Pricing Page unPacked - SALESFORCE: Pricing the Agentic AI Era (and Why It's Complicated)

Rob Litterst (0:26)

Welcome to Pricing Page unPacked. I'm Rob Litterst. I'm 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. I'm excited for this one, Ulrik. We're talking about Salesforce. From what I understand, you don't really have too much to offer about Salesforce, don't really have too much to talk about.

Salesforce: the poster child of seat-based SaaS

Ulrik Lehrskov-Schmidt (0:54)

So Salesforce is it's the it's I bring them up again and again because they are like they're the poster child of seat-based SaaS. But then when you dig in, it's like, are they really right? Because they do 35 and some change, about 35 billion of revenue last year, they do six billion bottom line. So it's like, okay, it's like 20% margin, even a bit less, right? But then the idea is that the sales and marketing expense on that is like 13 billion.

And the year like the growth that they had year over year is like three billion. So it's like they're one of the companies that they literally spent four and a half dollars to acquire one dollar of revenue. And I think the reason for it is that they were like one of the first SaaS companies that grew quickly, and they were sort of trying to figure out how to do it. So they made all of the mistakes so that the rest of us don't have to, right?

And one of those things was that they just incurred a huge amount of what I call commercial debt. Like extremely complex packaging, extremely complex pricing, lots of sort of sales-led, bespoke deal motions that just like this layer of legacy that is Salesforce, right? And then I think for the conversation today, we can talk a little bit about that and how they got there, but they're also trying to really reinvent themselves with AI and Agentforce. And I think for me, the question right now, and I'm actually really curious to hear your take on it, is that is it just is it really sort of a fundamental rethink of what the model is, or are they just gonna stack on even more complexity and legacy now just with an AI layer on it?

Rob Litterst (2:43)

Yeah, that's a great question. To your point, I think like they are definitely kind of like the creator, or at least like one of the initial creators of like seat-based pricing in SaaS. And I think like if you look at their model today, you can still see they're charging per user per month. So like they still have a model on their CRM that is per user per month, seat-based. But then they do have and they have like different levels here. To your point, I think Salesforce has always had like a lot of different levels of pricing, and there probably is a pretty phenomenal amount of commercial debt going on underneath the covers.

But where it gets really interesting, I think, is what they're starting to do with Agentforce and how they're trying to kind of shift their model. Like I think last week we talked about Intercom and how Intercom like really kind of like burned down the boats, right? And like really kind of like went into the new AI era and were more than willing to kind of just like throw their old model away. They invested marketing dollars in Fin. They like basically put all of their positioning and marketing on the website around Fin and just completely shifted.

I think Salesforce has been super aggressive with Agentforce, but I don't get the sense that they're doing the same level of kind of like burning the boats as Fin. It seems more experimental and more kind of like they're iterating on Agentforce to see what works. And I think they've taken a lot of heat. Like when I look at LinkedIn, people talk about how you know, Salesforce tried this model and then they went back, and then they tried this model and then they went back, and then they tried this model and then they went back.

And like sometimes it seems like people are kind of like insulting them for not knowing how to price agents. But I kinda I respect it, honestly. Like it's it seems like they're really trying to iterate, they're really trying to push in the new direction. And right now I think they're kind of like pioneering this kind of idea of flexibility as a massive benefit. They're like kind of letting people pay how they want, they're letting them buy how they want, and they're just giving customers a lot of different options.

And it seems like that is probably gonna become a wider trend. The thing to mention about Salesforce is they're basically like a SaaS holding company at this point. Like they're basically like an index fund of SaaS. So like, well, we're gonna focus on kind of like Salesforce core, like their sales software and Agentforce today. I think the interesting thing with Salesforce right now is this kind of tension between seat-based pricing and whatever the future holds, outcome-based pricing, credit-based pricing, wherever that's going.

And so right now, like looking at their sales pricing, I look at what they're doing with the sales section right now, and it just feels like kind of traditional SaaS with a little bit more bells and whistles and kind of some more differentiation around different seats.

