Skip to content
Back

Office Hours: Migrating to Credits

Rob Litterst · November 12, 2026 · 65 min

Watch the full session

Free PricingSaaS account required. The transcript below is open to everyone.

Sign in to watch

About this webinar

Rob Litterst moderates and Ulrik Lehrskov-Schmidt answers questions from attendees on moving to credit pricing. Ulrik frames credits as two ratios: dollars to credits, which stays fixed in the contract, and credits to product actions, which can change over time. The session covers whether to attach a rate card to an order form, how to migrate from location-based pricing, whether the market is swinging away from credits, where to start, how to combine actions of very different value, how to pass on token costs, whether non-AI features belong in credits, and when to use multiple credit types.

Transcript

12,015 words · 99 speaker turns

Rob Litterst0:14

Not really. Well I I just kicked it off. let me see. I think we probably have a few people. Yeah, we got a few people who jumped in. I figure we could give it a few minutes and and let people jump in and we can go from there.

Ulrik Lehrskov-Schmidt0:31

Yeah. I I was just on a call with Jeff who you introduced me to. He's a great guy, yeah.

Rob Litterst0:37

nice. Yeah, great guy. Yeah, we I worked with Autodesk a bunch at ProfitWell and I don't think but I don't think he ever was like one of the POCs that I worked with. I I I think he worked with somebody else from our team. They had like a few projects going on with us pretty much at all times.

Rob Litterst1:05

Cool, we can let people pile in. Hey everybody, we're just gonna give it a couple minutes and and let people jump in. and then we'll start getting into some of these questions. Seeing some familiar names and faces. Thank you everybody for jumping in. Hi, Radika. Let me see. Good to see you back in here.

Rob Litterst1:29

Still getting used to the chat in Riverside, Ulrich. It's like a different different animal than Google chat.

Ulrik Lehrskov-Schmidt1:37

Yeah, a little bit. See, Henry Body. It's yeah, we we've we've whenever you use Riverside, you have to talk about Riverside. Like it's just one of those so it's like, is it good, is it not good? Like pros and cons. Like it's it's it's a little it's it's nice, but but has this kind of formal feel. I guess a little bit like it feels like I'm using a piece of technology. It's like there is not friction, just just

Ulrik Lehrskov-Schmidt2:07

yeah. we can't hear you. I I

Rob Litterst2:11

It's yeah, that's unfort an unfortunate an unfortunate thing with the with like going with web webinar style on Riverside and without making this entire thing. You're right, Ulrich. It's like the opposite of Fight Club. When you when you're on Riverside, you talk about Riverside. It's it's just part of the game. But it's an unfortunate side effect of webinar style in Riverside. I believe there is a chance for you, Rudika, to request Sorry, one sec, what's going on? Sorry, my audio started playing on my headphones. There should be an an ability an ability for you to like call in live. here, let me see if I can And then if I see that I will allow you in and we can chat.

Ulrik Lehrskov-Schmidt3:05

But otherwise, so the way it's it's gonna it's gonna work is that we have the question, so we have the chat, and then we have some of the pre-questions that you and others sent in. So we're gonna go through those and then we're gonna monitor a little bit sort of how how it's worked with sort of like you get to sort of like ask follow-ups and so forth. But but if we can get people in with voice, that's also you know not bad. So so we're gonna try. as you can tell both Rob and I we are experts at Riverside. So we're actually we're we're at least we're expert learners, so we're trying to navigate it. Okay.

Rob Litterst3:45

Exactly. awesome. Why don't we why don't we kick it off? And Rodika, since we've been talking with you, we can kick it off with your question. so I'll read it here and feel free if for now you want to just chime in from the chat. That's totally cool. If you see the ability to jump in live and want to request that, what whenever I see it, I will let you in. But I don't see anything yet. So Rodika's question we are introducing credits for some of our consumptive AI products, and just to kind of like set the stage. Rodika works at LinkedIn, so everybody kind of understands what the product is. We are introducing credits for some of our consumptive AI products. Is it common practice to attach a rate card to the order form explaining how many credits each unit of output costs? We have approximately ten different types of outputs generated by this AI product. And Ulrich, I will let you take it from there.

Ulrik Lehrskov-Schmidt4:38

Yeah, so so common practice. may attaching it to an order form, I think that's optional, but but disclosing it somehow, yes. I would say that what you want to be careful of is disclosing it in a way where it gains any form of permanence, right? Like writing it in the contract or having it as a sort of sort of an attachment where where people feel that these are the prices. This is the rate card. So I would prefer a model where You direct people to where they can find the at any time, like sort of the the rate card that is currently sort of in effect. And then you get to change that rate card whenever you want, right? So so one of the good things about credit models is that you have these sort of two ratios: you have dollar to credit, so people with their dollars buy your credits. And then inside of your product, you have the credit-to-product credit. relationship which is sort of the rate card like 10 credits for whatever and edit, 2 credits for that thing, and so forth. So the thing that you contract, hey, there we have your big the thing that you contract, that should be that has to be stable because people want to know how many dollars they give you and how many credits they get. But the but the credit to product relationship can and should be volatile. So you get to introduce new line items in the rate card.

Rob Litterst5:47

We figured it out. There we go.

Ulrik Lehrskov-Schmidt6:06

You get to pull them out, you get to change the number of credits that something cost and so forth, depending on whatever mood you're in that day. That's the sort of ideal because this is what allows you to have the flexibility on that end. And this is sort of one of the great advantages of credit models.

Radhika Kanuga6:23

Got it. Thank you. Ulric. I I totally so we are ultimately doing what you said. The question I posted was some weeks ago, but we're essentially adding a link that, you know, whoever's contracting with us can go to and they'll see what the rate card looks like per unit of output, how many credits per unit of output. And I understand the point around adding additional units of outfit because we're gonna keep doing that, which is why we wanted the link rather than to send it to an order put it into an order form. But the thing that I'm I struggle with is i how can you change the amount of credits something costs after the customer has signed the contract?

Ulrik Lehrskov-Schmidt7:06

You change the number on the website. Like I I know that it sounds like there is no spoon kind of thing, but but it's kind of I I think it it is that simple. You j you just you you don't you don't put it in the contract and you refer to it and then you go to your rate card and you change the prices. That's it. Yeah.

Radhika Kanuga7:23

Do other companies actually do this? Can you give me examples?

