B2B SaaS Marketing Snacks is one of the most respected voices in the SaaS industry. It is hosted by two leading marketing and revenue growth experts for software:
B2B SaaS companies move through predictable stages of marketing focus, cost and size (as described in the popular T2D3 book). With people cost being a majority of the cost involved, every hire needs to be well worth the investment!
The best founders, CFOs and COOs in B2B SaaS work at getting the best balance of marketing leadership, strategy and execution to produce the customer and revenue growth they require. Staying flexible and nimble is a key asset in a hard-charging B2B world.
Resources shared in this episode:
Brian Graf: Welcome to episode 85 of B2B SaaS Marketing Snacks. I'm Brian Graf. I'm the CEO of Kalungi, and I'm back with Antoine Vial, one of Kalungi's fractional CMOs who has led marketing for over 20 B2B SaaS companies.
In today's episode, we're circling back to a big topic we touched on in our last AI conversation: pricing. The AI landscape is moving fast, and with 14,000 startups launched in the last five years, pricing has become a critical way to stand out and communicate real value.
Antoine and I unpack the major pricing models we're seeing in the wild – from usage-based to value-based – and we explore which strategies work best at each stage of a company's growth. Why some strategies build credibility while others create churn risk, and how to align your pricing model with your ideal customer profile.
If you're working on pricing for your AI-powered product, or even just looking to future-proof your B2B SaaS go-to-market motion, this episode is packed with useful info.
Brian Graf: Let's get into it. All right, Antoine, welcome back and thank you again for joining me. I wanted to follow up on our previous episode. It was supposed to be about pricing and positioning, but we ended up spending about 90% of it on positioning. I wanted to make sure we came back and talked about pricing a little bit. Just to give a quick recap – actually, let me take a step back.
For those of you who didn't listen to the previous episode: Antoine Vial is a fractional CMO. He's worked with 20-plus companies in the B2B SaaS space and is at the forefront of AI marketing for us at Kalungi. I thought he'd be a great person to have on for these topics.
To recap what we talked through in the previous AI episode: we explored how the AI vertical is both hyper-competitive and also very new at the same time, so it's a really unique landscape for a company to play in. We dug into positioning and pricing for companies in that landscape – how do you navigate this really complex and rapidly changing market dynamic?
We covered that there's been an explosion of AI-powered tools and LLMs have almost become a bit commoditized as quickly as that has happened. Because of that, positioning and messaging become even more critical than they usually are – equally important, if not more important, than technical differentiation.
Being able to communicate value to the right people in the right way is what's going to set your company apart, rather than just being a little bit faster or other small technical wins.
That's the technical battlefield for AI right now. Just to reset the stage (and some of these stats might repeat from last episode): 14,000 AI startups have launched in the last five years – 40% of them in the US.
And beyond just those new companies, it feels like every other company right now has an AI tag, because everyone's incorporating it into their businesses and offerings. So not only is positioning a key differentiator, but now pricing is as well – it's how we communicate our value as effectively as possible to customers.
Pricing is a huge lever to pull when communicating that value.
Brian Graf: Let's dig into pricing a little bit. Maybe we can start with some challenges and mistakes that companies run into when they're building out pricing for the first time in this AI industry.
Antoine Vial: Yeah, of course. Thanks for having me. I'm excited to talk yet again a little bit more about AI. I think one common mistake is assuming that because AI is new, there are entirely new rules – which is only partially true.
If we look back at SaaS as an industry and how SaaS platforms have been sold in the past, you can see different trends. Take very traditional SaaS – CRMs like Salesforce or HubSpot – the pricing models have evolved over time, but the idea was that these SaaS solutions were selling better workflows, improved visibility, collaboration...
It was kind of hard to value the dollar amount, right? At the end of the day, it's always about saving time and money. In those cases it’s more complex to explain that value. Then as we progressed, we saw trends where pricing was reused as a repositioning tool – pricing and packaging became part of the positioning of these solutions.
