Pricing as Product: How Snowflake Aligns Strategy, Execution, and Customer Value

[00:00:00] PREVIEW: Pricing is downstream from what is your product strategy and what is your company strategy? You know, how do you wanna be perceived? Where do you wanna play? You need alignment across those things. For example, let's say Snowflake wanted to bring in more data. Well then we wanna make it cheaper for you to load more data into the platform.

We may charge you more for other things that you're doing on the platform, but we wanna make it as frictionless as possible to get data into the platform. Having those principles that tie back to your product strategy. It just helps make decision making run faster and... and not get bottlenecked up at the executive level.

[00:00:34] INTRO: Welcome to Unpack Pricing, the show that deconstructs the dark arts of SaaS pricing and packaging. I'm your host, Scott Woody, co-founder and CEO of Metronome. In each episode, you'll learn how the best leaders in tech are turning pricing into a key driver for revenue growth. Let's dive in.

[00:00:58] Scott: All right, awesome. Well, I am super excited to invite Ryan here and he's gonna go super deep on pricing packaging and he works at one of the, kind of, premier companies in the usage-based pricing space at Snowflake. I think there's a number of interesting topics we're gonna dive into. We'll talk a little bit about PE firms, we'll talk a little bit about McKinsey. We'll talk a lot about how finance product and go-to-market, all work together to make sure that your business model is successful. But with that, maybe Ryan, would love for you to just do a quick intro of yourself and then we'll dive in. 

[00:01:33] Ryan: Yeah. Thanks Scott. I'm excited to be here today. So, a little bit of background on me. August will be about three years since I joined Snowflake. I lead our product finance and pricing team here. So thinking about kind of end-to-end finance business support for our product organization. Prior to Snowflake, I did about four and a half, five years at McKinsey. Worked with a number of technology companies and payment companies across go-to-market pricing, and even did a lot of private equity investing work there. And then prior to McKinsey, I did five years in the Air Force and did things like help the government think about satellites and putting different missile systems in space. And so, excited to be here today to chat about my journey to pricing and what we're seeing at Snowflake.

[00:02:20] Scott: Awesome. Well, maybe let's start... I think a very common path that I've seen for pricing experts is they work in, like, a company like McKinsey, or consulting, maybe Simon Kutcher. Maybe talk a little bit about, kind of, your path at McKinsey. I know you've worked on a lot of really interesting projects, but kind of help us understand how, like, McKinsey led to working on pricing and packaging.

[00:02:43] Ryan: Yeah, so I came into McKinsey again, coming out of the military, was a bit of interested in a lot of things. Knew I wanted to work in technology. I was really interested in the growth side of things. And at McKinsey, if you think about, kind of, our go-to-market practice, we had kind of four different pillars of which pricing was one of those pillars.

I did a growth transformation when I was there and worked kind of on a number of different growth topics. And that actually had me build a connection with a partner who ended up pulling me into a private equity portfolio company opportunity where they were thinking about a complete transformation.

And the first phase of that was really focused on all things pricing. And so that was kind of my first foray into pricing. I got the opportunity to work with a number of the experts at McKinsey and really learned kind of from the bottom up all things, everything from designing the pricing model through rolling it out and implementing it, and even writing copy to customers myself.

So, that was my first foray and I learned a ton there when I was at McKinsey. 

[00:03:49] Scott: That's really interesting.I think one of the things that I've realized, I've talked to a number of PE firms, is just how common it is actually 'cause, like in the private equity process to kind of use the private equity umbrella to kind of reset a pricing model or a business model for a business.

Maybe, talk a little bit about kind of in that specific instance, like what were the core elements that you were kind of tasked with taking care of, and then how did that process work? Just like logistically, like what were the core elements of making that transition successful for the business?

[00:04:25] Ryan: Yeah. So, you know, we were brought in by the private equity sponsor, so we had direct access to the management team of the company. And worked directly with a number of folks across their leadership team. You know, I think, the most urgent call to action was they had not raised prices in almost three years.

And so, the board and the management team felt like there was a tremendous amount of value that had been added into the product during that time. And they had not necessarily done a good job of capturing a lot of that value. But as any good consultant would do, we wanted to look exhaustively across all of the different opportunities.

So, we started with, kind of, a deep packaging exercise where we looked at the entire product portfolio. We tried to understand what personas each one of those products was going after. We tried to figure out was there any kind of logical grouping that we could think about in terms of the packaging exercise itself? Did all of those things necessarily fit?

And then following that, we did a bunch of. Market research where we talked to a bunch of experts in the industry. We did a comprehensive survey of over a thousand respondents- mix of customers and potentially new customers. And really tried to quantify not only willingness to pay, but also how folks were perceiving the current kind of pricing metric or value metric that that company had.

And we ended up kind of going through, you know, a diagnostic phase where we tried to like, expose what were all the core problems, where were there opportunities. Ultimately, we decided that we were going to change some of the packaging and we were going to do a pricing increase, as well as change kind of the go forward plan on how they would do future pricing increases which was let's set expectations with customers that we are going to ship and deliver X amount of value over time, and you should expect that you will pay on average, you know, 5% more every year. And let's build that into the model. Let's set expectations that that was true.

We decided to hold off on the pricing metric conversation. And the reason for that was the pricing metric was well understood and it was common in the industry. And we didn't feel like the product had enough net new capability to justify a big change in the pricing metric.

We did a bunch of exhaustive workshops with a bunch of the sales teams and go-to-market teams as well, though, to lay the groundwork to eventually do that in the future. And then we spent really the rest of the time figuring out how are we actually going to roll this out? This was a company that did not have Salesforce or an elaborate ERP system. We're managing pricing sheets in, like, Excel. And so to do kind of a wholesale pricing change and make sure that every customer got charged the right amount at the right time. This was a company that had quite a few different SKUs and different combinations of SKUs. So we really had to think through, not only the system side of it and the mechanics of how do you charge people the right amount, but then in parallel also, the story of why are we doing this? How do we kind of reinforce the value that you've gotten over the last three years, but also the value you're gonna probably get for the rest of this year? And you know, I'm sure we may go dive into this a little bit deeper, but this was not a seamless exercise. There was a lot of internal debate and pushback.