Ulrik Lehrskov-Schmidt7:28

I it's almost easier to give you examples of people that don't do this, right? So so so so if you go to so so the way that that that any of the large LMs will do this, like they they they will notify you that they're changing it, but they will usually change it within like effective immediately. so so I think and

Ulrik Lehrskov-Schmidt7:55

Clay does this, for example, just to give you an example. I think I think they have the ability to push a pricing change within a few hours. So whenever, like say Claude or OpenAI, they change their pricing model, Clay will usually change it within a few hours. OpenRouter does the same thing. and so on. So so I think it's what you're dealing with more likely is some form of let's say

Ulrik Lehrskov-Schmidt8:24

Legacy thinking either in the heads of your sales team or in the heads of your customers, where they are still uncomfortable with the fact that AI pricing is is is more changeable than traditional sort of license-based SaaS pricing. And your fundamental argument for this is that is a few things. One, it has to be because the fundamental economics and cost models are changing. So we we we cannot be caught in the middle.

Ulrik Lehrskov-Schmidt8:54

and two, which is more esoteric, but that is your overall narrative from a product perspective, which is say, regardless of whatever the technology will do tomorrow, we at LinkedIn are committed to bringing you these business outcomes. And we will always make sort of the right product and commercial decisions to do that. And fundamentally, if we if we create something that is unsustainable for you and the rest of our customers, you will not do business with us for long. So So it is like it is not in our or your interest to do so so essentially this sort of north star vision mission kind of directive that that lets you always explain why you're changing a price. Like why are you changing it from ten 10 credits to 20? Well, because these guys did that and also we want to keep delivering the best outcome for you. So so if so maybe you have like underlying

Radhika Kanuga9:46

What do you mean when these guys did that?

Ulrik Lehrskov-Schmidt9:51

So maybe maybe it can be from anything from underlying model providers are changing their pricing to something else is happening in the market to to the product got better and so forth, right? So so I think it's most customers can can accept any how as a child price increase if they understand if they fundamentally understand and accept the why.

Ulrik Lehrskov-Schmidt10:16

So so the narrative is actually like a really strong guardrail that you should have around your your your pricing and and product roadmap.

Radhika Kanuga10:26

Okay. Helpful. It's gonna be it's gonna take a lot of changing of a lot of minds at our end to get to that place, but it's really helpful to have your point perspective on this. Thanks, Ulrich. Okay. Appreciate that.

Ulrik Lehrskov-Schmidt10:36

Start small, it will accelerate. Yeah.

Rob Litterst10:42

Awesome. Thanks for Dika. okay, so up next we have Megana. I'm sorry if I didn't pronounce that perfectly, but I will read Megana's question. And if you want to try to I think it's a request to call in, Megana. If if you see that button anywhere within Riverside, feel free to call in and I'll I'll let you jump in here. But I'll read your question first and we can get started on that. So Megana's question is what are options For the credit migration process, how do you plan and decide which is best for existing customers? What analysis supports and how long does the process take? And Ulrich, Megana works at Bird Eye, which I believe is Bird Eye Marketing Solutions. lots of different products in here. I'll share the I'll share the URL in chat.

Ulrik Lehrskov-Schmidt11:35

So So what are the options for the credit migration process? So so I'm taking this to understand, Megana, that that you're like you have it, you have whatever, another pricing model, and now you want to either introduce credits as well, or instead of like those that can be a little bit different scenarios. But there you are. Hey. So maybe give some context to what you're doing.

Meghana Gudur12:01

Yeah, for sure. So I would say like currently a lot of the products in the portfolio, they have a ton of products, are charged sort of like location based. So there are customers who, you know, are sort of power users who use it, you know, way more than others. So I would say like some of the customers are like in the restaurant industry, they get a lot more reviews or have a lot more online visibility versus let's say like a legal firm that we're selling to. So the thought is that to sort of streamline like the different ways value is delivered across these products, that maybe we can go to A credit model. But I think you know, existing customers are very used to location-based. It's a very competitive space where there are a lot of other companies that sort of have stuck to location-based. But I think we're trying to find a way to sort of monetize the more like agentic AI solution. So I guess just trying to think through like how do you do this migration process with like the least friction and you know churn risk to the existing customer base.

Ulrik Lehrskov-Schmidt12:57

So it's not just the AI products, it's actually also the let's say regular outcomes that the product already delivers. Yeah. Okay.

Ulrik Lehrskov-Schmidt13:10

Is it okay I use your case a little bit like an example to explain like a broader point here? Yeah. So so a lot of companies have an existing pricing model that prices relatively flatly. Like let's say per location or per seat is also like a very traditional model. And then The value is actually not tracked really well by this metric, right? So even though you track per location, some locations are just like way better than others. Like they they get like a million reviews and others get like whatever, a hundred. Or if it's seats, some customers just get a ton of value out of the seats they have, others get a little bit less. So so you have this kind of like the the metric isn't there's like this hockey stick distribution of value, and whatever metric you have is kind of a little bit flat, so it doesn't capture sort of the the the tip of the hockey hockey stick in the value. So you need to sort of come up with another way of measuring value that sort of better tracks that the actual value that is delivered in the account. So maybe you want to track reviews instead of locations. Or maybe you'd want to track ideally something like revenue, or maybe revenue generated by the marketing, whatever. Like you could you could sort of imagine something that would like really capture the value of your solution. Maybe you can't measure it. Maybe, maybe it's it's sort of something you can't agree on the customer with, maybe it's hard to define. There can be all these sort of operational issues with measuring value, but you can at least sort of imagine that there might be better ways than just counting locations. Okay. So So so so that's sort of step one, is just to say, okay, so sometimes you you want to go to to have a different metric that is that is better capturing the value. A credit system is usually good at pulling some cash up front for a consumption that then comes over time and maybe is also very variable over time. It might be seasonal, it might be project-based, it might so it's not smooth, right?

