Monetizing AI: How To Roll Out Major Price Changes Correctly

Surprise, surprise, pricing is a huge topic for basically every startup. I'm just astounded at how much push is necessary to get founders to kind of just pull this lever that will obviously make them so much money. That's kind of lesson number 1 I want to share is just around priority, making sure that you're not accidentally deprioritizing something that could be your biggest growth lever just because you're not used to changing it. This one change alone was the single most impactful revenue driver last year in terms of year-over-year growth. Why didn't we do it sooner? Again, prioritization and kind of like an organizational collective blind spot around making it a priority. My name is Gaurav. I am a startup advisor. I specialize in growth. I'm part of the founding team at Superhuman. So I spent best part of the last 10 years building Superhuman out with my brother who is Rahul Vohra, some folks may know him in the tech community. Over the last couple of years, I've started advising startups. Some of the most notable ones include Clay, Replit, Whisperflow, and a host of other ones. So I work with about 10 different startups at a time. What I get to see through that work is a really broad spectrum of what's going on in startups kind of day in, day out, and a lot of cross-pollination of ideas, challenges, and concepts. And surprise, surprise, pricing is a huge topic for basically every startup. I personally have a ton of experience with this from before Superhuman, where I was a consultant, a strategy consultant, did a lot of pricing work for our clients. At Superhuman, I led all of the pricing thinking, both at the very beginning and then quite a bit more recently. We rolled out a major pricing update that was pretty exceptional for business growth. And at probably about 3 or 4 of my clients right now, I am shepherding them or maybe just herding them through pricing changes. And in the clients where it's landed, it's been exceptional. In the clients where it's yet to land, it will be exceptional. I'm confident of it. And I'm just astounded at how much push is necessary to get founders to kind of just pull this lever that will obviously make them so much money. I think about pricing like all day, every day because I price myself constantly. I probably pitch myself about 5 times a week in different ways, shape, and form. And so I'm very much always thinking about what does it mean to add value? What does it mean to then capture some of the value that you create? So it's a topic that's always on my mind and I'm super excited to be in this format because I hope that you can grill me, ask some great questions, and I can share some of these hard-earned lessons with everyone.

I've been a Superhuman customer since 2018, and I don't think you made a pricing change until 2024.

We left it far too late, which is— but you're correct.

And now I think I'm paying you $40 a month. I was paying you $30 a month.

Yeah, that's absolutely right.

Okay, why did you wait 6 years to make the change? And what was the change? You know, there's a $10 change for me, but tell us about the change more broadly.

Yep, okay, so The original price point, actually, to go deep into the history books, was $29. And then we had an intern join us. His name is Cam, and he was like, this is annoying, I'm making it $30. And everyone was like, sure, you just made the business 3% more valuable.

Great.