There were salespeople that were trying to make their number that year and were worried about churn. So,it was a really fantastic kind of end-to-end learning experience for me. I mean, it was on a very small scale, but I looked at that, kind of, in hindsight and was like, what a great opportunity to learn and kind of figure out all the nooks and crannies of how to do a wholesale pricing update.

[00:08:12] Scott: Yeah, I mean, I think right now this is the kind of story of Silicon Valley. There's so much disruption being caused by AI in this space. Maybe would love to hear like a little bit of like your... maybe a couple key takeaways from executing that process. And if you were to go back and do it again, what advice would you give yourself? Like how would you approach it? Like what are the kind of key things that you eventually got right that like, you wish you had done early in the process? 

[00:08:43] Ryan: Yeah, so I would say one of my first lessons was really on more of the people side. I found that in the beginning we had almost too many people involved in the process, and it caused a lot of churn. It caused a lot of debate over, not even just the pricing change itself, but also the timing in which you would do that. And as I'm sure you know, these private equity firms are on very, you know, cadence, structured cadences of when things need to be done on a time schedule. So I would've said, you know, Ryan, you probably should have had a smaller group of people to kind of really iterate and refine the, why are we doing this?

What are the real core issues that we can solve today versus things we maybe need to do later on? And then at least kind of starting to put together the first pass of what would we go tell customers around why we are doing this? I think that would've put a lot of people at ease, especially in the sales organization.

And parts of the marketing organization that I feel at the time were nervous about not only hitting their number, but also having to go in front of the customer and explain why this was happening. So I would say, that was kind of my first lesson. Second lesson, which I learned resoundingly and has actually even helped me at a current thing I'm dealing with at Snowflake today, is never change the price without delivering incremental value or new value to customers.

This is something that, you know, when you're so focused internally can feel like the right thing to do in terms of the financial outcome. But it's really critical that you line up that story of, 'Hey, we may increase price, but we are gonna reinforce not only the value we've given you, but also the value that you can expect going forward'. So that would be my second lesson.

And then my third lesson just given, this was kind of my first end-to-end experiences. Man, there's a lot of details in rolling this out and getting it right and those things really matter, right? I mean, the design of the pricing model itself, the pricing level, doing all the research, incredibly intellectually stimulating, really fun to brainstorm and iterate on. But really when you think about what the customer is going to see and the impact that that, you know, change is going to have on that organization, it really is kind of those details at the end that end up leaving a lot of that first impression. And it's a lot of stuff to get done. And so, you know, making sure that you are thinking about that early on, so you don't miss components that could be critical there. But also making sure you leave enough, kind of, energy in the tank to go think those things through and be excited about those things at the end. That was a big lesson for me too, as well. 

[00:11:28] Scott: Yeah. So if I were to summarize the three: One is, kind of, keep the team focused like in the beginning, like, like who's the working team?

Two, like don't do a price change that is like, couple your price change to increased value. And then three, really kind of make sure that you understand that the details matter. One thing you said in there that I definitely have seen is that it's incredibly easy to think of pricing packaging change as almost like a theoretical exercise, but like, what I've found is that the majority of the work definitely in measured in time is actually happens after you've made the decision of what you're moving to.

So maybe actually in this case, maybe talk about just, like, roughly take the full length of the project, how much of the project was spent like in the kind of theoretical, like here's the right pricing model, here's the right measure, here's the right dollar value. And then how much was spent in the like, okay, now let's roll this through the org.

[00:12:26] Ryan: Yeah, so it's funny. At McKinsey, we think of these, kind of pricing work or these transformations in, kind of, three phases. There's a diagnostic phase, there's sort of a bottom-up planning phase, and then there's an implementation and, and rollout phase, right? And I would say in this project, which was probably eight to 10 months long in its entirety, just the pricing work stream itself, I would say that the diagnostic phase, which was like helping us get to the right answer, was probably four to five weeks.

The bottom-up planning was probably another 60 days, so two months or so. And then the implementation was the remaining amount of that time. And yeah, like, exactly like you said, there's so many details to work out. And you know what was also really interesting is not... being sort of in the consulting seat, not only were we worried about the actual success of that rollout and the impact on the customer, but we were also, you know, working with our client to coach and enable and help them feel confident too.

And so seeing them through all of those phases and all of those steps was really important. I almost felt like it was kind of my own extended team that I was working on. But to answer your question, Scott, yeah, it's that long tail of implementation that can really take a lot of the time, but it's those details that are super important to get right.

[00:13:46] Scott: Yeah, I honestly, I think that that breakdown where like, you know, at least half, if not, like three quarters of the work is actually the detail sweating, I think is something that I think most people don't think through or don't understand unless you've been through it. And then if you've been through it, you kind of know that, like you prep for them appropriately.

I would love to talk about the second thing you mentioned, which is kind of this idea of pairing product value increase with price increase. But specifically you mentioned something that I think is really interesting, which is you kind of needed to reset expectation with customers that the price... we are going to be continuously improving the value that in the product, and therefore you should expect that some amount of price increase is kind of a natural motion for this business.

Would love to hear about how you thought about communicating that, and then secondarily like, you know, as someone who's like lives in the bowels of like pricing and business models, I'm actually curious, like just mechanistically, how would you increase the price? I think a lot of, you know, in a usage model, I totally understand how you can increase price in a seat based model, it's kind of harder to understand 'cause you know, per seat prices, there's certain thresholds where if you're above them it starts to be like, this is like crazy expensive. So talk to me a little bit about like, both those angles. Like, how did you communicate it and then how did you mechanistically make it possible to like, kind of build this durable price increase?

[00:15:16] Ryan: Yeah, no, it's a great question. So I'll talk about the mechanics first, and I probably should have clarified this upfront, but I. This was kind of a unique company. They were sort of in the environmental space and they would basically do certifications to say that you had passed all these environmental standards and they were a body that said, you know, we're gonna go do all that work to run down all the regulations and certifications you pay us.