Ulrik Lehrskov-Schmidt15:04

And also, you get the cash on day one, not on day 365, like at the end of the year. So that's what the credit model is. And also the credit model allows you fungibility. So it allows you sort of one pool of credits that can then be expensed across 10 different outcomes or whatever, like a lot of different outcomes. So your credit model is usually good at sort of having A transitioning from a more license-based model where you get the cash up front. So you say, hey, credits also get the cash up front. So that is that is that is then more similar to a license-based model compared to a usage-based model where you're paid in arrea. So that's one of the benefits. And also you have you have all the different metrics, like reviews and let's say whatever footfall or whatever. You might have a lot of different things. So that's where you have it. What you'll usually do is you can you can you can you can come up with all these different metrics that would better track value and that you can reliably measure and that like that you believe in, let's say, then maybe that reviews. And then then you'll do sort of you'll you'll you'll track how many, let's say, each restaurant or or or client does. And then you can just rank or it from high to low, like most reviews to least. And then probably some of your locations are gonna have zero reviews. Whatever, like they they just don't, and others will have like a ton. So you probably have some sort of 80-20 distribution where 20% of the locations have eighty percent of the reviews. So you can then sort of thin slice this. You can say, okay, so you can have like 99th percentile, 88th percentile, 80 like 95th percentile, 90th, 80s, and so forth. And then you can say, okay, so so you could you could include a model such as we price per location, and then you get Whatever, and then you mention a number like 10,000 reviews for free. And then 10,000 is picked because you know that let's say none of your customers, or 99% of your customers, or 95% of your customers, whatever, like percentile customers, are untouched by it. So you introduce the the bar is so high that almost no customers touch it, and the only ones that do are the ones that get tremendous value out of the product.

Ulrik Lehrskov-Schmidt17:20

This is especially useful if you have some form of underlying metric that is growing, such as maybe your AI consumption. Right? So if you're putting in AI consumption in a product and you're seeing that people are starting to use it, you can say, hey, AI consumption is free up until whatever, a thousand actions, because nobody is using a thousand actions yet. But you can see from the behavior that they will soon. So you actually put the pricing in at a point where nobody is priced. But then the behavior and the product will get them there over time. So it's kind of it's like a strategy, like I called it like boiling the frog, like you turn up the heat really, really slowly. But you do it in a way where it isn't sort of a shock to the ecosystem. You can sort of put in the pricing and then make zero dollars, but the mechanism is there, people know that it's there and then they go into it. So it like not the same, but a similar mechanism is when Gmail came around, it was free, and you had whatever it was, like 10 gigabytes of storage. And then nobody who at that point used 10 gigabytes of storage for the email. Nobody. And then everybody did at some point. And I think now it's like 90% of users pay for Gmail because everybody needs more storage. So they kind of grow into it over time. So Figma did this as well with their credits. They said, hey, we now have AI. Now we're going to measure AI based on these metrics with these credits, but credits are not priced. And then at a certain point, they actually put in thresholds. And they started to enforce the thresholds and sell you additional credits above those thresholds, and they were pretty high. And then as consumption grew, you can see that the vast majority of credit consumption revenue that Figma has sits in a very few of their total account base. So it's only the few accounts that spend a lot of credits that actually pay for that whole thing. So it's like instead of thinking about a credit system and something that you like, now we need to make money on all our customers. Think of it as a way to only monetize like the top five percent of customers that are the ones that are getting disproportionate value out of the core product.

Meghana Gudur19:24

Yeah, that makes sense as sort of like a platform of like included usage that has enough for most customers and then only

Ulrik Lehrskov-Schmidt19:29

Yeah. And then you can you can just like either either the product takes them above that threshold over time or you lower the threshold gradually, but you do it in a way where customers get customers need to get used to just the fact that things are counted first and that it is a metric that they need to be aware of, and then at a certain point like there is a paywall essentially or or it gets drawn out of their account. So so gradually is especially if you have an existing revenue base and you're maybe you don't want to rock the boat and nobody likes, you know, you know, to to sort of shake the tree. So the doing it in this way can be can be very effective and actually doesn't need to take a lot of time. Like you you can you can you can go from no monetization to like pretty extensive monetization in like nine months. but but start by monetizing maybe nobody, then one percent and then You could you can sort of figure out where the right level is.

Meghana Gudur20:25

Got it. Okay, that's helpful. Thank you.

Rob Litterst20:29

Awesome. Thanks, Megana. next up we have James. Okay, James, let me find your question. So James Lally, who runs pricing at Intralinks. I'm gonna read your question, James. In the meantime, if you want to try to do like the the the call in, that would be awesome. Would love to have you on the screen as well. So James's question, are we seeing the pendulum swing back away from AI credits and towards something else? If one is starting towards AI credits, should you persist or should you pivot and shift towards licenses, fees, caps? How are folks solving for predictability?

Ulrik Lehrskov-Schmidt21:09

Okay, so I don't know if you've been able to join James, but so so you actually have a few questions like wrapped up into sort of one question here. So the first thing, are we seeing the pendulum swing back away from AI credits and towards something else? maybe we have you on here now. Hey James.

Ulrik Lehrskov-Schmidt21:30

Context, like what are you trying to get done here?

James Lally21:33

sure. So just sensitive to what's going on in in you know, the industry of, you know, people blowing through you know, whether it's tokens or credit limits, and then putting in you know, limits. So whether that's stuff that we hear from, you know, I think PWC, it's like every employee gets fifty bucks a day or, you know, however it is people are trying to manage the cost of AI utilization. has the pendulum swung too far with complete consumption of somebody who's just going to complete credits or tokens, whatever the variation is. and now there's a demand to like hey look, I have no idea how many credits I'm going to use. And so therefore I need to have a way for me to manage that. So I need to be able to say that James only gets a hundred tokens a day or, you know, whatever versus Ulrich you get a thousand because of, you know, capabilities or you know, utilization, etcetera.

Ulrik Lehrskov-Schmidt22:32

Yeah, yeah, all that. So I think has it gone too far, will it swing back? I think so of all the different sort of pricing modalities like flat fees, like licenses, user space, credit, credits are like growing really fast. Like just we monitor like a bunch of pricing pages through pricing SaaS. So is it is it gonna swing back? I think more people are gonna People are going to learn what it is good for and what it is not good for, right? So I think right now a lot of people are doing credits because it's sort of in fashion and also it it matches costs really well. And then what we're going to find probably is that a lot of products have actually turns out to have really predictable consumption sort of measurements and actually not. And not sort of super variable value. Like it's not one customer is getting whatever, a billion worth of value, and another one is getting like $200 worth of value. So if value is sort of the same and and you might have a few cost spikes, it actually makes sense to have a more of a f like a license-based or flat pricing model. You just say, hey, we're just gonna do that. And then you just control the costs instead of having a credit model to to to to monetize it, you just control it with fair usage. So I think there are some of these sort of product models and value yeah sort of sort of value templates that that are that are better monetized and managed by combination of simple pricing models with with fair usage to control for the costs. That being said, for the ones that does have like this I talked about this sort of value hockey stick where some customers get a lot of value and others don't, you you usually need some form of At least metered pricing, that is like a price per something that that where some customers really can't spend a million times more than another. And whether that is a user-based pricing or a credit-based pricing, that sort of depends on just when you want you to have the cash. So so you for both of those models, the the problem with credit pricing is that the thing that credits often price, the actions that the NC, like that technology stack is still unknown.