And then about five-plus years ago, you had a wave of FinTech SaaS companies growing super fast because they focused on the job to be done, and they were very good at expressing that value through their pricing while being super clear about what they were helping to solve.
You can think of companies like Stripe and the like, which grew super fast with very strategic pricing. Now with AI, there's this tendency for your SaaS tool to position itself as "saving you money," which is embodied by the pricing models AI companies are using.
The idea is to follow that trend. One of the big mistakes we’ve seen is AI companies forgetting that the way you should use pricing is to show that you're providing value. The companies that are really successful right now with AI – Cursor, Perplexity, Ramp AI, Blindspot AI – beyond the fact that they're solving a real pain point and have extremely good UI/UX, a main reason for their success is the pricing.
It's very clear what you get when you sign up for Cursor: you get access to an engineer for $200 a month instead of a full-time employee. Same with Perplexity: you get an analyst in your pocket for 20 bucks. Not using value-based pricing (or at least showcasing the value you provide with your tool) is one of the common mistakes we've seen with these AI tools.
There is definitely a lot of complexity depending on different pricing models (which we'll discuss a little later in this conversation). But really, I think the big mistake is forgetting that if you want to be successful when launching your AI SaaS or AI platform, you need to focus on the value – be super strong in conveying what's in it for the customer and quantifying that in dollar amounts.
That all comes through in the pricing. Regardless of what pricing style, tiers, or strategy you use, that must be the focus. I would say that's probably the biggest common mistake we've seen so far.
Brian Graf: Yeah. At the end of the day, you have two basic pillars or boxes that you need to check with your pricing. You need to simplify, simplify, simplify – be able to very cleanly and clearly communicate what differentiates the tiers and what value you provide. And to your point, you absolutely need to back up the value you're providing with your pricing.
One thing I've seen – an area of opportunity for a lot of companies that use subscription models – is that they end up using subscription models almost because they can hide behind the subscription, right? They count on the price fading into the background and getting lost in a company's expense tracking, more than it actively conveying value.
Of course, that might deliver a bit more money in the short term, but in the long term you're just delaying churn. You’re setting yourself up with a system where you're disadvantaged against someone who comes in with a really active pricing strategy and actively communicates value through that pricing strategy.
I think that might be one reason why, with this AI boom, we're seeing usage-based pricing come into play more. But of course, there are many more models. Let’s talk about usage-based pricing for a second. Are there instances where that can backfire in a B2B SaaS scenario? Is there a case where it wouldn't be beneficial?
Antoine Vial: That's a good question. Usage-based pricing is a preferred option for a lot of companies at this moment. But it almost feels like going back to the days when you had to pay for extra credits for SMS or when you hit your limit. It's a model that works really well for B2B because companies are used to these kinds of models – less so on the B2C side of things.
I think maybe a hybrid is the best combo here: seat-based plus usage-based (credits).
Some pros of this kind of strategy are that you allow your clients to balance how many people have access to the tools and identify power users versus non-power users. You give that flexibility while still having seat-based pricing for predictable costs. And you allow some team members – power users – to increase their usage.
This gives an opportunity to scale over time, which is great for your clients as they get used to the tool and figure out if it's solving the use cases they planned to solve, then increase usage as they prove the value is there.
There’s definitely a flip side: with usage-based pricing, it's a little harder to predict your costs. If you think about a monthly subscription, if you have a surge in usage you may get a bit of a price scare. That's something to be cautious about, especially with customers that are sensitive to price increases. It's also a little harder to explain or even implement.
You have to explain that there might be tiers depending on usage: how do you get access to more, what happens when you haven't used your credit? All of these things make it a bit more complex to explain and provide value.
So I think for all pricing models, there's often a combo of different pricing styles, or at least a type of customer that is used to it or more prone to engage with that kind of pricing. It comes down to a fundamental question relevant for any B2B SaaS company: who is your target audience? Do that research prior to choosing your pricing model.
Ask, is this an industry that is used to this kind of pricing? If so, is it something they like or dislike? How can we position ourselves to match industry expectations and always provide value through our pricing model, whether it's flexibility or something else?