We're gonna validate not only all of your internal processes, but also all of your supply chain. And so we can basically certify you as you're gonna have like this green stamp of approval, which is good for, you know, your investors, your company, your customers, all that. And depending on the level of certification you would do, you would pay a different rate.

So think of it almost like a licensing model. You were paying kind of a one-time fee every year to go get certified, and then depending on the scope of your certification, you would be in sort of different tiers. And so what we did was, look, there's a lot of things that the company had invested in terms of covering that scope, not only of all the different certifications that they would add, but also different components of your business. So not just your internal processes, but all of the supply chain, et cetera.

So what we did is said, look, those tiers, kind of that license price that we charge you, that's what had not been changed in almost three years. So we were gonna do a wholesale upgrade on the pricing related to those licenses. And then what we did is set expectations that look every year, when we add more certifications and when we add more scope to covering more of your end-to-end business or ecosystem, you should expect that the price kind of on maybe the highest tier of that license is gonna go up by 5% every year. And a lot of that from, you know, the investor side was really focused on how do we cover inflation? How do we cover kind of some of our costs?

But in terms of how did we get that communicated to customers, I would say there was kind of three ways we did that. One was kind of a marketing-led distribution of a bunch of different enablement across the website, webinars, email, et cetera. And you could think of it as we wrote like a really, I would say, powerful recap of here's all the things we've done for you over the last three years. Here's our roadmap over the next six months and all these exciting things that are going to be coming.

And with this, by the way, there's going to be a pricing increase and you should expect that in the future as we continue to deliver more value, we'll continue to, you know, ask us to share in some of that upside or some of that value that we're creating for you. So that was the first motion.

Second motion was they actually had an existing kind of customer council. And this was a bunch of their top largest, highest spending customers. And in that arena, we basically did a workshop where we all sat down and kind of walked through what had changed in the product, what all the value was. We asked a lot of questions about what they wanted out of the company in the future, and that's where we also kind of walked them through the rationale for the pricing change and then set expectations with them as well, that there would probably be future pricing increases, but we would want input from them ahead of doing that.

And we ask that they would sort of be sponsors or, you know, people in the ecosystem, in the community. That would sort of defend the pricing change and back up the claim that we had added all this value. So think of them as like champions that were gonna help kind of go out and support that change.

And then the last one was really kind of a sales motion where, you know, on their battle cards, on all of their scripts, all of things that sales was using to kind of do regular customer outreach. We had kind of distilled that down into kind of a few punchy sentences. And for all of the existing customer reps, so people focused on kind of account management, they had a script that they were gonna follow to kind of reach out to their customers and make sure that every customer was contacted before the pricing change went live.

[00:19:28] Scott: Awesome. Any, I guess, any general thoughts about this idea of like, you know, I'm a software business, I'm gonna be continuing shipping code. It's not just gonna be kind of staying static, and therefore I should be increasing value over time if I'm doing my job. And therefore, like maybe I should share on a little bit of the upside.

How do you think about in general, what do you need to do in that business to make sure that that becomes a kind of core part of how that business runs? Or maybe what are the blockers that, like you see when companies are kind of thinking about this? Like, you know, I've noticed just personally that Dropbox cost $9.99 in 2009, it cost $9.99 in 2025.

And I'm curious like, generally how you think about kind of baking in this expectation that like, prices will go up because value's going up. Like what are the lessons for the software industry or that you would advise like folks who are thinking about like this problem.

[00:20:32] Ryan: My first reaction to that, and maybe it's a controversial hot take, but I would say a lot of people are always scared of increasing price. And you know, a lot of that comes down to, you know, what do we tell customers? You're afraid that there's going to be churn, you're afraid that your competition's gonna run with it and take advantage of it.

So it can be very hard to lean into those types of things. So, you know, I think to get confidence internally to do those sort of things, you need to do all of the right customer research to validate that what you are delivering is actually adding that value and that people are willing to pay for that value.

And at the end of the day, you know, you can make more money, even with slightly fewer customers, if you feel confident that the right customers are willing to pay more. Now, I caveat that by saying, you know, I'm in an industry today that's incredibly competitive. The expectation in the industry is that prices go down and prices never go up.

So I think you also need to be realistic with where you are, what is your competitive differentiation? The company that I spoke about earlier, in terms of the time I did at McKinsey, it was in a niche space, right? It was offering a product that not a lot of other people were offering. That's a little bit of a different dynamic than, you know, operating in kind of the cloud or the data world where, you know, you're dealing with the cloud providers and there's a bunch of competition across the board.

So some of this is, you know, where do you play? But I do think that, you know, if you do the right research and you are convicted in the value that you are offering, there are places where you can find that capturing more of that value is reasonable, especially in a world where some of the stuff that's happening is incredibly expensive, right? And I think customers resonate with that. They know that, hey, if your costs are going up, I understand why potentially you may need to charge a little bit more. So hopefully that answers the question, Scott, but 

[00:22:33] Scott: Yeah, no, I think it does, I mean, I think you said exactly what I... what I believe, which is I think people are kind of rightfully afraid to raise prices. But the way that I think about the problem is if your product isn't getting better, you don't have a right to raise prices. And, and so in a sense it's almost like, inverted. It's like the price raising is like, it's the outcome you should have a right to access if you are doing your job as a product leader or whatever.

And to your point, there are certain market dynamics where you can increase the value dramatically and not be able to raise prices because competition is so fierce. I mean, I think that's like totally, that's like another part to the calculus, but what I would say is more often when you examine it, it's more like people are just kind of intrinsically, they're leery of getting rejected and they underappreciate the idea that you said, which is that actually, you know, there are a lot of markets where by raising the price, you make net more money. Yes, you may have fewer customers, but those are customers who are much more value. They'll actually see and feel the value that you're presenting. And... and over time, you know, those kinds of businesses end up being quite successful.