The anti-PLG motion and commercial debt

Ulrik Lehrskov-Schmidt (5:31)

What I'm thinking is that it's sort of the anti-PLG motion, right? So and that's I think part of the culture of the company is that they they've always been very sales led. So when I'm looking at a pricing piece like this, and we have sort of three tiers, and then they have three enterprise tiers, and they're like one is free, and then they go up to 550 a month. I don't know how many people actually go to the website and buy Salesforce.

Like how many like, hey, I'm now gonna be a Salesforce customer, I'm gonna like click try for free, and then I'm gonna like get started, right? I just don't see that happening a lot. I see salespeople going in having a say a CRM conversation around, yeah, well, the system you have, they they're sort of not letting sort of enterprise ready or whatever it is, and then try our thing, and then we sort of we stitch together a solution, and then that's the idea, because that solution stitching is now based off of a an extremely large like collection of different sort of product Lego blocks that that salesperson has to sell, right?

So it's like okay, you're gonna get like you're gonna get a CRM system and you're gonna get like maybe Slack and like five other things, and then I'm gonna stitch on some add-ons, and then like I it's all it's it turns into solution selling even at the small end of the market, which means that this what we're looking at here, 25 bucks a month is in itself super simple, but it's not once you scratch the surface, like because there are like too many products behind it, and any kind of sales process that I engage in with Salesforce is going to take me through that conversation.

The historical norm of Salesforce is that they have some discount because they don't have any discounting here, right? So that means that that a lot of the discounting now happens after this fact at the sales level. So they don't have any structural discount. So it's like, hey, the salesperson has a mandate to like charge the price or give a discount. And then that also means that every account, even the ones that are at 25 per user per month, like smaller accounts, are now being closed on varying degrees of pricing, which means that you just have a tremendous amount of complexity, even in small accounts.

This is sort of the reason why it takes them four and a half dollars just to grow another dollar. Part of that is that most of those 13 billion that they spend every year is actually not spent on selling that new 3 billion of revenue, it's spent on renewing the 30 billion of revenue they have. Right? So it's like they have created this situation for themselves where they like every year you have to spend $10 billion to not shrink because every year customers show up and you have to sort of resell them all over again, right?

And because they have so many different products stitched together, so many different discounts and so many different ways, even though each individual piece is simple, the collected sort of solution that you're selling is super complicated and not even that, but also bespoke, it's individualized. And that's really what's creating the situation for Salesforce.

Aggressive discounting as a sales weapon

Rob Litterst (8:41)

And I think like one interesting thing going up against Salesforce, because I used to sell HubSpot, as we mentioned on a previous episode. The kind of nature of what you're talking about, where they have like all these different hubs and all these different modules, is that they can get very aggressive with discounting. Like I was trying to sell this like back in the day, so it's like over 10 years ago. I was trying to sell HubSpot to the New York Times, and they were obviously considering Salesforce.

And we got like to the one yard line in the sales process, and they basically said, We're going with Salesforce because they're giving us part of it for free. Like I discounted down to a certain level for them, but then they were like, Well, Salesforce is just gonna give it to us. Yeah. And I was like, Oh, okay, well, yeah, I guess that takes care of that. So like it allows them to really get aggressive with pricing.

Ulrik Lehrskov-Schmidt (9:30)

A friend of mine, he was like a pretty high-level executive in a bank, and I won't say which bank, but they were the largest Salesforce account in Europe. So, and he's he was the one that signed off on them getting like he ripped out the entire CRM system, they bought all the banks and they were sort of consolidating his thing. So he was like, I'm gonna do that. And he said, and he like they flew him to San Francisco, he had a dinner with like he met Woody Hells and then met Mark Benioff, like they like even all of these accounts.

Like, okay, we're gonna they're good at that stuff, right? Yeah, and then he said, you know what? Like the way it works is that we have like there was a price, but they were just like discounting 90% year one, and then the next year would be 80 and then 70, and then 60. He said, It's fine, I'm gonna have another job before this gets serious. Right, right. Like, I don't care, like it's a like yeah, like the end result is gonna be like expensive, but probably worth it.

But it's gonna be a decade before we even hit that level, and you know, right next couple years when it's my PL, I'm gonna be gold, right? And that was sort of the idea. So just like the whole discounting structure was sort of set up to like to play into this sort of corporate politics career game where you sell to an executive that like gets to like create this like momentum in the organization with the new system, and then the bill is sort of the next guy, right?