For usage-based, Data.com is a great example where you can have different tiers of usage and stay within that pocket so you don't have strict price tiers. That's my take on usage-based/credit-based pricing.
Brian Graf: At the end of the day, right, usage-based pricing almost depends on how you look at it. In one sense it reduces friction, but in another sense, it adds friction to the decision. It definitely adds complexity and variance in payments.
The nice thing about a subscription is you know what you're getting – you can put it into your budget as a consistent line item and not worry about it. But particularly with something like an AI service, where costs differ dramatically depending on usage, it's an interesting problem. As a company, you don't want to hit someone with a complex contract negotiation when they hit a ceiling.
That doesn't help anybody and it's not going to make your customers happy. You definitely don't want to paywall them and say, "Hey, you had a great month with a lot of people using your product, but now you're shut off until you renegotiate with me if you want them to continue to see value."
That's not a good interaction with your customer.
So, in many ways it makes sense to go usage-based, but from a CFO perspective I could see it being a big blocker – like, "Hey, I have no idea what my costs are going to be given how much variance this introduces into our cost structure."
Brian Graf: What about pricing from a credibility standpoint? Do you view pricing as something that can reinforce credibility and reduce adoption friction? How do you see that playing out across some of these pricing strategies?
Antoine Vial: Yeah, this is one of my favorite pricing models. Prior to all these AI solutions popping up left and right, I think value-based pricing is definitely one way to reduce the risk during the buying process for your prospects.
When you tell a prospect, "I'm taking a portion of the output that I provide," – a simple example: let's say it's a cloud cost optimization platform and you're selling a solution that will save them 25% of their cost, and you take a percentage of that – that's a very strong statement that you are confident in your ability to deliver results.
It's a win-win: you save money, I get a portion of it, everyone is happy. It also significantly reduces the risk. If it doesn't work, then there is no payment. So value-based pricing, regardless of the solution, is very relevant for this kind of scenario. It can be risky and has to be implemented properly and also proven.
I highly recommend running pilots, proving that you know exactly who your target audience is, showing that it works and that you can have predictability in your outcomes.
But if you do that, then it's an extremely powerful pricing model that will outcompete almost everyone if done properly and competitively. I absolutely believe it was relevant before, is even more relevant now in the AI context, and will continue to be relevant – because at the core of these pricing models, you are delivering value.
You are making a statement: "I am 100% confident I can get you X amount of savings, and therefore I will take a portion of those savings." It's pretty hard to provide a more compelling story for any prospect in the buying process, because you reduce the risk essentially to zero.
Brian Graf: Yeah. There was a FinTech SaaS company I talked to a while ago whose pricing was "1% of all the money that we save you." It's super crisp, super simple, and adds a ton of value. The ROI for the client is tremendous, right? But of course, like you said, you have to be really confident in your product to go to bat like that.
Honestly, you could argue – it's not always that simple, but – if you're not that confident in your product, you should probably be looking at your product and its fit, right? And figure out how to improve it so you can deliver something (maybe not that aggressive) but where you can really put your money where your mouth is and communicate your value through your pricing.
Antoine Vial: Yeah, and I think it's especially relevant in the AI context because of all these AI agents and productivity improvement value propositions that are really put front and center by these companies. It makes sense.
But yes, it's always hard to define and measure value and outcome, right? If you aren't confident in defining what that value is or what that outcome will be, then it can become a risky strategy. Also, with this kind of pricing, you have the complexity of implementation – how do you define it across all of your accounts?
Your accounts might differ; can you standardize it in a way that works for all accounts? For these AI agents, it might make sense – you're basically reducing a full-time employee’s work to a dollar amount of productivity, and that's how you quantify the value.
But yeah, the risk is always in defining that value and outcome.
Brian Graf: And I think that model actually works really well for AI. This might be just a personal opinion, but the general snag that I hit with AI (and I assume a lot of other people do as well) is whenever I get introduced to a new tool, I'm like, "Oh, that's a really cool idea. What a cool concept. But how good is it actually going to be? How much value will it actually deliver in practice versus just being a good idea?"