Ulrik Lehrskov-Schmidt24:48

And people are discovering the products and the products are changing really fast. So it's more that we don't have we don't have a lot of experience with it yet. Like if you're pricing, let's say your your electricity company also charges you usage based. But you have a pretty good handle of how much electricity you're gonna use and like it's it's a it's a known phenomenon, so you're not worried about it, right? So I think I think a lot of products are sort of their electricity but they don't Nobody knows yet, right? So so it just it just haven't been discovered. So so because it's the unknown factor, not the actual variants in the consumption. So so this is where we then see, I think across the board for a lot of projects we do and also in the market in general, we see a lot of people spending a lot of product resources generating sort of a primary guardrail around measurement, control, and optimization. So this is step one. Tell people how many credits they're spending on what, make it create an audit trail, like and have it like a measurable unit that people can agree on. Like like it sounds like basic things, but like still a lot of companies essentially don't tell you how many credits you have. Like it you get an email maybe once a month and it's outdated and you have no audit trail and so forth. So but being able to, as a CFO, let's say, or or whatever, like a buyer. To be able to drill down and say where am I spending my credits, what am I spending them on, who is spending them, and being able to sort of estimate whether I'm getting value from that is a good first step. Next step is then to be able to put in place controls. Mike gets 100 credits, Sarah gets also 100 credits, Mike can approve more credits, Sarah cannot. They can only use them on these products, whatever it is. And then The final step, optimize, which which only a very few companies have gone to yet, is to sort of build in sort of modeling and simulation and recommendations in the product on when to spend the credits on what.

Ulrik Lehrskov-Schmidt26:52

So the best analogy I have is if you if you've ever been in a sales meeting, you probably like met a salesperson that has this little calculator where it's like how many employees do you have? And then you're gonna spend, you're gonna say like 20 minutes per employee. So if you buy our product, you make like X dollars, like these kinds of things. It's it's it's the same logic just built into your credit model saying, hey, if you spend like all these credits on running these actions. We will generate these outcomes for you that will generate this many sales or this many, whatever other things, right? So the credit model's ability to tie itself to value and build little ROI-type scenarios for you so that you get to spend your budget on the thing that moves the needle the most for whatever your business case is. is also starting to become for for some scenarios, Snowflake does this, for example, for some scenarios like a like it like an integrated part of that sort of control layer inside the credit system. So I think that will, in my opinion, get more and more sophisticated. And I think it's it's a it's almost like an it it's a part of the product track that has been Like a lot of CPOs have been like annoyed with it as sort of, I have to spend roadmap time with it. Like I why I I want to build like the sexy product features. Why do I have to build this sort of boring like infrastructure financial stuff? But I think a lot of people realizing that, this is actually where the money is. Like this is this is what allows us to take that estimation risk that sits in the sales process and hand it to the customer. Say, I don't know either how many credits you're gonna spend, but If you buy a million, then I'm now giving you all this tooling so that you can own that risk. Like you get to track it, you get to control it, you get to optimize for it. So if by giving you this tool set, hopefully the customer is now okay with with owning that risk. And that also means that now we don't have to account for it in the contract. I don't have to discount for it. I don't have to put in caps, I don't have to put like all these sort of contract mechanics can be left like alone.

Ulrik Lehrskov-Schmidt28:55

Because the product sort of takes care of the risk for the customer.

Rob Litterst29:06

Awesome. Up next we have Jordan from EverPro Solutions. So Jordan's question, where do you start? The product, the billing system, customer research, leadership buy in. juicy question. Jordan, if you want to try to get in by calling in, we can we can get you up here. Otherwise, feel free to to use chat. there we go.

Ulrik Lehrskov-Schmidt29:34

Hey, I sounds like you're considering a credit system.

Jordan Lieberman29:41

We have a lot of product talked about this ages and can't get off hence the root of my question.

Rob Litterst29:50

Jordan, your audio's coming in a little it's it's like it's a little slow and choppy.

Ulrik Lehrskov-Schmidt30:00

Okay. So I think so if I if I just I just have to sort of take what what you what you gave me there saying, hey, so where do you start? Like so you're cons if you're it depends a little bit of like have you committed to a credit system or are you just considering it? But if you are so depending a little bit on where we are, I think I think the the reason for doing a credit system in the first place should be value is tied to maybe multiple metrics or a single metric that has That lacks some sort of smoothness. Like it has seasonality, it has project base. So we don't really know when and how that sort of consumption is gonna be. So by having a credit system, you can sort of smooth out the cash flow, you can collect up front. Or if you have multiple different metrics, people get to buy one thing and then they they just need to estimate and pay for that one thing instead of having to estimate like 10 different things. So credit systems are good. To handle that kind of complexity either in the product or in the consumption pattern of both. They they tend to so, really simple example: Audible audiobooks, Amazon's audiobooks, they have a credit system. It's one of the sort of the the the the original sort of B2C credit systems. And the reason it works is that consumption of audiobooks is super unstable. They want to put it on a subscription and But I'm not like, you know, I don't have a schedule that I'm listening to one audiobook a week. Like I'm gonna listen to like seven in my holiday and then none for the rest of like the time, right? So in order to take that instability of consumption and sort of put it into a sort of smooth curve, I buy credits every month and then like the consumption like smooth out. So that's that's how it works. So you kind of have to start with the customer. We're saying, okay, so it's a consumption pattern like unsmooth.

Ulrik Lehrskov-Schmidt32:01

Or do we have all this complexity that we want to solve for? And do we have some untapped value, like a spiky hockey stick of uneven distribution of value? Then a credit system might be the right thing for us, right? And and and this sort of, I think this then takes you to sort of to to the product. And then so I would say what I would then do is that. So now you're saying leadership buying, billing system, and so forth. I'm saying nothing gets done without leadership buying. But I think it's not necessarily. I don't think that leadership buys into a credit system per se. They they buy into a credit system as a solution to another problem. So you kind of have to buy the leadership into like what is it that we're trying to do here from a monetization point of view, from a like growth point of view, from a from a market expansion point of view, whatever it is. And then The credit system has to be the answer to the problem. Like it isn't it isn't sort of what they're buying into. So it's more like, hey, we need to monetize all of this AI and we need to mar have to margins not go below whatever 60%. Like that's the problem. And then credit systems, they do both of those things. And also we might get all these other sort of value capture benefits out of it. Okay, awesome. Let's go. Like th I think that's sort of the sequence that you need to sort of run it in. So so so I think I think that makes sense. The billing system is usually and sort of every every other tooling, which actually includes other things as sales compensation, like telemetry, like all these adjacent CPQ tooling. I I've had pretty large projects like hundred million AR plus companies suddenly like press pause for several months because they felt that their CPQ tooling wasn't like built for credit systems. So I think My approach to this is that you kind of have to push through this and and keep things together with like duct tape and scaffolding to get to some because if you if you wait until every system is like ready and you know re live live, you you can spend months or even years getting to that point. So you have to say, hey, we're just gonna run this in a spreadsheet for like a month.