Thanks, Cam. So that was $30 from 2017 until 2024, like you said. There were some hidden prices, like $10 a month for students and charities. A different price for enterprises that was kind of behind closed doors. But the thing on the pricing page was $30 for the longest time, and the product just got better and better, right? Like to the point of adding value and creating value. Everything from collaborative features that teams can use to AI to sales features like integrations with HubSpot and Salesforce. So much value was added in. We never increased the price. We never updated the price. Obviously, inflation also happened over the course of 6 years, and that's yet another reason why the price could have gone up, should have gone up, but ultimately we didn't prioritize it. And I think that's kind of lesson number 1 I want to share is just around priority, making sure that you're not accidentally deprioritizing something that could be your biggest growth lever just because you're not used to changing it. You're not used to the code or the ways of thinking around it. One of the things that I think is bad advice, like very bad advice out there in the pricing world, is this concept that when you do pricing, you need to assemble kind of like a roundtable of pricing people across the company or, you know, a campfire group. Or I've heard all sorts of BS framings around this where you need someone from product and someone from engineering and someone from marketing and the CEO. That's nonsense. That's going to slow you down. Assembling a team of cross-functional people like that with a ton of different priorities is the slowest way to execute on a pricing change. There's probably 1 or 2 people who are very close to the problem. And in Superhuman's case, it was me as the head of growth and Alex as the head of product marketing who really just know what they need to do to go and make a good pricing change. So anyway, back to the change we made. The product got better. We therefore shifted everyone on $30 up to $40 a month. It was a forced shift. This is maybe unconventional. We'll talk more about why and what we considered and how we decided this. We introduced a backfill price at 30, which was a weaker version of the product, taking out maybe like 2 or 3 years worth of innovation in that SKU. And we added the enterprise SKU more deliberately on top to frame 40 as a great deal. So 40 came out with 1 or 2 features, I think, at announce time. And then we gave customers the choice to basically just go with it from 30 to 40, like, like you went, or to respond, say that they wanted, to stick with the old price but a less good product. And I think in the mix as well, there were off, uh, options around moving to annual for a bit of a discount as well. We can talk a little bit more about some of those details because they matter, uh, in the context of each product. But at a high level, this one change alone was the single most impactful revenue driver last year in terms of year-over-year growth, both as a one-off shift because we just made tons of revenue from the migration and as a trajectory change because the introduction of a higher price SKU results in kind of faster growing revenue than before because there's now more people electing into that price, you know, from the get-go. There's now more teams that are growing on that price with a seat-based model. And so, you know, it was both a one-off jump and also a rate change. Maybe the second part of your question is like, why didn't we do it sooner? Again, prioritization and kind of like an organizational collective blind spot around making it a priority.

I'm curious on two points. How did you determine the price point lift? And how did you communicate the forced change?

Great questions. Okay, I'm smiling because as someone who's been with the company since day zero, You know, in my head I was like, it's going to be €40. Like, I just knew that from the start of the project. And I was hella busy with a bunch of other projects and Alex was kind of like driving a lot of the execution. What actually happened is we brought in a pricing expert who went and did a whole bunch of conjoint analysis and Van Westendorp type surveys to go and like speak to a bunch of users, get a bunch of qual and quant data, present a bunch of options, figure out what features need to be where. And I was like, I was like week over week seeing this kind of unfold for— took about 3 to 4 months, but I was patient about it. I'm like, fine, whatever. This is like not my, not my thing to just like force. And surprise, surprise, it came back as like $40 a month. And it's like, okay, but why did it come back as $40? Well, it's a combination of what is the upper limit, the reasonable upper limit of what customers are willing to pay. In our case, we're a premium software. We want to be at that upper end. We're sort of more, um, Buyrite, less Walmart. And that is a philosophy that's been with the company in terms of pricing ever since the beginning. And, you know, we wanted it to be something that kind of comfortably straddles the line between something that individual can pay for, because much of Superhuman's growth starts with an individual and then turns into teams, but also is, is relatively high for teams. I am pretty confident that there is headroom in that number. I think it could be $45,000, maybe even $55,000. Especially for price-insensitive companies that just don't care. And that then maybe gets more into the enterprise pricing and so on and so forth. But anyway, there was a long version of what we did. And also, I think, I think at most companies there is a much faster version, which is someone is kind of like the benevolent dictator who just decides, okay, and just gets it done. Now, this works, I think, in companies where the customer is relatively price insensitive, usually businesses. We saw earlier a whole bunch of snafus where companies like Cursor and, you know, I think it happened with Spotify recently as well. They are B2C. They have very price sensitive customers. You don't just want to shoot from the hip and make a pricing change. Second part of your question, how did we communicate the change to customers? In our case, we gave them 3 months notice. I think it was an August email for a November change. And we said, this is happening. You're on X, you will be on Y. In the case of Teams, like X is, you know, 10 people times 30, it will be 10 people times 40. And hey, good news, like we're not charging you today. We're not charging you historically. You've been getting all of this functionality for the longest time. This is like a gift. The framing is really important here. And on this date, it will shift upwards. You have some options. You can pretty much just like go with it or you can respond and we can talk about what a downgrade looks like to the, tier down. And for those customers who are on monthly, this, this sentence is missed for customers on annual. You can also actually lock in a lower price by going to annual, which works out for a $40 price point at close to the $30 amount. So they feel like they're paying the same amount-ish. And then for that exchange, you get people on annual billing. You know, you then have to like reason with oneself as to whether that discount is worth it. Because you've basically just given away some dollar value there. You may not necessarily want to include that line. So I would always say, think deeply about whether you want to include that offer. But it just gives customers the feeling like there are a lot of paths forwards for them without them feeling like, wait, my price is just going up. What's my recourse here? We didn't do any in-app notifications. There were no banners. There was nothing that was like warning customers. This is something that comes up with clients I advise. I'm like, don't waste your engineering dollars on putting notifications in the app. Just send emails. It's fine if they miss the email. That's on them, candidly, right? Like, things like receipts and like privacy breaches and stuff, you just send emails, right? So billing changes should be no different. And we did not send that many reminders. I think if anything, we might have sent one email at best, maybe one month out, and then just did it, right? And we braced for impact. We're like, wow, what's going to happen? Are we going to churn a bunch of our user base out? Like, is this going to be terrible? And it was astoundingly quiet. You know, a couple of questions, a couple of support inquiries, maybe a 1 or 2% above baseline sort of spike in churn that lasted maybe 60 days or 90 days, but really nothing that was alarming. Like, we were over the sort of like scariest part of it within a week or two. Just a couple of questions and a couple of customers who were frankly already unhappy, who were like, fine, I'm churning.. But for the most part, just a ton of revenue streaming in.