I mean, I think, like, you know, Hermes sells scarves for thousands of dollars and like, you know, and for, and that's a really good business. And you need to find your niche and where you can do that. Not all markets fall into that, but I do think it's like something that I see people making this kind of, I guess, fear-based error a lot.

And also the other thing I would just say, especially as an entrepreneur, is you don't actually know the true value of your product. If you're doing something brand new. Like no one does, the market does not know, and the only way to discover the price is to price it and see what the market will bear.

And I think if you're doing something truly net-new, you should be. Especially in the beginning, quite, quite willing to experiment here. Awesome.

Well that's actually I think a pretty good segue into Snowflake. So maybe talk a little bit about kind of your role at Snowflake. And I think a live question that almost everyone asks is, where does pricing sit in the organization? Is it in finance, is it in product, is it in somewhere else: Strategy, BizOps? Like where is pricing at Snowflake? 

[00:24:54] Ryan: Yeah, so I joined Snowflake about two and a half years ago. I originally came in on the pricing team and we actually used to sit in the sales strategy and ops organization, which was on the go-to market side of the business.

And you know, we still closely partnered with product, but I think the design at the time was to kind of connect the go-to market and the product organization more tightly. In that role, we did a lot of work, too on kind of the rollout and the implementation of a lot of that pricing, and that tended to be more of a go-to-market facing function.

When I was about six months in, we went through a re-org and we actually moved our pricing team over to our finance organization, which at the time I was excited about because a lot of Snowflake in the early days was, you know, really more of a cost plus model in terms of understanding, 'Hey, we have all this infrastructure that we're using to run customer workloads. Let's think about how we price on top of that infrastructure'.

And so by being in the finance organization, I got a really good understanding of our architecture, our infrastructure, what all of our underlying costs were. I learned a lot more about the product in a much more nuanced and intimate way.

What was also exciting about that role is that I got a much more broader relationship with the product and engineering organization. It was now no longer just pricing, but it was really serving in terms of like an end-to-end finance business partner, of which pricing was an important element. So, we still live within the finance org today.

You know, I report into our head of FP & A who works under the CFO, and then we support all of our product leadership team directly. 

[00:26:38] Scott: Awesome. What do you think are like, the core pro con around... actually pricing living in go-to-market is something that I haven't heard that often.

It seems fairly rare, but like, what are the pros and cons of kind of those two organizations from your point of view, and then from the just like macro business point of view? 

[00:26:56] Ryan: Yeah, so I would say. I like the finance structure because in my mind, finance, at least internally, is a very objective partner.

You know, really at the end of the day, they're focused on trying to do the right investment for the company.They're being very thoughtful about how the unit economics work, how those scale over time. If you think about pricing in either directly in the product organization or directly in the go-to-market organization, sometimes there are trade-offs across those two lines, right?

You know, product is focused on getting the product out, making sure that you capture all the value of the ip. Sometimes the go-to-market organization can be focused on, let's just get the deal, let's get the customer over the line. And so, I found that the finance organization was a much more kind of neutral spot to, I think, have good relationships with both organizations.

And also be kind of this objective business partner that had a lot of context on the economics and how that fit into the overall company performance. 

[00:28:02] Scott: Yeah.One trend I've seen, I would be curious to hear your reflection, is that the modern office of the CFO is... it's both kind of a compliance function, like an accuracy, but it's also increasingly becoming a data and strategic function.

And so I think putting pricing there kind of aligns with that. Maybe talk a little bit about, kind of broadly how you've seen finance become more of a strategic partner rather than just purely like an error checking process on top of other organizations. 

[00:28:33] Ryan: Yeah, no, I think that's a great point. I do think that as we move into more of the world of AI and other things, I think a lot of that kind of back office routine type of work will get more and more automated and I think that finance will create more leverage and more capacity to be more of that strategic function, that strategic business partner. And again, like I mentioned, you know, you have kind of that objective point of view of, 'Hey, I'm working with a bunch of different functional teams. I'm not embedded in any of those teams directly, so I could be fair and make kind of what's the right decision for the overall company'.

I will say we are a bit of a unique team within finance though, because at the end of the day, you know, we largely price a lot of our direct product offerings. Snowflake's a very technical company. You know, we sell infrastructure software at the end of the day. And so to understand not only how the costs work, but also what is the product? What is the value prop? Who's the buyer? How does this compare against alternatives?

You need to almost, you know, think like a product manager. You need to understand the technology, the market, those types of things. And so we've had to go out of our way to kind of hire a team that has a lot of strong financial and pricing skillsets, but also has a bit of a technical bend and a technical background to them, which I'll tell you is not the easiest team to recruit for or hire.

[00:30:01] Scott: Yeah, well it actually kinda dovetails on...kind of... A common thing that I see, but that feels kind of flawed in a way, which is, you know... I think most companies at your scale have like a pricing council, and it's kind of like, you just said it, right? Like, proper pricing is a very cross-functional, cross-disciplinary problem.

And what I've seen with pricing councils is they kind of tend to have like five strong-willed people who all have different views. And then what that kind of causes is like endless debate about things and then like action is relatively harder to gin up. So, talk a little bit about how you all approach that at Snowflake.

Like how do you make sure that your pricing council is moving at the right rate and kind of in the right direction while still balancing the fact that, you know, typically there's gonna be a lot of different perspectives in the room. 

[00:30:56] Ryan: No, that's a great point. We actually made a pretty meaningful [00:31:00] change to our pricing committee process about 18 months ago.

I'll explain a little bit more about why, but I'll start with, I mean, we do have a standard pricing committee today. It largely makes up a lot of our executive leadership team with representation from all of your standard functions.

I would say that, you know, the people in the room who largely have the most say or care a lot about what the decision is, tend to come down to the CFO, our Chief Product Officer.

And, you know, particularly if it's something that is very different or unique from what we've done historically or might include something like a big pricing drop, our Chief Revenue Officer or sort of our go-to-market leadership is also very involved and curious.