Ulrik Lehrskov-Schmidt34:23

And then see how it works, and then we can sort of we can we can do proper tooling after. So let's run the sales test, let's just run the like finance can can we'll have to do this manually for the first whatever hundred accounts, whatever needs to actually happen so that you can move forward. I think there is an excessive need to automate and streamline things just to begin. as opposed to beginning and then later on automating and streamlining things. And of course it's you know you have to there there is some there is some balancing that needs to be done here. But but I think sometimes people just never consider that they could manually like carry or bridge something as as just to to move a project forward. So so a really good example is I think a lot of telemetry is Is enough for you to start a credit system, decide that it is the right thing to do, but it might not be enough to optimize it once it's running. So you can use sort of your current telemetry and like do manual data polls, analyze customers and all these things, and arrive at the conclusion that the credit system and consumption are building it right. And then you can put it into the market and start to sell it. It might be that you need to do manual data pools to actually do your invoicing for a while, and that is heavy. But then of course you can catch up until it's like automated and real time. It just flows into your like quote to cash flow and then go there, right? But but sometimes just like get it getting to the point where you actually are in the market with revenue as quickly as possible and then like like downstreaming all the the tool stack is is often like a better way to execute.

Ulrik Lehrskov-Schmidt36:14

So start with problems first, get the management buy in, understand the value from the custom scenario, launch however it takes to build the tooling later. Rough outline. Like your mileage may vary, but that's sort of the rough outline. Yeah.

Rob Litterst36:33

Thanks, Jordan. awesome. Next we have Sana from Blinks. And Sana, if you want to try to dial in, we'll let you in up here. Her question is how to handle use cases with huge price and value differences in one model. So for example, action X costs ten credits, but Action Z one thousand. I'm excited to hear you talk about this, Ulrich. just given the content in the new book. There she is, I'll let her in too in case she wants to drop more context. Hey Sana. Yep.

Sanne van Deudekom37:10

Hi, can you hear me? Yeah. yeah, so maybe to to elaborate, we created a tiered subscription model with different tiers with buckets of credits mainly for our assistive agents.

Sanne van Deudekom37:22

And of course, the higher tier you commit to, the lower your average price per credit is. That works very well to incentivize commitment, right? Up front. and it worked very well where we had this assistive agent where which would cost like one or two credits per X. But now all of a sudden we have these more agentic agents that actually cost a hundred or two hundred per X. And if I had to add these to the model, the model wouldn't work anymore because you would give discounts way too early, and these people that are going

Sanne van Deudekom37:52

To use these agents directly burn through the highest tier, right? Because if you if you just use it like three three or four times, you directly burn to the highest credit tier. So now we basically have two models next to each other. A few of our agents are in this tiered subscription model, and then we have a few things that are just separately being monetized right now. So then my question is: how do you best combine these different types of agents with different heights of value into one credit model?

Ulrik Lehrskov-Schmidt38:20

So maybe you don't, is sort of the unfortunate answer, right? So so so it sounds like you have you have you have you have invented what I call a multi-pool credit system. So it's not one credit system, it's two, they sit side by side. Notion just launched a multi-pool credit system, Clay has a multi-pool credit system. I've designed several. So

Sanne van Deudekom38:23

Maybe you don't is sort of positive answer, right? So so so so it sounds like you have you have you have you have invented what I call a multi pool credit system.

Ulrik Lehrskov-Schmidt38:47

Usually like why do multiple and just not just do one pool, right? So it actually it has a little bit less to do with the fact that that hey this costs a thousand, this costs two. It usually has more to do with what is the underlying margin of each of the items, right? So maybe so clay without knowing their margins intimately, they have two credit pools. One is for automations and actions, that's very agentic, and then they have one for data enrichment. not so agents, right? So presumably the margins between these two types of actions is very different. So they kind of want to separate them, right? Because one, they might not want to offer a lot of discounts. Maybe the max discount on so I'll give you another example. It's more generic and maybe easy to understand that so sometimes we have We have like all the different sort of product features which we which maybe has like really good margin in it. So here we can actually we can have like a margin curve that goes from 0% to maybe whatever, 60%. So you're gonna offer really good margins for the largest of accounts. But maybe another pool that we want to run on the credit system is all the infrastructure, like virtual CPU, storage, these of things. And that we kind of wanna sell at cost, right? So so basically it's just like, hey, we we want you to be able to be fungible, like fluid with where your spend goes across the infrastructure, but we don't ever want to sell infrastructure at a cost. So it doesn't really matter if you buy a little or a lot of it, like it's gonna be pretty much the same price. Which is why we don't want the other credits to pay for the infrastructure, because there we want the the discount to incentivize the volume. So this is where you might sort of split it up and say, hey, you have like an infrastructure credit pool, which which is where we don't have a lot of margin, and we basically just want to like Give you a nice and easy way to purchase it. And then we have the let's say value-based part of the product where we do have a lot of margin and and and so forth. So this can be like a reason for doing a multiple credit system. Customers, if you explain it well enough and you just you you you yourself

Ulrik Lehrskov-Schmidt40:50

have a clear thinking around what it is and what goes into each pool and why, you're usually also able to articulate it quite well and then it will not pose a problem in the sales process. Like that's it. And you can even have different rollover policies, commit policies, all these kinds of things for the two credit pools, which might also sort of be beneficial to your sales process. So so yeah, so so it's it's it maybe do some soul searching for sure to see if you truly need it. But if you do, don't be afraid to just like operate with with two. Usually do you need more than two? Like, we'll have five credit pools. You know, do some more soul searching. Like that's like that's probably like not right. So but but but two can can definitely have have its reasons.

