Consumption-Based SaaS Pricing Models on Snowflake | Guide

[00:00:05]good day everyone and welcome to build 2021 my name is Frank Cerrone I'm a senior director in snowflakes Finance group uh with a focus on helping our customers strategize on their business and pricing models I'm joined today by my colleague Matt Hill hi everyone thanks for joining uh my name is Matt Hill and I lead our data application sales strategy at snowflake we wanted to spend some time today talking about SAS pricing models in the market in general but more specifically some of the experiences that we're having with our customers as they go to go to market on the snowflake platform first a bit of history on uh on the SAS business as many of you probably know the SAS the SAS pricing models have been around for decades they actually started in the 1960s but a company that came to Market in 1999 called Salesforce was the first pure play SAS subscription model in a public company that came to the market what's starting to happen now as these companies transform their business models and certainly the success the success that snowflake has had you're seeing this increase in Surge in usage usage-based pricings traditional SAS models have largely been subscription billing based models where someone pays for access on a recurring basis when in fact today the way snowflakes gotten to Market and some some of the other companies that are coming to Market recently many of these companies are migrating to a usage-based model which we want to talk about today give you a little bit of flavor of what's happening in the industry aside from Snowflake there's been several very successful companies that have gone public at a large multiples to their to their revenue stream most of which have been around this usage-based pricing model just to give you a sense of the magnitude of success today sending snowflake aside even there's been a probably about a 7x multiple increase for the companies that have gone public under a usage model the market has very much adopted this strategy as a leading strategy for pricing and as you are considering building your product it's certainly worthwhile to understand the premium that you might attract if you're able to build a usage-based pricing structure a little bit around sort of the customer benefits you know obviously we all enjoy premiums in our pricing and our stock price goes up which is great but from the customer's perspective it's important to understand what their experience is like and how positive it's been at least for snowflake in building a usage model customers typically struggle with some of the traditional subscription models and what they found is that if we come to Market with a pay for use model they've quickly can correlate product usage to pricing and can come up with a methodology that makes sense for them that isn't onerous or viewed as overpaying since they're only paying for what they use to control the budget suddenly shifts to the customer's backyard from yours as the developer or the go to market Product Company the control the control mechanism for them quickly becomes how much do I really want to expand this use case or other use cases with your product and since I'm only paying for what I use as I expand these my budget will ramp up or down almost on demand based upon my sort of dictating which direction I want to take my business it's taken a lot of risk out of the sizing of a contract one of the conflict points that's pretty common in going to Market with any kind of a subscription model is helping describe to a an end user customer how to guesstimate the number of users or the number of features or or whatever the other metric is that the company prices around how to size that in their environment especially for a new software product that they really haven't deployed yet taking the risk out of the sizing exercise allows our customers to sort of flex up and down as needed and almost on demand within their own environment and then the last point and maybe the best one from a customer's perspective is a very low entry point uh to the product they no longer have to think about you know I only want certain features or I want to narrow down certain capabilities to keep my costs down rather I'm I'm getting access to a pretty robust product at a very low cost based upon how you might structure your usage what are some of the strategies that we've seen in in deploying a usage model and maybe some things for you all to consider as you are contemplating how you want to go to market three of them come to mind that really Rose to the top as we started to talk about going to Market with a usage model first and foremost probably the most important you need to be really thoughtful around identifying you a usage-based metric that creates value for your customer I'm going to make an assumption that of course you're starting with a really good product to begin with and if you have a good product what drives value from your customers perspective our recommendation would be really try to keep it simple and very measurable so you can actually put some predictions around it but very much value based not Cost Plus base but a value driver metric that has customers enjoy benefit you also enjoy an uplift in your Revenue the second thing to think to think about is your your sales origination process and frankly maybe the mix of your sales team is going to change a little bit uh compensation structures are going to change a little bit no longer are you rewarding A salesperson for just finding a new relationship finding a new relationship in a usage model doesn't really generate any Revenue so the reality is until that customer adopts and expands the business you haven't really generated revenue for the company yet so be thoughtful around how you structure compensation for your sales teams certainly rewarding someone for signing up a new customer is is tantamount to success so you need to do that but it's a leading indicator to the Future think really hard about how you would support your sales reps on an ongoing basis to drive expansion within your customer base and the last one an important one is help your customer size their contracts it may be counterintuitive to think of it this way but you as a developer really own the ultimate responsibility to help customers be better users and that includes optimizing their environment or your product for their environment at first that might seem troubling because your Revenue might actually dip a bit but if your product's a good one and it's expanding the amount of use cases within the organization will more than compensate for whatever you're doing to help the customer optimize and more importantly customers will certainly appreciate you supporting them as a partner as I mentioned before there's been you know a bunch of successful companies that have come to Market as uh SAS providers lately that have done it under a usage model and they're building it around very simple value metrics this is just a sort of a snapshot of course you know snowflake you're hopefully uh you're building on our platform right now as we speak ours is a query based platform that's a metric that we found works as a as a value metric and consequently we've organized our entire company around that metric and delivering our our presence in the market today these other companies have done something similar HubSpot this number of marketing contacts and as you can see various other metrics pretty simple to understand across the industry are are how people are organizing how they go to market now many of these had a opportunity early in their life cycle to make this transition if they were transitioning from a traditional subscription model to a usage model of course snowflakes was a early stage you know sort of Clean Slate of paper a business model and so it was native from the get-go other organizations including yours might not be in that in that situation so some of these might require you to be thoughtful around how you might transition to a hybrid type model that includes both a base licensing subscription charge as well as a usage