I would say though, one of my big learnings since taking over pricing is the importance of doing kind of all of the meetings ahead of the meeting and making sure that you can kind of get alignment from not only those folks, but a lot of their leadership team as well, [00:32:00] and really flushing out what are the concerns, what are the risks, what do they care most about?

And I would say, I've learned a lot about there are distinct differences between what product cares about, what finance cares about, what the go-to-market team cares about. And so we've just integrated that into our process. We actually have a checklist that we go through that as we're even, you know, in that first kind of product development phase, you know, our private previews where we're thinking about getting customer feedback, we're actually taking that checklist out and saying, ok I know our CFO is gonna care about the margin, and I know they're gonna care about how our costs scale over time. So, let's think about that early on and let's make sure that our pricing metric aligns to our costs so that those two things scale at the same rate. You know, I know my Chief Product Officer's gonna care about keeping it simple, right? Not overcomplicating it for the customer.

So yes, maybe our competitor has a more sophisticated pricing model, and that might even be perceived as more flexible. But I know that one of our principles is, let's keep it simple. So, I'm asking those questions with individual PMs in the very beginning to kind of push our thinking so that hopefully by the time we get to the pricing committee, a lot of those kind of standard principles and things that people care about have already been kind of settled.

[00:33:18] Scott: Yeah. Actually, I think one of the striking things about Snowflake is how principle-driven the pricing process is, and I think that's... I think different companies have different levels of this, but as I understand it, Snowflake takes it quite seriously. Maybe talk a little bit about the benefit of having principles? Like, why not just YOLO it here? What is the point of principles from your perspective in the pricing process? 

[00:33:46] Ryan: Yeah. I mean, I think at the end of the day, it allows you to basically delegate and get leverage as the organization gets bigger and bigger, right? You know, I think the primary person at Snowflake who really feels  deeply about those principles is our Chief Product Officer.

If he had to be in every single pricing conversation, I mean, he would probably never get anything else done. And so, I think it allows the organization to operate off of a rubric or a certain guide, even when they start writing that first PRD, right? Like, let me think about how this product's gonna work. How do I think about the pricing?

I think it streamlines a lot of those decisions that we talked about in terms of the pricing committee. There's been a few where I've gone in the room and everybody kind of gives a thumbs up and everything is decided relatively quickly. And that's because, you know, it's not a significant change in terms of the pricing metric or the pricing model, but we've also followed those principles.

And I would say, and this is a really important, in my mind, pricing is downstream from what is your product strategy and what is your company strategy? You know, how do you wanna be perceived? Where do you wanna play? And so you need alignment across those things. If you're doing a pricing thing that feels counter to the product strategy that you're trying to drive. For example, let's say Snowflake wanted to bring in more data. Well then, we wanna make it cheaper for you to load more data into the platform because we may charge you more for other things that you're doing on the platform, but we wanna make it as easy and as frictionless as possible to get data into the platform.

If I made a decision on pricing that introduced a really complicated pricing metric to load data in, or made it really expensive, that wouldn't feel aligned, right? And so I think having those principles that tie back to your product strategy just helps make decision-making run faster and not get bottlenecked up at the executive level.

[00:35:45] Scott: Yeah, I mean, I think this concept that pricing is the product or is a key attribute of the product, or is best thought of as part of your product, is exactly right. And actually, one of the things that I find notable about Snowflake, especially as a usage-based business, is how much product effort goes into communicating the value, and the kind of dashboarding and control structures that you offer to your end customers.

And I think that to me is a really key part of like, look, you know, if pricing is my product and this is a consumption-based thing, how do I help customers feel at ease with the fact that they can go, you know, spend a large amount of money in a very short period of time. Maybe talk a little bit from your perspective, the importance of the kind of like, end customer experience of pricing, and how you as a pricing leader kind of incorporate that concept, like the Grockability or the kind of how you're gonna consume it in the product. Like how do you think about those things when you're thinking about a pricing change?

[00:36:50] Ryan: Yeah, it's a great question and I would say, look, this is a concept that is evergreen. Like, you will never have nailed this piece, right? Because at the end of the day the market is always changing. Your customer base is always changing. You know, the person who is the system administration person when you first sign the contract may not be the person that's there two years later, right? So, this is a constant requirement. It involves a ton of different services, right? It involves everything from, you know, how do you get up at your user summit conference and speak to customers, right? How do you enable your website so that it's intuitive and easy to find how much it's going to cost me for x, y, z?

It's even built in tooling into the product itself so that you can set budgets, you can set resource monitors, you can enable things like chargeback so that a customer can, you know, grab their users and make sure that they're sticking within their budgets, right? It's enablement. It's empowering your field to be educated and understand how our pricing works, not only for the core business, but also the new stuff that's coming out.

So, it is a big responsibility, I would say, to make sure that if you are going to say one of your core value props is simple and the platform is easy to use, you need to deliver that on the pricing side ac across a bunch of different channels. 

[00:38:15] Scott: Yeah, I mean, I think you just outlined what I find to be the kind of, one of the distinguishing things about Snowflake and companies that have true consumption base that kind of take this problem really seriously. Maybe talk a little bit, I, you know, I think one of the interesting trends in the industry right now is there's a lot of companies that are on kind of seat subscription business models, which value communication is relatively straightforward, right? It's like, and definitely price communication's pretty easy.

It's like, how many people do I have, multiply by a price? Okay, cool. Like my bill is easy. Snowflake isn't like that. AWS isn't like that. Talk a little bit about some learnings you've had, o r that Snowflake has had writ large around value communication, like the key things at least, that maybe the key surfaces, like one or two things that are like matter a lot to customers in these consumption-based businesses that might be unintuitive for someone who's kind of used to treating the billing portal almost like an afterthought of an afterthought, it's like a thing that you just don't care at all about.

I get the sense that from Snowflake's perspective, that's actually an incredibly important surface to kind of think about. So what are some learnings that you've had there? 