Sanne van Deudekom41:42

Okay, clear. Thanks. Thanks, Sam. okay, Michael

Rob Litterst41:46

Thanks, Sana. okay, Michael. I think Abby might have left. I think she jumped in and then had something come up. so I'll read your I'll read your question, Michael. I don't really think we need you to jump on. I think it's a good question and and we'll we'll let it rip. And if you want to type any more context in the chat, feel free. So Michael's question, Ulrich, is what is the most creative way you've seen a company package or pass on their token costs?

Ulrik Lehrskov-Schmidt42:25

I'm understanding your question to be, hey, I have I have AI costs. I want my customers to pay for the AI cost. what is the most creative way?

Ulrik Lehrskov-Schmidt42:43

I don't I don't even necessarily think that you need to be creative with this. Like you can just pass it on. Just like, hey, I have the cost, now you pay for it, right? You can of course like It's it's usually when when it it becomes this sort of poison pill that you put inside something else. Like, we're we are we are now increasing all the seat based pricing because of AI and then maybe you increase it more than you need to, right? Or and so forth. So so I think it's it's one of these where you it can become a sidecar to almost anything else. so so I think I think I don't even I don't it's a good question. I'll I'll consider maybe I'll you've got one. Okay. Go for it.

Rob Litterst43:19

I have one. I have like an I have like an anti-answer. and it's fresh because we just did we just did this on the the podcast this week. But linear I think is really interesting. They don't they they pass they pass along token costs straight through with no markup whatsoever and then monetize like a few things kind of like around the token costs, right? So like they have like the linear agent. The linear agent needs access to a virtual computer where it can do its work. So Linear charges you like in 20 minute increments basically for how much the agent needs to use the virtual computer. And if you look at like how much they're charging for the virtual computer, it's more than like list prices for a company that would like sell you that straight up. So like they're they're charging a they're getting a bit of a profit off of the virtual computer. And then they also have these like agent loops functionality, which they charge kind of like They they say the range is between like seven and twenty cents. But that like on that they're say they're not necessarily saying that they're charging with no markup on the token costs, right? So they're clearly capturing some margin there. Like you mentioned it in the podcast, Ulrich, but I really like that strategy because first of all, like if like bring your own key becomes really popular and companies start like bringing Their chat GPT, which I like Chat GPT, I think is like pushing way more into this than Anthropic is. But if that becomes like more the norm, it still gives you something to monetize, right? Cause you have like all of this stuff like outside of the token costs that is actually what you're monetizing. Like you're not even monetizing a margin on top of the token. So I think that's really interesting in and of itself. and also just think it's like been very helpful for them with adoption. Like if you like look at the top line stats on linear from an agent perspective, like ninety-five percent of customers use agents.

Rob Litterst45:13

Like the tickets with like pull requests are up like seven X in the last year. they're like the like number of like work that's created by agents in the platform's gone from like three percent to fifty percent in the lat like since the beginning of this year. so it seems like all that stuff is working pretty well. But sorry, I'll I'll stop talking. You can you can riff.

Ulrik Lehrskov-Schmidt45:33

No, so so so so and and so so okay, shameless plug for like Rob and I we do we do a podcast called Pricing Page Unpacked and which you should of course check out. But but so so on the last one we talked about sort of so what happens when when you do exactly sort of like maybe not a creative but if you if you if you argue that your AI thing costs money because you have AI costs. So we when you basically when you're very when you're trying to sort of say, okay, we'll monetize by passing on costs at a margin, right? So we'll we'll we'll mark up the AI and then that's how we're gonna make money. The problem is of course that that the people are gonna say, well I Totally get that. Like AI costs money. Our AI costs money too. It just so happens that I actually have my own AI. So instead of having your AI do the work, I can just like go through your MCP with my own AI because I have a great deal with anthropic or whatever. And also the AI, my AI knows me really well. So maybe that's even better, right? And then suddenly you're sort of What used to be your monetization model, which was, hey, they're gonna use my product and my AI, I'm gonna like monetize in that way, with the excuse of covering costs, has now been sidelined because people are bringing their own. So we're seeing a lot of companies right now being sort of caught in this argument. It's like they monetized using cost coverage as the narrative. And then because customers can bring their own AI, they don't have any cost to cover, and then they're sort of left without a pricing model, or at least a pricing logic. And now they have to sort of scramble to sort of reinvent or re-argue why their product is valuable when they've just said that they're just doing it to cover cost, right? So so so don't do that mistake, I think. So never

Ulrik Lehrskov-Schmidt47:20

unless it's just a partial like infrastructure part of it. But but for the meat and potatoes of your product, never argue that your pricing model is just covering cost. Like then you truly have to be a foundational utility like electricity, right? But for anyone else it's you you can get like people can just bring their own and and and and they will. So so you'll be left sort of without nothing to stand on. Yes we had that. Alright. Do we have

Rob Litterst47:51

We do. Yep, we got another one. Steve Blank from Files. I'm gonna read your question. There there should be the ability to try to call in. I I think you might have joined a little bit later, Steve, so not sure if if you heard that, but there should be the ability to kind of like call in and if you can if you can find that in the navigation, we'll let you in up here if you want to add some more nuance. But I'll read your question in the meantime. So Steve's question is how much non-AI related functionality.

Rob Litterst48:19

can or should be considered in the credit value.

Rob Litterst48:25

And he's coming in real quick.

Ulrik Lehrskov-Schmidt48:26

So yeah. I think I'm lying with this question, Steve, but but let's let's let maybe unpack it a little bit. Like in the credit value. So so that's the part where where it's a little unclear for me. What do you mean? Yeah, I do.

Steve Blanck48:39

Can you hear me? Yeah, so we have thought about this a little bit just about consumption in general. So talking to your point, I'm sorry all the bikes back here, use Rambler. yeah, like we have thought about this as consumption. And AI is part of that story. So some of this is like you've talked about MCP and you've talked about just agents that do jobs. Not all of our work is necessarily LLM-based. It's some is deterministic, some of it is just like we do a job for you. And it's bas and we feel I've felt like that's the same charging metric. But our cost bases are different. The thesis is it's simpler to have one pool of credits.

Steve Blanck49:23

I haven't seen a lot of market signal that, and we're early in this perhaps, but my question is like, my question is simply though is like, is this so against the grain that we should be considering other models? Or is this like a reasonable way, say you, Mr. Customer, can have this access to this whole host of things. we deplete credits based on some are cost plus because of LLM, some are because of what we think it's the value passed on, et cetera.