A product suite that isn't really a platform

Rob Litterst (10:51)

That's hilarious. I also do feel like one of the things that's been kind of like a glaring hole for Salesforce that I think they're working on, but like even though they have all these different product suites, it's never felt like they're super integrated with each other. So it's never felt like they have this like real kind of like integrated like platform play where everything is talking to each other in a really clear way. And it's like this kind of like integrated value prop.

Like, I think they finally started to really integrate Slack with Salesforce CRM. And I've heard that that's actually really, really good. Like I've seen some posts about that on LinkedIn. But for the longest time, it felt like Slack and Salesforce were like two completely different things that had absolutely no relationship to each other. And so like the it felt like their kind of like suite of products was, to your point, kind of more of like a it's like more of a kind of like a commercial thing, like a commercial value prop of like you can get these things for free, but it wasn't necessarily like a product value prop where like you know these things work really well smoothly together.

Ulrik Lehrskov-Schmidt (11:52)

They started in like the height of the internet era, like 1999, like in the literally like last millennium software company. And then I think they did a lot of the mistakes that you have to do. One is that they didn't really have the technologies really sort of build fully integrated products, and so it was just easier and cheaper and faster to just build a little bit like standalone and just plop them on as modules. So there was a technological side to it, but I think another side was that they had the let's say knee-jerk reaction to development is that hey, like if we develop this new functionality, well, we already signed the contract with all these other guys in these like weirdly bespoke ways.

So if we just add the new functionality into the thing that others have, we have no way of charging for it. Like, because it just like then it's just giving away free. So therefore, we now put it on as yet another add-on, right? And then we spin out the sales motion and go out and try and sell it, right? And then over time, it turns into this like many-headed dragon, this hydra of like a product that it just it has a tremendous amount of complexity and SKUs, and then you need that sales force, like the sales people to actually sort of navigate that complexity on behalf of the customer, right?

And then you don't need to have the like the clean like product platform at the back end that just ties everything together because you just do it on the commercial layer instead.

Rob Litterst (13:24)

No, it makes a ton of sense. It makes sense too why HubSpot has kind of tried to go the integrated product route as kind of like an answer to Salesforce. But to your point, I mean you look at the complexity, even within, and remember, like this is just one, this is basically like one SKU for that. Like the multiple SKUs within here, but like this is like one product that Salesforce has. They probably have like 50 to 100 products. Within here, you have six different options, just for CRM.

And so like, you know, it I think you take that and it compounds with all the different products and suites that they have, and it gets really convoluted really quickly.

Agentforce: from paying for access to paying for outcomes

Rob Litterst

Let's talk about AI pricing, Ulrik, because I think this is where Salesforce wants to go. I think like just to set this up, the kind of fundamental tension here is this idea that we're moving from paying for access to paying for outcomes. We're gonna move from paying for seat-based licenses for humans to paying for outcomes that actually drive work and get work done. And whether it's for humans or agents, that's where kind of SaaS is going. That's where all the lines are headed.

Salesforce has really tried to get ahead of this and they've been very loud about it with Agentforce. And they have been very quick to iterate on pricing, which I'll give them credit for. That there are a lot of companies in their position, incumbents, who have not really experimented much and kind of seem to be waiting. Salesforce is not. They've been playing around with this. Curious, like any initial observations before we jump into the specifics.

Digital labor and a startup-style hustle phase

Ulrik Lehrskov-Schmidt (14:49)

So, one observation is that I often I use Salesforce and OpenAI as two, let's say, large companies that actually they have they've sort of fundamentally sort of found out or decided that the product is one half of the equation, the commercial model is the other. So, and you just need to test out different commercial models and pricing models in order to find the right one. So, like, hey, let's like I think they changed their pricing like in big ways, at least 10 times in the last year.

Like they would roll up credits, then they would like pull them into seats, they do something else. So, and that to many companies of that size is just like an unthinkable speed to just have like something that big happen every like five weeks, right? So that's one of the high-level comment. And you can see here, like they say like two things that I notice on this space. One is that they say they call it digital labor. So this is like a really good price anchor.