I feel like that's a big gap that has to be crossed in selling AI tools.
Having your pricing back up your claims really helps establish credibility for those prospects and bring them further down the funnel.
Just to bring us back to positioning for two seconds and then we'll go right back to pricing – I always like to ask my clients: what's the guarantee you're willing to put on the line to showcase your faith in your product and the value you think you can deliver? This value-based approach is a way to bake that directly into your pricing, which can be really impactful.
Brian Graf: Okay, so let's zoom out a little bit and talk about some of the different pricing models you've seen in AI SaaS. What are the main ones you've run into? Give me some pros and cons of each and we can talk about them.
Antoine Vial: Yeah, absolutely. I think we can focus on six of the main ones that we've seen most often.
First, you have your traditional SaaS pricing: per seat, per month, subscription-based. We mentioned this earlier. Advantages of this model are that it's very simple to explain and it's very predictable for your customers. It makes it easy for a CFO to approve – does it fit the budget, yes or no – which facilitates the conversation, probably shortens sales cycles, and makes forecasting easy.
On the revenue side for your own purposes, it's fairly easy to know, based on tiers or subscription levels, "If I hit X amount at this pricing tier, I will have Y revenue," so you can forecast with this model. Of course, a downside is that it's not as focused on value – as we were talking about. It's a very common model for something like CRMs: "Hey, we facilitate [some process]" more than directly tie to value.
One tool that uses this pretty well is Jasper. It's user-per-month and gives you access to content generation, and then you let your users use the tool infinitely. So that's one – probably the most common and still very relevant for quite a few use cases in the AI space.
Then you have tiered feature-based pricing. Here, one of the big pros is that the upgrade path is very clear and straightforward for users. It allows users to align the pricing with the value the tool is delivering. If they're on Tier 1, they can stay on Tier 1 as long as it provides enough value, and they can clearly see, "Oh, the cost of additional value (Tier 2) would be this – is it worth it or not?"
Then they make the decision. That's definitely a great way to drive upgrades.
Brian Graf: And you can use something like that as much as you want, right? If you're worried about usage variability or you know you're really going to use the hell out of a certain feature, a feature-tiered approach is a very clean way to accommodate that for the user.
Antoine Vial: 100%. Next, you have hybrids – like we talked about – where it's seat/subscription-based plus usage-based. I think you get the pros of predictability from the subscription, and the pros of the usage/credit-based model, combined into something fairly predictable but that also has the ability to ramp up depending on usage and perceived value. It's definitely a great type of pricing for AI solutions.
The downside is it is harder to predict the pricing, and a CFO might push back a little more during the sales process because there's variability. Companies sensitive to price surprises will be more wary of these kinds of models.
Then you have purely credit-based. Here one big advantage is it's basically pay-as-you-go, which gives flexibility – if you want more of a tool, you pay more; if you don't need it, you don't pay. You probably want to implement it in a way that allows customers to easily increase or decrease their usage or payment, to make this strategy successful. But that's the big value prop.
We've seen a lot of AI tools using this, especially those with heavy API usage, focusing on this kind of pricing.
Brian Graf: Just to back up on credit-based – and this applies to a few other models – in AI, a land-and-expand strategy is often pretty applicable. A credit-based or usage-based model is a nice way to reduce friction in the beginning: get yourself in the door, let people try it without the CFO freaking out because it's low cost to start, then allow them to see the value and start to sell you internally, opening the door for expansion and upsell. I see a benefit there.
But to your earlier point, it definitely brings back memories of, like, "Oh, I can't text you until 9:00 PM because I'm out of minutes or credits," which can be frustrating.
Antoine Vial: Or even worse – "I'm allowed to text and use all the credits I want," and then the bill comes in and you're like, "What?!" (laughter) Absolutely. This model is definitely applicable for a lot of AI automation tools – per-token usage-based or credit-based pricing, which is very relevant right now.