I have one more question about the Superhuman story. Does anybody else have questions about Superhuman?

I'm curious, when Simpuri Huvis came and just started charging a category which was historically never charged, how did you decide even on that $29, which to me was like a little bit higher when I'm thinking. How did you decide like, oh, people will get $29?

So that is a, that's a very specific lesson, I think, in a situation where you're introducing a much better product to a world where there is a much worse and free or cheap alternative, which in AI often is not the case. It is often the case that there's a lot of greenfield, kind of new, you know, blue sky sort of like area to deploy the product and everyone's figuring it out at the same time. Sometimes those situations resolve to a race to the bottom where everything has to be given away for free because that's what OpenAI is doing, or that's what, you know, Lovable and Replit are doing. So take this lesson with a grain of salt because I don't know how many people out there, if you're working on a product that has a free crap alternative. But in that case, we decided that the wedge we wanted to attack was kind of like the Tesla or the Apple approach. It's the much better version that will cost more money to begin with than the free alternative. And so that basically anchored us to this idea of being at the upper limit of what people are willing to pay for the product. The Van Westendorp questions, this is 4 questions, something to the tune of, you know, the one that we anchored really the most strongly around was, at what price point would you still buy but it feels like it's expensive? That's the kind of like, benchmark that we really focused in on. And we had to segment quite aggressively because operators and ICs were telling us $15 to $20, VCs were telling us $45, and founder CEOs were telling us $30. This was when we ran those questions in 2017. So we went with the $29, should have been $30, but basically anchored around the founders and CEOs. And we knew that we were probably leaving money on the table for VCs. They could probably be paying 85 or even like 100 and just wouldn't care. But we wanted something that just was a simple price like 30. And we definitely knew that we were also going to exclude some operators because they were only willing to pay 15 or 20. But that is then resolved later on with team billing and operators not having to pay themselves. And then of course, I still think there is money on the table for the upper end of the market.

So when you guys were doing this, I had the opportunity to speak with Alex. Yeah. Did you work with Alex?

He was running product marketing for Superhost.

Document.

And he said something that was really memorable to me, which was— I think I asked something to the effect of, how did you persuade the business that this was the right direction? And he said something to the effect of, well, I first— sort of, Garvin and I first isolated on a set of principles around how we would price. And then we articulated those principles to the business. And that made all the detailed decisions a lot easier. Can you react to that and talk a little bit about it?