They're saying, hey, we are like you have labor now, but the new thing is like digital labor. So like those guys that work for you that you pay a lot of money that you don't really like get rid of those guys and then use our AI. And then, but now so what we're doing is we're anchoring to the salaries of the people that we're replacing, right? So that gives them sort of like price anchor to what we're doing, and then they're saying, hey, you can choose from consumption pricing, Flex Credits, whatever Flex Credits are, conversations, per user licensing, and like so they're just they're basically saying, Hey, we are behaving a little bit like a startup.

Usually I say a startup is allowed what I call like a hustle phase between like zero and five million ARR, where they get to like try different pricing and see what works. Salesforce just sets this out, like, hey, we're at 35 billion, but dude, we're just gonna do that. We're just gonna like see what works for the next one to five billion, and then we're gonna probably throw out the 99 models that didn't work and keep the one that did, right?

Which is like pretty awesome, actually.

Rob Litterst (16:45)

To your point, I think like calling it digital labor, like basically setting the stage so that you're pulling from the headcount budget rather than the software budget is really smart. I think, you know, we haven't really seen a lot of companies go super deep down that kind of positioning route with AI pricing because I think there's a lot of tension around AI replacing jobs. Like I think we saw this pricing change a few weeks back from this agent company called Lindy, and they basically added this human to their pricing page.

It looks like they actually got rid of it. But they literally used to have like this human decoy plan over to the left that was basically like making fun of humans, like works some sometimes, doesn't really understand Excel, like all this sort of stuff. So I think like this is an elegant way to kind of position yourself that you're tapping into the headcount budget rather than just software spend.

Flex Credits, conversations and buying models

Rob Litterst

Well, so getting into how they have actually started to charge for Agentforce, they've kind of tapped into consumption-based pricing, they've tapped into credit-based pricing, they've tapped into sort of outcome-based pricing. It seems like they've kind of tapped into everything. Like they're really trying it all. And it seems like they are basically, and I don't mean this in a negative way, they're kind of throwing spaghetti at the wall and seeing what sticks with this and seeing what people actually want and how they want to pay for it.

But you see, they allow people to get started for free. They allow people to pay for Flex Credits, what they call it, when you pay per action, or you can pay for conversation similar to some of the other kind of outcome-based pricing, customer support companies. What's your read on this kind of strategy where you throw all the models out there at once? Because the one other thing that I'll mention is they also offer a bunch of different buying models. They allow you to pre-purchase and draw down.

They allow you to pay as you go, they allow you to pre-commit and commit to a certain amount, but then pay in arrears. So they're really allowing people to kind of like choose their payment model and then change, choose their buying model as well. And actually, it seems like for different types of payment models, there's different buying model flexibility. So Flex Credits, you get like fully flexible conversations, you have to do pre-purchase only.

Too many options: analysis paralysis for buyers

Ulrik Lehrskov-Schmidt (18:55)

Yeah. So as I said again, like this is just testing what works, right? So in a normal situation, you might have had like imagine that you had like an internal project with someone like Salesforce say, hey, we can run this pricing in a lot of different ways. We can run like on individual actions, we can like we can do another more sort of all-in, like higher-level metric than it's per conversation. What do you want to do? And then you can run the numbers and like you come to sort of modeling.

They look they they've essentially just said, hey, let's not make the decision, we'll just push it into the market and see what gathers revenue, right? I think, so I think that's really the interesting part. The problem with something like this is that you're if we take the like the free foundation out, now you have we have two models. Okay, you want to pay for credits, you pay for conversations. You're like, okay, I don't know how I'm gonna use this product. I'm gonna use it in a way, and that means that one of these two pricing models is gonna be the right one for me, and the other one is gonna be wrong.

So that also means that right now I have the ability to regret something. I can make a bad decision for myself, right? I can choose conversations and I should have chosen Flex Credits or vice versa. And as the customer is actually not aware how they're gonna use Agentforce, they might actually stop and pause and think and say, I don't know. Like, well, the first question you're gonna ask is, well, how many Flex Credits am I gonna use per conversation? Because I'm gonna try to like math them together and then like and like arrive at a like a conclusion.

