useful, subscribe to the podcast on your podcast provider of choice or find us on LinkedIn, the PricingSaaS community, or on YouTube.

Ulrik Lehrskov-Schmidt (30:41)

And remember, your pricing page is not just a page, it is a strategy statement. See you on the next Pricing Page unPacked.