Businesses are more used to it than end consumers are. I think it's a great way to scale up and expand within accounts. The value of your tool has to be, of course, beyond the price you're charging – as long as you do that, this provides a very scalable pricing model when implemented properly.
So that was the fourth one. Then you have per SKU (or per feature) pricing. This gives you the ability to provide great transparency into the performance of your pricing and gives you control over cost and quality of what users are doing with the tool. You see some major players using that strategy. It can be a layer added to your pricing; it doesn't have to stand alone.
Depending on features and capabilities, certain ones may fit very well with this strategy – not relevant for all AI tools, but definitely something that has been very successful for some AI companies.
Brian Graf: It almost combines feature-based pricing with credit-based or usage-based, right?
Antoine Vial: Absolutely. And you have the traditional tiered pricing, right? That allows users to ramp up as they see value and also go down. Right now, ChatGPT is a great example of different pricing tiers: you have a free version, a $20/month version, a $200/month version, and maybe eventually an even more expensive enterprise version.
Brian Graf: Sure, it'll go up from there. And that's not even getting into the API costs and everything like that.
Antoine Vial: Exactly. The complexity of these pricing models is exponential as the use cases expand. But yeah, this provides a lot of flexibility for users.
And the last one is the one we already talked about: value-based pricing. This was around before, remains during the AI wave, and will likely remain supreme after. The ability to show that you tie your pricing to an outcome is very powerful – it's a strong proof of confidence in your ability to solve the problem you say you solve. In any sales conversation, it helps shorten the sales cycle and lower risk. It's just that much more powerful.
Brian Graf: I mean, there's a reason that one kind of reigns supreme. A couple of reasons actually: (A) it's the easiest way for customers to directly link the value they're receiving from your product to the price they're paying for it – so of course it's the most obvious to them. But it's also the hardest to pull off. It's very easy to talk about, but to actually do true value-based pricing, you need to deeply know your customer and your product.
You need to be really confident in your product. You probably need product-market fit. All these things require really putting pen to paper and rubber meeting the road – you need real traction and an established product actively driving value to accomplish this. There's a reason it's in its own league at the top, but also quite elusive for some companies that can't check those boxes.
Antoine Vial: Yeah, and it also illustrates where you are in your growth journey. You hit on a few points: are you at the MVP stage? Product-market fit stage? Have you proven you can handle these kinds of pricing models? There are still some downsides, like lack of predictability even with value-based pricing. If you save 10x more than you were planning on, you’re still going to cost the customer way more. It might be an easy sell (because it's all savings), but you're still increasing their cost.
So there are always pros and cons for all these models. It really depends on what you're trying to achieve. Some pricing models can be a means to prove you've reached MVP – like, can you reduce the friction for early adopters by using a very simple pricing model? Yes/no. Can you keep your customers? Yes/no.
Depending on those answers, you can play with different variations or combinations of these models. Pricing and packaging is definitely part of the positioning and messaging conversation – they go together.
That's something to take into consideration when planning the implementation of a new pricing model.
Brian Graf: Is there a link in your mind between different pricing models and the maturity of the product? You mentioned MVP versus PMF. Do you have a recommendation for, say, an MVP-stage product that's just getting started to break into the market, and once they've hit PMF and want to scale? Of course there's a lot of intricacies in there, but any general rules of thumb that people should take a look at?
Antoine Vial: Yeah. I think it's fairly safe to say that when you are in MVP (untested) mode, pricing falls under that umbrella – you need to test. You need to test what lands and what resonates with your target audience. At that stage, a simple subscription model might make a lot of sense. Or even a usage-based model, just to see if your customers are actually adopting your tool.
That can include tiers with a free version, and then trying to push them – depending on your go-to-market strategy. Are you going PLG (product-led growth) or sales-led go-to-market?
If you're an MVP-stage product with a PLG strategy, you probably want to reduce friction as much as possible and increase usage as much as possible to get data and see usage behavior. So you want to reduce the barrier to entry and friction as much as possible. Any pricing model that fits that goal should be implemented.