[00:39:24] Ryan: I would say one of the things Snowflake wrestled with probably two or three years ago was how much of that does Snowflake, you know, build into the product itself and innately control that messaging and that ability to monitor and budget and resource against versus how much do you rely on the ecosystem and kind of other partners to help also do that?

And I would say that one of our big learnings probably two or three years ago was, you know, we need to do both, but we need to lean in and do more of it in the native direct Snowflake surface. Because you can't always control the perception and the way that things are positioned when you rely on partners only.

And so we invested heavily in a ton of cost governance tools that made it very intuitive and easy to understand how much you were spending, how much you were gonna project to spend by the end of the year, at what point in time would you run out of, you know, remaining capacity on your contract and therefore would need to renew.

So I think we've invested really heavily on surfacing a lot of that in the product itself directly. I still think there's potentially opportunity to not just surface the information, but also provide information on the so what or what you can go do about that. So, if you are concerned about how much money you're spending, either reinforce the value of, hey, you might be spending a lot, but you're getting a lot of work done as a result of that. Or, let me give you three or four things that you can go do if you are concerned about spending too much to kind of reduce your bill and bring that back into a more tamed run rate.

But I would say that we are looking at all that stuff and then consciously also spending a lot of time talking to customers to figure out, you know, is this just an issue where people need to understand how to control cost? Or do we actually have a fundamental pricing problem? Or maybe we do, maybe we are charging too much, or maybe we don't have the right metric in the terms of as they use it more, the amount they're spending isn't scaling with the value that they're getting out of that.

And actually, a great example of that is we recently announced a pricing change for Snowpipe which is a product we use to ingest data in. And we moved it away from a model that was based on how long it was running to a model based on how much data you were ingesting. And we made that change as a result of getting a lot of feedback from customers that said, look, I know how much data I wanna put into Snowflake.

I have no idea  based on how many files, what type of file it is, and what type of compute you're gonna run for me, how long that's gonna take. And therefore, I'm hesitant on how much I'm going to put in. And so we said, 'look, let's do something that's simpler. Let's do something that's more predictable and let's make that change.'

So, you know, I do think we've invested heavily in trying to better surface that information for customers and make it actionable, but then also being responsive when we do think there's an opportunity to maybe change the pricing model or change the pricing lever. 

[00:42:31] Scott: Yeah, I mean, I love that example because it approaches it. It kind of starts the idea in like what is the actual value and then how do I communicate that to a customer? How would they understand the value? And it's really easy in that case to think, you know, to think actually cost forward. It's like, okay, our cost is probably a proportion. It's like it's proportional how long the job is running or whatever, roughly.

And then data is like kind of a function of that too, but like, okay, but if I reason about it from a value perspective, it's like how much data does Snowflake reliably ingest into our product and store on my behalf? Okay, that's actually a much more Grockable thing for end customers. It's simpler to predict ahead of time, and it's kind of just leaning into like what's the value and then how does my metric measure to the value? And then the pricing kind of is downstream of that. I just think that's like absolutely like the kind of north star that I would incept in everyone. It's just like, start with the value, start with the customer, understand that. If you understand that, then at least you're on the right path. If you start without that understanding, there's no way you're gonna like find the right answer.

[00:43:38] Ryan: Yeah. And the, and the other big lesson out of that initiative was the rest of the industry had really coalesced around charging based on the number of terabytes to come in, right? Based on the number of the amount of data that you're ingesting. And so, you know, another big lesson for me in trying to be responsive and understand where customers are coming from is if you are going to deviate from the industry standard, there is quite a burden on you to then help enable and educate customers on why that is needed and why it should be different.

And you better have a pretty compelling reason for why you need to operate that way. Now the reason we had done that historically is 'cause that aligned with we, the way we charged the rest of Snowflake. And so we wanted to keep it very simple. We didn't wanna have fragmented different pricing metrics across the platform, but we ended up deciding, look, this is the way the rest of the industry charges.

It's much more straightforward and intuitive for customers and we feel like that's the right thing to do. And so we made that change. But that's another kind of big lesson is, you know, when the industry has a very strict way of doing things, really make sure you understand the reasons why you might be changing thatbefore you go ahead and do that.

[00:44:48] Scott: Yeah, I cannot agree more. I think like, pricing is Darwinian competition and sometimes you like, have some weird mutation and you go off… off on the side and you are right, and the industry converges on you. But more often than not, that's like not a good branch and like you should just like adopt the industry norm.

And so the way I think about it, actually, the way I thought about it from Snowflake's perspective is just that it's if you're gonna do something unique, be prepared to pay the like marketing tax for it. And like, my best example of this is if you look at Mike Scarpelli's blog during the pre and post IPO window, it was roughly a continuous refrain on the benefits of usage-based pricing. But essentially what that was… it was training the market how to think about usage-based pricing in a snowflake context. And that marketing expense was incredibly high. And it worked. And it was right at the, you know, it like took a long time for the market to understand it. But I just think it's like, unless you're ready to devote that level of focus time, attention, you probably should just stick with the standards 'cause I think customers understand the standards. So.... 

[00:46:03] Ryan: Yeah. And it's not all or nothing either. Like I mentioned, you can have different models at different points in the portfolio and you can make those trade offs relevant to that respective area. But yes, totally agree. It's definitely a marketing tax if you are going to deviate from the norm.

[00:46:20] Scott: Well, I had two last things I wanted to go into. The first is kind of, you know, the evergreen problem with usage-based pricing is forecasting. So, talk a little bit about how you all have focused on, or how you center the forecasting challenge that comes with the fact that like, your product is extremely flexible. New workloads can come on and come off with seconds' notice. Like how do you think about forecasting and predictability for running the business? 

[00:46:46] Ryan: Yeah, that's a great question. So, I would say first I'll start, it's incredibly hard and one of the benefits, again, of being in the finance organization is that I get to sit right next to our finance data science team, and that team is responsible for all the revenue forecasting at the business.