Steve Blanck49:51

It's sort of I'm I'm looking for like a permission structure in the absence of being able to find like competitive dynamics that support or or reject that idea.

Ulrik Lehrskov-Schmidt50:01

So credit systems were a thing way before AI. Right? So so so that's just like so let's start there. Like the in the nineteen nineties, Brad what were we called? Brad and Dunhill, I can't remember, like there there like a marketing they sold market data on a credit system, billion dollar business. exact it ended with it it ended catastrophically with there's a story around a lawsuit where So which is like maybe good for everybody to sort of read up on in that the the salespeople would overcommit their customers and then they would lie about how many credits they they had left and so so people would buy like basically like inflate their their salaries by selling people like a hundred thousand worth of credits and they only needed twenty thousand and and so forth. So it got tested in the course and and it kind of you know it it backfired quite a bit. Anyways. When I was doing credit systems for customers, let's say five years ago, one of the one of the tricky parts was the people like credits, what is that? Isn't it complicated? How does it work? Like all these things. So you had to sort of you had to have a really strong case when credit was just like the right model for it to really work because you had to overcome this communication barrier of explaining to a customer base what a credit even was and how it worked. That's now gone. Like now everybody uses credits. So so now people are like, let's just use credits because because everybody knows what it is, we can just use it for whatever. Inside a credit system. So think of it this way: it's just a layer on top of which, or inside of which, you can actually run any kind of other pricing model. You can run a flat fee, you can run a license, you can run seat-based licenses inside a credit system. Say, hey, a user costs 100 credits a month. That's it. Why not?

Ulrik Lehrskov-Schmidt51:49

You can run, of course, a consumption-based model. Hey, whatever, one credit per API call, and then we'll just, you know, we'll take it off your account whenever you do the thing. So all the other modalities, flat fees, licenses, users-based, actually runs inside a credit system. And then, of course, you also get to like the metric, whatever it is that you're counting, doesn't need to be a Like an AI cost-related item. It can be anything, such as, hey, you're setting up a workflow, or it's per step for workflow that is set up, or it is per thing that runs through the workflow, or it is for a user, or it is for an API call, or it's for however many V CPU hours you're doing, whatever it is, right? So feel free. What I would say is it can be a really good idea to have in your credit system not the things that not just the things that cost you money, but also the things that are actually valuable to customers. And because credit systems have been so like they became popular because they were a great way to frontload cash so that you could cover cost. So AI has cost, I need to charge up front, and then so so it worked. Like so that is the sort of not the original, but that is the recent use case. People kind of forgot that, there are things in our product like the deterministic parts that just are valuable to customers. So let's charge for those too. So you have my full blessing to just go ahead and do all of that. It sounds great.

Steve Blanck53:17

permission seeker by nature. Yeah. Okay. But no, that is refreshing to hear though. Because I don't think it's illogical, but it's been hard to it's been hard to pin down because if you look up like if you research MCP consumption pricing, they just the best you can get is people are charging on an API call basis. And we don't view that as the relevant charging or or reference price. It's just something that a competitor may do. But to us, and and whatever for what files does or what our our company does, like it's really about being a companion product to a system of record or something like that. And in that sense, like, you know, we're working through a lot of different scenarios, both good and bad, of what this enables, but also what it constrains in our seat model, which is why we're very like hybrid focused, but also why I'm trying to figure out pricing but also like recourse and s and broader strategy based on how a consumer one consumer uses it in a traditional way and one consumer may figure out another way to use it that changes the balance between the seat and the variable piece. And I think that's so all to say that that's why my question has so much importance right now is we're trying to plan today, but also whatever the heck happens in the next year or so or more. So

Ulrik Lehrskov-Schmidt54:29

So go so what what I'll suggest that you do, and this is gotta sort of goes a little bit for everyone, is that when when you're thinking about metrics, like what to count. So I I divide metrics into two buckets, and each bucket has two kinds. So metrics I have measurement metrics, so these are things that you measure, and they have two kinds, value and cost. So you measure different things that cost things, and you measure different things that have value. And you usually should measure like a lot of things that cost and a lot of things that has value. And then out of the pool of things that cost, that has cost and things that have value, you then choose which deployment metrics you want. Like those are the ones that you actually use in your commercial model. And those are either pricing metrics, so those are things that you decide to charge money for. So they say, hey, I have a cost metric, I want it covered, it becomes a pricing metric. I have something that is just a value, I want it priced, so it becomes a pricing metric. And then you also have control metrics. So these are the things that you run your fair uses on. Not pricing, just fair uses, right? So you kind of have like measurement metrics and deployment metrics. Measurement is like value and cost, and then the other is like pricing and control. So this is sort of how I view the universe of like metrics and what you should measure. And you have way more measurement metrics than you have deployment metrics, right? Usually. So kind of keep like a catalog of things you measure just to keep a track on it. And sometimes you have like five things that could be a pricing metric. If they all bring value, but you say, hey, if we give this one away for free, this other one will grow because it like sits later in a use case. So we kinda wanna like encouraged consumption of this thing because we know it will lead to this other thing. So you basically choose the pricing metric that's sort of at the end of that chain. And then the other things you just you know they're valuable, you measure them, but you actually don't price them, right? So you can have these kinds of considerations when you're choosing your pricing metrics, but you should measure all of it.

Steve Blanck56:27

Yeah, that last piece is key. and I'll I'll let the rest of the folks know. But that is the biggest part for us is that we have this, I feel like we have this indifference between we don't really care how they use. There's a block of activities. I've kind of thought about it like gym passes, is like we have we used to sell gym passes and now like we actually know the reps and weights that each individual person uses. We don't care how they're distributed, but from your fair use point, we can measure it. Once it becomes imbalanced or b in balanced in a way that favors one or the other, we want to make sure we have recourse. course isn't the right word, but we have a strategy in place to make sure we monetize on either side of the equation. And that's been the biggest hasn't happened yet, but we can see which way the wind's blowing. So anyway, that's really helpful. caller, I really appreciate the the insights there.

Ulrik Lehrskov-Schmidt57:11

Great. All right. Cheers. Thanks. Right, Rob.

Rob Litterst57:16

Awesome. I think that's all we have time for, man. I think yeah, that was that was great, Steve. so yeah, I think we can probably let everybody go a couple a couple minutes early. Just ca I feel like it'll be really hard to keep keep an answer at time. unless you want to do one more, Ulrich, what do you think?