I think Alex has a much better recollection than I do of what those principles were, because I think they mattered a lot more to his success than to mine. But I can articulate the principles that I share with my clients when we talk about the exact same thing, which is that we need to be customer first. So we need to think about what is genuinely actually keeping the customer's best interests at heart. So we're not just increasing pricing, we're also thinking about the product value. There has to be some concept of the value that is being received by the customer. If you don't start there, then it's very easy to make a bunch of mistakes downstream. So start there. Then there is business value maximizing. We're talking about capturing value. Pricing ultimately is capturing some percentage of the value that you've created in the world. The extreme form of this statement run in the opposite direction is, why isn't everything free? So essentially, that's all it is. Those are the two principles. You could argue a third principle is to keep things simple and to not burn a load of time and especially engineering brain space on the problem. So what is kind of an Occam's razor way of approaching many of these questions, not overcomplicating things. At least when I talk to clients about pricing changes, and for context, like, QONDO is Superhuman for LinkedIn, a client I advise recently rolled out a similar pricing change. Whisperflow is, is one that I advise, and we're thinking about what we can do to evolve the pricing model there. These are the principles that we're talking about.

I got 3 questions. All right. We gotta hold them back, right? No, your story is warm. I love the fact that you were able to make the change. You had the data, the support, the guts to go out there, make the change, make the structure beneficial. Like fantastic. So I have 3 questions. First one, oddly specific. All right.

I love an oddly specific question.

Do you remember what your NPS or CSAT was when you made the pricing change to $40?

We've been hovering around 50-ish on NPS for a long time. Which I will say is like kind of lower than I want, but there's also some segmentation to be done there. Right? Like I think for the customers who love it, it's, it's way higher, but there's a lot of people who are kind of tire kickers who answer that survey. Is the next question, what did it go to?

My second question is around that annual, that annual plan. You said it was discounted to 30. Did you get a lot of takers to that?

Not very high. It was like a 5% take rate at best of those offered. Yeah, it was really low. So, if you're stressed about the cannibalization there or the revenue loss, at the end of the day, the take rate's going to be sub-10% probably. So, you could even just allocate the amount of brain space to it according to that.

Last question. Would you churn at 45?

That's a great question. Uh, it's not that hard. Don't think about it too hard. I think I told Alex that I was like, 40 was pressing. Do you pay personally or is Schematic paying?

I pay personally.

Well, let's fix that. Let's fix that first of all. Let's get you on a company credit card. And Gio, you can also have an account.

Yeah, we use Teams here.

Yeah, there's a Teams account that one person should not have to stress out about the money. And you know this, right? Like, the value of your time in the business context is— whatever, I'm doing the sales pitch right here, but 45 times 3 or whatever.

Sort of a team's upsell.

I'll give you the commission later. I've got one last question. Anybody else?

When you guys were doing the strategy, you said it took 3 or 4 months, which is mind-boggling, but fine. When you were rolling it out, you had given this option of like a downgrade option, right? Talk to us, talk to me about what made you do that, meaning were you concerned that there would be a huge churn risk? Is that why you guys thought of a downgrade risk? And how many people actually went downgrade?

Um, again, it was sub-10%. It was, it was in the region of like 5%. So it was a really high percentage just went with the flow. In the case of Superhuman, there had been a lot of accrued product value that made sense for the middle tier. It's not like we were launching a new SKU with new features and pricing that higher and moving everyone into it. So I think before you get into the price change, you kind of have to take stock of where you are in terms of the product as it exists today. What is the accrued product value? What's about to ship? Is there like something that just eclipses the current product value that you're about to ship? If so, go and price it at way higher. See also OpenAI's $200 a month price point, right? Like they came out with that because it just eclipses the existing thing. And so that's very case-by-case specific to the product and the speed with which you're releasing value. In our case, it was, hey, in order to run this price change, this is mostly correcting kind of a built-up pressure. And that pressure comes from the product being so much damn better than it ever was when we originally priced it at this level. Was there a second part to the question?

No, that's it. The down period option was my biggest thing because it sounded like you had a backup plan in your pocket when you went live.

Yes, exactly. Yeah, so in the context of the product landscape. We wanted to give customers who felt price sensitive the option to peel off and go back to the lower SKU by not having features that one might argue are power features they don't want.

And if you were to do it again, would you do that?

It would depend where the product is, but if we were in exactly the same position, then yes, I would do the exact same thing because it is the cheapest for engineering. You're not arbitrarily finding new features to put into the higher tier. It's astoundingly easy to do it this way.