You can imagine the CFO wants that team very close to him. And so, what I do is I work very closely with that team to understand not only what are we forecasting as the business kind of runs steady state, but as we start to entertain either new products launching or making pricing changes to existing products, I work very closely on doing very detailed analysis on what we think the impact is to revenue and then incorporating that into the forecast.

And typically that's done kind of as a manual adjustment to begin with, but then over time the forecast models pick a lot of that up. But we work very closely there. Now, as you probably can expect, Snowflake, it's not just about the list price, it's also about the performance of the underlying platform.

And so in many cases, it's not just understanding we're dropping price from one to 50 cents, it's also, well if the service is getting 20% faster, what do we think the net price is that a customer's going to pay at the end of the day? And, you know, snowflake is committed over time to wanting to give customers better economics.

And so, we wanna pass a lot of those savings on to customers when the platform gets faster. But it is a very, I would say, weekly, if not daily, exercise to stay in sync with that team, especially given the recent product velocity and a lot of the things going on at Snowflake. But it is a very close partnership that allows us to kind of stay within the right level of accuracy, forecasting-wise. 

[00:48:36] Scott: Let's say you're a smaller company and you obviously don't have that level of data, nor that level of resourcing to kind of go approach it. How do you advise like A CFO who's kind of maybe entering usage-based pricing and kind of, they're coming from this world where SaaS, you know, you kind of can build cohort retention curves and you can kind of roughly understand what my revenue will be when they're entering a company that's like, has the kind of inherent, I would say, volatility that comes with usage.

Like, what's the advice for kind of making that transition and what are the like 80-20 investments that you'd put in place to kind of make sure that you're able to run that business?

[00:49:14] Ryan: Yeah. Well, I mean, first I would start, you could do a lot of this on Snowflake, but also, you know, it takes time, right? So when we first kind of rolled out our more ML based type forecasting process, I remember talking to my boss and said, this took better part of a year to really kind of nail down.

And there was a lot of testing and validation that went on during that process. So, my advice would be, look, start simple, especially if you're transitioning off of a non-consumption based business. You know, be thoughtful about kind of a slow rollout and for the people that, whether… either it's the customers and or the parts of the platform that are moving to the consumption-based model, keep that small in the beginning. And to be honest with you, you can probably get away with. Some simple regression models and those types of things to start, you may pay a tax in terms of level of accuracy, but as you get more data, hopefully those models get better and better. And if it's a very small percentage of the business, then that's a tax and something that you can live with as you kind of iterate and get better and better over time.

I do believe in the world that we're living in today with AI and with, kind of, these advanced cloud native data platforms, a lot of this kind of like, time series forecasting type of thing is getting more self-serve. It's getting more democratized. And so, you know, you don't need to necessarily have a PhD in statistics working on your team to probably be good enough to get started.

I would say though that you also need a lot of capabilities from a systems perspective and making sure that you have the ability to adjust these things and make sure that you're learning over time. And so I would do… I would start small. And then I would scale up over time as you get more and more confidence that you can be accurate.

And I would try to control the amount of variables that you can, if at all possible, whether that's different features, different geographies, or just the sheer number of customers and the amount of usage that they are doing on the platform. 

[00:51:16] Scott: Cool. I'm also curious because you have a very large customer base and you're on a consumption model.

Like, are there any unintuitive either segments or cohorts that you find particularly useful when thinking about forecasting? Like, if you are starting to introduce a little bit of complexity, a little bit of segmentation, like where are the areas that you'd point that finance team to go look?

[00:51:38] Ryan: Yeah, I mean, I would say customers that are using a lot of different features on the platform tend to be more complicated. Those that you know, tend to have very different usage patterns, right? There might be things that only run, you know, once in the morning and then never again. There's stuff that's always on, it's always running in the background.

It might be some sort of, you know, load or cron job that's kind of always operating. So, a little bit depends on what exactly is the customer doing from a workload perspective. Actually, interesting story I was diving into this week a little bit of better understanding our seasonality adjustment. We have a lot of holidays going on right now, and actually holidays can have a huge effect in a consumption based business. So, depending on where you are in the world, how many holidays that are going on there, that can also play a pretty large role in your ability to forecast accurately. And you know, I would say too that I.

The size of the customer also matters, and how penetrated within that organization also matters, right? If you're talking about we've got one team using Snowflake and they have five analysts, that's a little bit of an easier problem than, you know, a multinational corporation that's got offices all over the world and is using all sorts of different things on the platform.

So, I think to summarize, it's probably the penetration and or the feature adoption, the size and complexity of the business that the customer is operating, plus a lot of like where you are in the world and some of the seasonality that, I think, drives up the complexity from a forecasting perspective.

[00:53:09] Scott: Very cool. Awesome. And the last area I would love to talk a little bit about is obviously the thing everyone's thinking about, which is AI agents, all of that good stuff. So maybe talk a little bit about what you're seeing in your market from an AI pricing packaging perspective, and how you're thinking about kind of navigating this moment.

[00:53:28] Ryan: Yeah, it's a topic that I think about every day. I think it's one of those things that this space is moving so fast that one, you're gonna get a lot of bats at this, so you probably have a lot of opportunity to... to again, evolve and kind of go with where the market is going. I see that there's a lot of opportunity to learn from some of the bigger players in the market too as well.

What I'm seeing a lot of is it's very hard if you are kind of selling infrastructure to deviate a lot from where the industry's at in terms of the standards. You know, people generally have an expectation of how much I'm gonna pay for a million tokens on a llama model. And so really where I think there's room to innovate and or do different things when it comes to pricing is what are you doing on top of the infrastructure, right? If you think about the compute plus the models, a lot of that is, is fairly standardized, I would say, in the industry, but it's the, you know, the data agent or the talk To Your data solution that's being built on top of that, that I think there's a lot of room for, 'Hey, how can we be thoughtful about, is it really just tokens we want to charge for? Is it the number of messages or the number of questions that we're answering? Does every question have the same amount of value? Is there an opportunity to maybe charge differently for different types of questions, different types of prompts?'