The answer to that is well, that depends on how you use the system and what like and so okay, so right. So then you're sort of back to square one. I think while they're testing something, I think actually they might also be facing a little bit of like analysis paralysis with their customers because it's so clear that I have the option of choosing something wrong when I'm entering into this model.

Rob Litterst (20:46)

Okay, I think this goes back to what you said in the very beginning, that they are like the quintessential like sales-led growth company. Because like adding this kind of complexity to your pricing page doesn't lend itself to PLG at all. It basically makes it so that you have to jump on a call with a sales rep and like diagnose what makes sense and figure out the best model for you going forward, which just plays into their model and allows their sales reps to feel like they're you know educating the customer, building a deeper relationship that way, those sorts of things.

Ulrik Lehrskov-Schmidt (21:17)

And I think if I look at this, I'm gonna assume that they actually want people to buy Flex Credits. Like that seems like the most comprehensive thing. That's also the one where they give the most different options, right? So conversations is pre-purchase only, it doesn't have employee-facing agents, it doesn't have the voice. So it's like, okay, so it seems like if business gets like really serious, and if I'm a really like an enterprise customer or have a like a complex use case, we kind of want you to buy credits.

But if like you're not that, then it's okay buying conversations. But I definitely also see this model where somebody would start with conversations and then suddenly like Oh, now I want employee facing agents. Like, oh, now you just like shift your entire pricing model to something else, right? And then you get into these. Well, can't I can I keep my customer facing agents price per conversation? And then the other ones price per credit. And then you get into all this kind of mess, right?

So it's one of these like I would really want to be in the room when they're trying to sort of consolidate all this and like figure out even like how we when you get data from something like this, how are you gonna make conclusions based off of it, right? So okay, we have a bunch of customers, some chose this bucket, some chose the other. Like this is the money. You have no way of figuring out what the outcome would have been if you just had forced one bucket.

So you just have a lot of data, but you actually don't have the answer to the question you want, which is which one is best? It's one of these where props for testing things out, but the way you're testing it actually like it just punts the question and doesn't really answer it.

Rob Litterst (22:52)

Yeah, it shows the it shows little conviction and kind of like the right way, right? It's like very it's like there's this idea of kind of like opinionated software. It's like you should use our software this way, you should buy our software this way. This is like the exact opposite of that. It like lets people kind of do whatever they want. And in a time with a lot of uncertainty, I don't necessarily know that's what people want, right? Like people kind of want to be told what to do in some ways right now, I think.

That is actually happening.

Ulrik Lehrskov-Schmidt (23:18)

But what Salesforce has done is again, they're the quintessential sort of sales-led companies. They've said instead of centralizing the decision from a product perspective, we're pushing it out into the sales part of the organization. So the salesperson has all these models at his or her disposal, but now it's their job to talk to the customer, understand what they want and need, and then propose something, right? One, so like I think you should go with credits, like it sounds like you could go with a whatever prepay model, but I would really want you to on the whatever commit model, and that's gonna be best for you.

And then we like do this, and I give you this discount. What do you say, right? And then you go on it's on a salesperson's part of his or her job description to reduce all this complexity, and then it becomes actually like that relationship, then starts to matter because as an individual buyer, I have no way of figuring this out, right? To need that almost like a doctor to like help me diagnose the problem and like find the right solution for me, and I'm just gonna trust that you know what you're doing, and then like we'll get through this together, right?

It creates this incredible dependency on this whole like labor-intensive part of the organization, which is the sales organization.

Salesforce alumni and CRM best practices

Rob Litterst (24:34)

One of the interesting things about Salesforce in the world of SaaS is Salesforce was very early and they got very big. They kind of have this like PTC type thing where like a ton of the leaders from Salesforce went on to be leaders at other SaaS companies. Yeah. And so those leaders knew how to use Salesforce and would talk to their old buddies to get their new company set up on Salesforce. And then that would just like waterfall down, right? It's like a Salesforce leader came and like ran the sales team at HubSpot for a while.