You may even "enterprise" your tool (provide it for free) and realize it'll cost you more upfront, but you can always adjust. That's part of the learning process of getting a solid MVP.
As you move along the journey and start to see customers who pay you stick around – happy to give you their money – then you can solidify your pricing a bit more and play with models like success-based pricing (outcome-based) once you know you have predictability in the outcomes you provide. You can introduce hybrid solutions.
You can also play with the visibility of your pricing as a positioning tool. For example, being very clear and transparent in your pricing – you see a lot of companies providing pricing models with a free-forever tier and a lot of transparency. This approach helps user acquisition. People are willing to engage with these kinds of tools because they know they will always get value out of it. And if they believe they can get even more, they'll be happy to pay.
I think you have a ton of clever ways to play around with pricing strategies, but they always need to align with where you stand in your growth journey. At the beginning of the growth journey (MVP stage), try to reduce friction as much as possible. When you hit PMF, try to attract more of your ideal customer profiles – the customers that you've seen are easy to serve, low friction, and high fit.
Your pricing strategy should allow you to bring more of those in, and maybe filter out more of the ones that don't fit. Later on, as you scale, try to increase value per account as much as possible. Whatever pricing model fits those goals at each stage will be my recommendation on how to use pricing depending on your growth stage.
Brian Graf: Yeah, I agree. I don't have a ton to add to that. For MVP, you need to get people in the door, testing the product, giving you feedback – so your pricing should be very approachable (like a simple subscription) or use something like usage-based to allow an easy ramp. The con for you there is you're playing with your own revenue variability a bit, so it can be hard to project your business's growth off of it.
Then, yeah, of course when you have PMF, build your pricing around your ICP and push towards value-based if you can. As you start to scale, you're really just playing around with – you could stay with the same model if you want, but now it's about things like adding more credibility, having more brand, having more pricing power.
You should be able to either stay in the same model, or start playing around with margins: how much do I want to capitalize on the market in front of me versus slim down margins and go for market share so I can capitalize later? Those decisions come into play as you progress.
I think that's a good place to wrap. I feel like this has been good. I appreciate all your insights on it – this is super useful. I hope this gives some founders, executives, and marketers good cards to play in terms of pricing and some models to explore. Any last thoughts from you, Antoine?
Antoine Vial: Not really – thanks for having me. This was a great conversation. Maybe one last thing: all of these pricing strategies, we looked at them through the lens of how to apply them for AI solutions, but they were relevant for non-AI solutions before. Learning from how other tools have implemented SaaS pricing is also very relevant for any tool that includes AI-powered features.
This conversation was through the lens of AI, but it's applicable to any other type of SaaS.
Brian Graf: Yeah, 100%. In general, sometimes there are big pricing changes that change how everyone looks at pricing. I don't know that AI is doing that – it's kind of refocusing us on usage-based pricing and bringing that back to center stage, but really it's a new and dynamic market to apply a lot of the same pricing theory that's worked in the past, just with a slight tweak.
A lot of this can be applied to your standard B2B SaaS company as well.
Antoine Vial: Okay.
Brian Graf: Well, thank you very much.
Antoine Vial: Thank you, Brian. Talk to you soon.
Brian Graf: Thank you to Adriano Valerio for producing this episode and the Kalungi team for helping us make this whole thing work – and of course, thank you for choosing to spend your time with us. As a reminder, all the links we mentioned in this episode can be found in the show notes.
If you want to submit or vote on a question you'd like us to answer, you can do that at Kalungi.com/podcast. Every time we record, we take one of the top three topics and jam on it.
Thanks again!
.jpg?width=192&height=192&name=saascmosnacks2%20(1).jpg)
Head back to the B2B SaaS Marketing Snacks home page for more.
Navigate the latest SaaS trends such as AI's changing role in marketing, to other significant changes in personalization, content strategy, and...
How do you make sure that your relationship bank account with prospects and customers is always positive? If your SaaS company focuses only on what...
With the right setup, maintenance and usage, your CRM strategy will be the key that unlocks your B2B SaaS startup’s true revenue machine.