Ulrik Lehrskov-Schmidt57:37

So let's do one more. So pe people dropped

Rob Litterst57:39

I know I was gonna I'm I'm trying to spare you because it because it's late where you are right now. Okay, there you go. I love it. Okay, cool. Let me see real quick. I think if we have Dan still on here, Dan Bolkowski had a good question. Dan, I'm gonna find yours and if you want to try to call in. and this is very relevant to me and Jordan. We were just chatting in the in the thread.

Ulrik Lehrskov-Schmidt57:44

No, it's okay. I this is not my last call for today. So it's fine.

Rob Litterst58:05

Is there ever a right time for creating multiple credit types? Actually, that's not the same thing as what we were chatting about, but yeah, ever, ever a good time to have multiple credit types.

Ulrik Lehrskov-Schmidt58:17

So I think this is sort of what I what I talked about a little bit earlier, like the the what what what Notion is doing and what and what Clay is doing with this sort of multi-pool credit system that I'm talking about. But but Dan, do you have like a specific problem you're trying to solve?

Dan Balcauski58:30

yeah, I mean I think you already mentioned it earlier in the call or at least maybe Rob did. so I know, you know, clay they have enrichment versus agenc types. the other one I was looking at the other day, I think Replit has build versus run credits. and so credit models are confusing enough. and I feel like as soon as you're like yeah, I I've been trying to use this analogy. It's like if you went to the gas station, it's like, do you want going to the grocery store gas or do you want family road trip gas? And you have to figure that out, like when you're filling up at the gas station. I feel like we're making customers do that. And that feels insane to me. So I'm trying to understand your perspective if if you think that's ever warranted, given the complexity and what you're asking of the customer.

Ulrik Lehrskov-Schmidt59:22

Yeah, so so so I mean you can always like doctrine analogy to do what you want. Like you have to have gas in the car to do the thing, but you also need to have the snacks, otherwise the kids will whatever, like things will go all right, right? So so you like it n one resource isn't enough. I think that's that's my point, right? But so I think Try to keep it as simple as possible, not simpler. Like that's sort of the original OG advice from Einstein, right? I think I think it sort of also applies to credit systems, is that if you can if you can do one credit pool, do that. But if you have fundamentally different margin structures or consumption structures and they're just like off from each other and you wanna be able to discount one a lot, but keep like really stable price points in the others, splitting it up is actually easier. So here's sort of the rule of thumb that I would use is that. Does does it make everything easier to have two credit pools? Like does does it just solve like a whole host of problems for you? Because then the communication to the customer, hopefully, if you're if you're really if you're really clear about why you have the two pools, like hey, they have different margins, like they have different consumption patterns, whatever, then you should actually be able to narrate that to customers. Say, hey, this is gas and this is snacks. Like they're two different things. Like you need both, but you know. Yeah, like so so I think I think that can become really good like like yeah rule of thumb for for when to do that. But if if but but if you if you can't really articulate that well, that will also spill over in your sales process and sort of the value communication around your product and and yeah, so bad things will happen, right? So so I think sort of the first step is just like caref clarifying for yourself what it is that you're actually what the what the source of friction really is for why you even considering to have two pools.

Rob Litterst61:11

Yeah. I think I I was just gonna say like I think like clay it makes sense 'cause like data credits and like actions and orchestration are like two different things, I think, and like it's it's pretty easy to kinda like differentiate between those two. I like the thing with Notion is like I think their differentiator is like who chooses what model to use. Like that like that's the That's like the difference between their two pools of credits. But then they have a third pool of credit that can fulfill either of those two. Like that that just doesn't really make sense to me because if you have this one pool that can fill both, like why not just do one? And also like who who chooses what model just doesn't seem like a clear enough reason to have two different pools to me.

Dan Balcauski61:53

Yeah. I think on clay, clay's rationale landed for me in as much as I, you know, looked at it. in as much as like data enrichment, I think the argument is like that's a very highly competitive commoditized market, but their agentic actions wasn't weren't.

Dan Balcauski62:15

And so they had a a complete difference in pricing power, but also, you know, it's like, look, if it's too much, customers aren't even gonna consider you at all. And so kind of that market based perspective, I think, makes sense. I I don't understand it in the in the Notion case or in the Replit case. I think you probably could get like if you had one pool, right? If you're talking about like margin profile, so if I think like the Replit build, so it's one set of credits for me to build the app, another set of me for set of credits that you take as my app is being served served and run to users.

Ulrik Lehrskov-Schmidt62:33

No, so i if you look at the two

Dan Balcauski62:50

I don't understand how you that makes less sense to me 'cause I feel like you could just, you know, change the magnitude of of how many credits each cost.

Ulrik Lehrskov-Schmidt62:57

You can. So so if you if you think about it, the the the two underlying price points are gonna change a lot over time, right? So so the build credits is essentially just like whatever underlying LLM is doing to sort of do the the coding, right? And that and the and the and the models will change, the capabilities of the models will change, and the cost profiles of those models will change a lot. Fundamentally what they're trying to do with the run credits is like, hey, we're gonna monetize the value of the thing that you build. Like more s like s it's a proxy for it. So They wanna be able to essentially very closely like follow the underlying costs of their infrastructure providers with one credit type, while they actually want to keep the other w which might go down like quickly over time, while they actually want the other ones to be stable and maybe even potentially go up over time. So if they tie them together, they've kind of taken like they might be able to find a balance right now, but they're gonna work in opposite directions. Right. as they move forward. So I think like right it's actually smart to split them because then they can aggressively drop the price on the build whenever it's necessary to stay competitive, while actually like keeping the the core monetization of the of the run credits sort of intact. And I think it's just it it it removes a lot of like decision headache whenever they they want to move those two prices because they can do it independently of each other.

Ulrik Lehrskov-Schmidt64:22

All right. I think that w was the last question, right? But it was a really good one. Yeah, there we have it. Thanks, Dan.

Rob Litterst64:23

Thanks guys. The real last one. That was that was great. Thanks, Dan. thanks, Ulrich, and thanks everybody, for hanging on. I'll post this recording and you guys can check it out if you want to and appreciate you guys bearing with us as we learn Riverside.

Ulrik Lehrskov-Schmidt64:41

Yeah. Like like once you learn it, it works pretty well. So it's great that we have people on. I I enjoy that part the most. Okay.

Rob Litterst64:45

Yeah. Yeah. Yeah, I'm gonna I'm gonna figure out a better instruction then figure it out on your side and I'll I'll actually tell people what to do next time. So thank you everybody.

Ulrik Lehrskov-Schmidt64:55

Yeah. All right. Take care, guys.