They got us set up on like a really complicated instance of Salesforce. And then his, you know, direct reports went on to other companies and set up those companies on Salesforce in the same way. It's like in the legacy seat-based model, Salesforce had this kind of like inherent advantage in that like they were kind of like best practices for setting up CRM and your sales operations and enablement. I really wonder in like the new era of AI, like I think AI has changed SaaS products and pricing so much that it's really going to kind of like chip into that advantage that they've had, which has obviously been kind of like deteriorating over the last few years, I think, as more players have joined the space.

But it gets rid of that kind of like immediate understanding that Salesforce's best practices in CRM. So it's really interesting. It's gonna be really interesting to see like what the new era of AI sales leaders are choosing as kind of like their go-to platform.

Credits as a Salesforce-wide currency

Ulrik Lehrskov-Schmidt (26:01)

And I think one of the things that that I was looking at Salesforce for a while and I was sort of hoping that they would go in that direction. I don't I think the jury is still out. And that is that in when they were introducing some of these Agentforce credits, they introduced them as sort of giveaways as part of the seat. So it's like, hey, every seat has can't remember the number, but let's say a thousand credits, whatever. There's a number of credits.

So if you have 100 seats, you're gonna get 100,000 credits, you can just use them here. And then there was a big discussion, and a lot of people were looking at Salesforce because, like, hey, these credits are they gonna be tied to the individual seat so that user has a thousand credits, or are they gonna be pooled across seats so that the organization as such has a hundred thousand credits? And that was like jury was out. I think actually they made some changes to that.

Salesforce had and still has this opportunity to actually clean up a lot of their pricing using a credit model in the following way. So if this takes off and they start to sell credit, actually what they can do is they can say, Hey, you can buy the credits, and then you can actually buy the seats with credits. So we could then say, Hey, the front end of the model is you buy credits, and then the credits become this sort of Salesforce currency that I can then use to buy the seats or like buy any of the other products, etc., etc.

So now they actually have this opportunity with the model to actually just like push everything like through the same entry point with like, hey, credits, and then credits give you access to anything inside the product. So you approve of sort of an upfront spend amount, you get X million credits, and then you purchase what you need inside of that, right? If they got out of the way and they just let the salespeople do that and then focus more on consolidating some of the products, making it more PLG friendly to try things out, because the commercial part, paying with credits was already taken care of.

I think they could use this new pricing model to actually evolve the commercial model as well and actually clean up a lot of that legacy that they that they've built up over the past two, three decades.

Rob Litterst (28:07)

I love all of that. And like thinking about Salesforce, like it kind of like breaks my brain to think about like how far that credit model could go across their ecosystem of products. But I think just staying with like CRM, just giving people the ability to buy credits and then say this can be for a human seat or it could be for an agent, like you're in full control over what that is. We're not even going to distinguish between the two, like you can just do whatever you want.

I think it's like a very cool way to offer flexibility. So yeah, I'm with you on that.

The verdict: buy, hold or sell?

Rob Litterst

Before we close out every episode, we always do our kind of like Wall Street version of Wall Street analyst assessment of the pricing strategy for these companies. So buy hold sell if we're going in order. What are you doing with Salesforce's pricing? What do you think about kind of how they're approaching Agentforce and where they're going? For me, it's a sell.

Ulrik Lehrskov-Schmidt (29:01)

I think it's one of these where it's hard to argue with someone that has six billion in net income, right? So on that aspect, like, okay, like I could just shut up and let them do their thing. They're growing 10% year over year, even less, and that's probably gonna get tougher and tougher. So I think it's they're fighting against like a gravity that they created themselves, right? And they are showing some signs of like wanting to sort of change, truly, but they haven't made any of the tough calls where they're really trying to let's say get rid of users or consolidate some of the products, and like so I think they would have some tremendous opportunities if they were more courageous on that end.

And I think that they're gonna run out of time in the next three years or so. Yeah. So they are still one of the larger CRM systems, like I think you have Microsoft and others like that are in there as well. I think they have so much holding them down of their own creation that that I don't see anything other than a sell. Sorry.

Rob Litterst (30:09)

What I would love to see is I would love to see them launch an agent that helps you actually figure out how the f*** to buy Salesforce. That would be great. Just launch the Salesforce buying agent to help take a customer through this complicated web and maze of pricing options. That is it for this episode of Pricing Page unPacked. 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 (30:36)

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