Did you do any tests?

We just rolled it out. There was a lot of discussion though, as per the previous bit around, you know, talking to customers, some, you know, decently high number of customer conversations across individuals and teams to make sure we kind of understood what that real willingness to pay looks like. But I think unless you are at the scale of like a Grammarly with that volume of people kind of hitting the pricing page and going through different pathways, it's it's hard to run, I think, real A/B tests on this sort of thing because the power, you don't get to significance across your different variants in any sort of reasonable timeframe.

So, maybe a slightly different question. I'm just curious, for a product like Superhuman, when you're selling to teams who might be paying for Gmail or Outlook or whatever that might be, what's the pitch to get them to pay for a premium email product?

So, Most, most people don't think about the money they're paying for Outlook or Gmail, right? That's one of the first things that a founder signs up for. And your Google Workspace bill is like deeply buried in some, like, you know, it's probably still on the founder's personal credit card. So, so most people don't think about that piece. So really the question is like, how do you get someone to pay for upgrade software or premium software? It doesn't really matter if they're already paying for something kind of foundational, I think.. And there are a number of examples like this, like Loom or maybe Otter or, you know, I don't mean Whisperflow. These are all productivity tools that help kind of the workforce be better at what they do. A massive part of it is individuals feeling it and loving it themselves. So they get that just visceral feeling that it makes sense to them. And they're like, well, if I feel it, then I'm going to— my company will feel it. It's very helpful when you have a champion or evangelist who who gets that, because then they just do the work of bringing everyone around them on board. But when you get into larger teams, you need to have some proof points, right? There's a buyer there whose neck is on the line for the results of their team. In our case, it's the head of sales or the VP revenue, CRO. They have a sales force that needs to hit goals. The sales force is more effective when they're better at email. So we can then show that VP revenue or CRO usage stats in terms of responsiveness, emails responded to before and afters. And obviously, you know, any number of other metrics we'd like to pull out of our systems. And we can show them that there's actually ROI. So you just like then turn it into a, you know, a time saved calculator or alternatively a revenue generated calculator if they're revenue generating end users. And that's actually consistent with what I see at a lot of these PLG companies where the, you know, the product helps someone do their job better and save a bunch of time.

I think you articulated the communication aspect really well, but when you think about timing and packaging, and maybe this isn't apparent to Superhuman, maybe to the other folks you've worked with, when you're dealing with large, you know, year-over-year contracts, for timing, are you looking here at changing pricing? Are you looking at mid-contract renewals? How does that factor in efforts on different schedules?

Yeah, so timing, timing is like a weird one because if you think about it, people on month to month, you have the same problem there, right? Like, which is that there is, you could be one day into the month with 29 days to go and like you want to capture the value sooner, right? Like we did a November 1st rollout or something. There's a bunch of people on monthly who are kind of mid-cycle in air quotes, but it's just monthly as opposed to yearly. As someone who likes to think about problems in a very mathematical abstract way, I kind of just want to solve the problem for both those cases at the same time, annual and monthly. And if you find the solution, it conveniently solves for your quarterly and your multi-year people too. So that was like my nerd way of saying like, there is only one problem to solve, which is how do you get everyone onto the new thing? My actual answer, which is a bit more useful, is I think people stress out a lot about putting an invoice out to customers on annual. Probably they stress out too much on average. And my recommendation is to just cut an invoice mid-cycle to get them onto the right price points. If you're doing PLG in particular right, and even if it's more of a sales-led motion with, you know, some expansion built in, you should be cutting invoices in, you know, intra-cycle. You should get your customers used to the idea that they pay more in the middle of that annual contract if there's some reason for it, like some justification. That could be usage-based or it could be some improvement to the product. So if everything has been done correctly up until that point, someone on an annual contract should expect to receive an invoice. Maybe it's every— it could be as frequent as every month or, you know, maybe for logistics it's easier if it's every quarter with some justifiable reason as to why they're receiving that. And I think in the cases where people are afraid of it, it's because they fought so hard to get this annual contract in and now they're just terrified of churning the customer. So they don't want to do anything to charge them any more money. Those are different problems that kind of need direct solutions. So I took a look, Clay's got a free tier.