And another thing that we've also been wrestling with a bit is, you know, customers are very excited about these things. And so they, when they think about future usage, they think about, oh my goodness, 'Well, what if every person in my company was asking five or 10 or 15 questions a day. I think I'm gonna spend so much money'. So there's a lot of thought going into, you know, how do we get people comfortable with expected usage? How does the pricing model scale as the amount of volume increases? What is the right amount of spend that feels reasonable for these things?

So that's a lot of things that I'm thinking through right now. I would say Snowflake is still kind of early in its journey here. I think we are learning a lot from what other folks are doing, but also hearing a lot from our customers.

It's early days. I think a lot of stuff is still kind of in that POC experimentation phase. We haven't seen a ton of things kind of move into production, yet. And so I think we've got a lot to learn still. 

[00:55:49] Scott: Yeah. I'm curious, I think a question that I commonly get is like, who are you looking at such, what are your info sources? 'Cause I think, there's a lot of hype around, like outcome-based pricing, for instance. And I'm like, you know, if I talk to these companies and I dig into the actual contracts, I'm like, you're not really doing this.This is like good marketing. It's like, actually this is just straight up consumption.

And I think that's fine. Like, I think that, like that is part of the marketing. But how do you advise folks to get beyond either the list price or the hype price, and kind of get to the ground truth and make decisions grounded in reality? 

[00:56:25] Ryan: Yeah, it's a great question.

So, I mean, I would say first and foremost, you need to start with what I kind of call like an inside-out view here, which is what can we actually measure today and what is actually reliable? Because I'll tell you right now, if you move to an outcome-based pricing model, and every time I achieve that outcome, I potentially charge you a different rate, or I have a slipping definition of what that outcome is.

Over time, you lose trust very quickly with the customer. And I'm still convinced that while people see a ton of value in these offerings, at the end of the day, people want something that is predictable, that I know what I'm gonna spend over this amount of time, and if I need to change that, I have the controls and the tools in place in order to do that.

So I very much start with what is in our control of what can we actually measure? How can we be clear with what the definition on these things is? And then I start to think about what are the types of ways that people are gonna get value out of this? One thing I've been very much thinking about recently is often, in these things, there's a lot that the customer does on their side, but there's a lot of things that happen in the backend that the customer's not necessarily in control of. And those things drive cost as well. And so, you know, I think an important principle here is if you're going to charge somebody a certain amount based on a certain metric, they need to then have the ability and or the control to control that price or control their spend as well, right?

Like, if you're gonna tell me I'm gonna, I'm gonna pay for a certain outcome, or I'm gonna pay for a number of agents, and all of a sudden I start spawning out, you know, multiple agents that you didn't ask for, and you're getting charged for those, right? That's not an ideal outcome. So I think you really have to be grounded in what can we measure today, what's in our control, what's in the customer's control? And then you can start thinking about value capture and what's the right level of pricing to charge and what's the right metric that people see is closely aligned to that value? 

[00:58:33] Scott: Couldn't agree more. I think this concept of auditability visibility control over spend is one of the things that I think is just so critical once you're any amount of revenue scale. Like, the example that sent it home for me was, I talked to Segment like six years ago and they basically had a PayGo model and some intern at some startup in India that had like $3,000 in the bank accidentally hooked their entire chain, like basically event pipe through a CDP and got racked up an $80,000 bill.

And I think the truth is that variable, like usage-based pricing is variable value. Like, and that also means variable spend. And that's like a good thing in a lot of ways. It aligns incentives, but it can also be a very risky thing for customers. The more you're kind of, moving into this automated agentic stuff, you just gotta make sure you're really investing in those control surfaces and those visibility surfaces. Otherwise, you're gonna have a lot of really unhappy customers, or you're gonna end up eating a lot of overages because you're just like customers, like how can you possibly charge me for something that I have no control over? That makes no sense, and I think that's just a really critical thing to think through as you're designing these products.

[00:59:48] Ryan: Yeah, and I think the stakes are higher here too, because the costs of these models and the cost of this compute is higher than the traditional cost of software, right? And so, you know, you're not always in the position to afford, Hey, let's just forget about that model or agent that ran over the weekend that you did forgot to turn off, right?

I mean, these things are expensive. They're supply constrained. And so yeah, you have to be very thoughtful about the audibility, the controllability, giving customers the ability to make sure that they have reigns over that.

[01:00:20] Scott: Awesome. Well, last question is, you know, what is something in the kind of space of pricing and packaging over the next six months that you're, particularly like, watching with a close eye?

[01:00:30] Ryan: I am watching kind of what is the kind of settled on pricing metric for not the foundation models themselves, but in the snowflake world we're really focused on kind of this talk to your data solutions. So whether you call it agents or whether you call it like a… an analyst in a box or a copilot. You know, these things that can kind of answer questions for you, use natural language, and then also maybe go off and do work on your behalf.

I'm really curious what is the right pricing metric there? I'm not fully convinced on outcome-based pricing yet, although I'm sure in the future we may find the perfect way to do that in an audible, scalable way. But that is kind of one thing that is also in the back of my mind right now.

I am also curious about in kind of the cloud world that is a consumption-based model. How much of that continues to remain? How long the service is running versus potentially, you know, units of work or units of value that you are delivering? Because in many ways, the customer wants the outcome, right? They want the job done, obviously they want it done as fast as humanly possible, but you wanna align incentives. And so I'm really curious where that goes. So those two things would be the things that I'm most thinking about and most excited to see as AI and data evolves. 

[01:01:55] Scott: Very awesome. Well Ryan, thank you so much for the time. I think there's like a ton of really, really good insights here and really appreciate you spending the time with us.

[01:02:05] Ryan: Awesome. Thanks Scott. This was great. See ya. Awesome.

[01:02:09] OUTRO: Thanks for tuning into this episode of Unpack Pricing. If you enjoyed it, we really appreciate you sharing it with a friend. We'd also love to hear from you. Feel free to email me at scott@metronome.com with feedback and suggestions for who you'd like to see on a future podcast.

‍