Yeah.

Well, it's optically a free tier. It's actually a free trial.

Ah, it's actually a free trial.

Okay.

Superhuman doesn't.

As far as I could tell.

Talk to me, so I assume you had something to do with both of those decisions. Talk to me about the thought process there and why you wouldn't want one with Superhuman, or was it simply the timing and when it came out?

And did you consider a free tier when you moved to that $30 lower feature price point? Yeah, so this again goes back to the idea of releasing a premium product in a space where there's already a free alternative. I think Putting a free product out there is important when there is high competitive pressure for the thing that you're selling. So the reason that, let's say, OpenAI gives so much away in ChatGPT for free, yes, they have the mission, which is, you know, AGI benefits all of humanity. That's, that's kind of like the, the overarching force. But if you just look at it in a competitive landscape, if they don't do it, Anthropic will, right? And then they'll lose their lunch to a competitor. In the case of Superhuman, we didn't have a competitor to worry about. In that case, we were deliberately carving out a separate space, giving away our own product for free, basically as a self-own, because it's like, why would we do that? The whole, the whole reason to exist is to have something better and charge money for it. Same with Tesla, same with Apple products, right? It's like, it's a top end of the market where there is something that you can use that's much worse, that's maybe cheap or free. Clay's example is really interesting. They are also in this position where if they don't give something away for free, someone else is going to do it. There are a lot of Clay competitors and alternatives. So they have to defend that market share right now. But it's interesting. It's not a free tier. It is a free trial, which gives them these 2,000 credits. And then once that usage is consumed, you have to pay. There may be also a time element in there. I need to go and double check. But they're doing it to basically kind of pointy elbows all the other players a little bit, but not so much to the extent of having a full-blown free tier. And the reason to have a free tier entirely, again, competitive pressure is a huge part of it. And if it's not competitive pressure, there may well be some network or compounding advantage that you— network effects or compounding advantage you're trying to accrue. So that could be a bunch of data you need. It could be that the product gets more valuable when there are more people using it. It could be that you want lock-in within an organization. So, there's a few other reasons to explore for whether it makes sense to have a free tier entirely that are very product-centric. But in the case of Superhuman and with Clay, neither of those things are actually true. There's no big data advantage or massive one, and there's no big network effects necessary. So, it's more about the competitive landscape. Great question. Can I proselytize about one quick thing?

Okay.

So, this is a mistake that I see frequently being made, and I would be a very happy person if it was basically never made. Actually, there's two, and they're kind of related to each other. So one is the idea that you have an individual SKU and a team SKU in a product-led motion. This is the most bonkers thing I've ever seen in my entire life. I don't understand why a company would ever do this, because if you have the capability for teams to use your product, it makes all the sense in the world for that to be available to the lowest price point to get groups of people using the product together. And in doing so, you then access logos, you access teams that can hopefully grow on their own over time, and your growth will steadily compound without you having to exert effort. If there's an individual to team kind of hop in price, you've introduced a lot of friction to your own growth. So that's like a self-own again. I see a lot of startups make this mistake. And then related to that is lowering the price on the SKU above for teams on a per-seat basis as opposed to increasing the price. If you buy into this idea that every SKU should be team capable, then the SKUs that are to the right-hand side of the leftmost one should just have a higher price point, whether that's per seat, whether that's for the whole product, potentially even if it's for credits. We could talk a bit maybe about whether it makes sense to discount, so like Klei discounts for more credits. But this is a huge mistake that I see a lot of companies make. And I think it comes from a place of fearing what customers might think if they move to the Teams plan and they're like, well, what am I getting for Teams? And like, you know, why should I pay more per seat? And I think founders just need the confidence and growth and marketing and product people all just need the confidence to know that something as simple as shared billing is the value that you get on Teams. And that's actually something you can charge for. So this is something I see a lot. You know, there's more startups in the world than I have time to talk to, to share that idea to. So I would like it if more startups did not make those mistakes.

Via LinkedIn, Chris, is anybody making that mistake today?

Okay, go fix it.