base charge so you really need to be thinking about how is it you want to approach that and your own specific situation education there's a lot of benefits of course to being uh you know being a shareholder in a company with a usage model I just thought I I'd share with you a little bit of specifics around you know what's the difference in terms of Revenue recognition as you start to think through a usage model first and foremost um you got to pick the right model as you can see on the headlines on the right talking about the right metric has been a lot of the buzz in the in the industry there was a conference recently but that Bessemer Health called state of the cloud in 2021 lots of discussion around what's the right usage metric to build into the into your pricing model companies with usage-based pricing are growing faster than companies in a traditional subscription model and we saw some of those headlines at the beginning of the presentation what's the biggest difference in a subscription model your revenue is recognized on a straight line basis that is whatever your contract is that you sign up with a customer in that year divided by 12 months if it's an annual contract uh becomes how you recognize Revenue in a usage model it's quite different in fact in a usage model when you sign up a customer the very very beginning you literally have no Revenue so until the customer onboards and adopts and begins to build out their use of your product within their Enterprise you will really experience probably less Revenue than you typically would in a subscription model however as adoption expands throughout the customer what we found and what others and usage models have found that the actual usage far exceeds any Revenue you would recognize in a straight line sort of a business model because adoption is getting wider within the organization which has this inherent retentive value around how customers use the product in their specific use cases I thought I'd share with you just some of the snowflake metrics this is public information so there's nothing proprietary here when we get in front of our investor Community there's a few points that really rise to the occasion at the very very top of the discussion around valuation obvious one is how many customers have you added in your in your business and so something that you'll you'll all be measuring no matter what your pricing model is how many customers are generating more than a million dollars and then the last one and probably the most important one is what's the net revenue retention or or sometimes called net dollar retention that you as a company are experiencing with your product in the market these are actual numbers and so in our example snowflakes experience in a 169 percent net revenue retention what does that mean that means that for a customer that's been in our Organization for at least two years or longer we're finding that their second year Revenue usage is a hundred and sixty nine percent greater than the previous year remember the curve I just showed you sort of starts low at the bottom traditional sort of uh bell-shaped curve has a tendency to ramp up as adoption expands that's the ultimate sort of sign that customers are not only enjoying the product but they're spreading it widely within their organization this is sort of an example of what we've learned as we've walked through usage models with other customers as well as our own model the life cycle of a customer although it would be somewhat unique to your specific product that you're taking to Market will follow a pretty similar profile you'll experience an on-demand or an entry-level relationship where you're developing proof of concept where you're really proving out does your product meet the goals of what the customer is trying to accomplish and as they come to the conclusion that it does you'll sign your first contract typically not a very long contract it might only be as short as a year it might be even less depending on your product and what you'll find is that adoption of the product during this period of time might be a bit slow customer needs to get fitted into their fit it into their development cycle they need to get that roll out to their employee base it'll take a little bit to get it started but as it accelerates and optimizes you'll find that you'll reach a sort of a steady state and then beyond that as the customer matures you'll see broader expansion across the Enterprise that will start to geometrically grow your Revenue the important Point around this is how do you predict what a customer is going to use in a usage model it's a challenge there are ways to do it you need to build a model that fits your go to market strategy but for every single customer that will engage in the relationship with you their usage data and profiles will be the will be sort of the inputs for any kind of predictive model you're going to do around usage each of those customers will always exhibit very unique behavior and in the early stages since you really don't have a whole lot of experience with these new customers yet you're going to use existing customers you might have in your business already to cross learn some information about a customer's usage pattern and to use that to extrapolate across the rest of the organization's um you know that you're you're doing business with to help you better predict what your usage profile might look like couple of sort of learnings from our experience in the usage world today generally speaking almost to a person usage customer usage will not follow a straight line pattern so your Revenue recognition will dip a bit and then it will start to come up the curve just as I displayed you previously customers exhibit very large variation depending on where they are in the technical deployment of the product uh organizational politics whatever is driving their adoption will create sort of unique variation within each customer customer usage over time will start to get offset as you optimize what that means is as you optimize your product you actually might short term take down some of your usage or Revenue but long term as the use cases expand you'll actually see an improvement in Revenue plan for the short-term dips and prepare for the expansion and the out years contract sizing is an art and not a science it is difficult especially in a new relationship to predict what a customer is going to do as much as you can learn from your other customers within your portfolio of customers But ultimately you need to get started building a history and a track record at the customer level and then lastly and something that sort of maybe might be intuitive there are sort of systematic and unsystematic factors that drive uh usage Trends there's macroeconomic trends that drive Behavior just like there is unique trends for customers for your product within a customer base um covid a hurricane a natural disaster all these kinds of things you know a recession they're going to drive usage behaviors within your customer base those type of systematic risks you really never get rid of however as you expand and mature and scale your customer base you'll find that the unsystematic or the ones unique to a customer will largely start to cancel themselves out across their own customer base giving you better predictability to your Revenue stream and with that I'm going to flip it to my colleague Matt to talk about an example of a model on snowflake thanks Frank and that's some really great information I'd like to build upon the snowflake example by taking a look at trulio who who actually spoke earlier at the build conference so you can check that session out uh if you if you missed it Trulia uses a hybrid pricing strategy so customers make a commitment based on usage and plan type and then they can exceed that plan with additional per usage charges importantly as Frank mentioned they also offer a free tier to encourage customers to try the service and see the value for themselves before actually making a commitment twilio marketing campaign has two value metrics they charge on number of contacts and number of emails sent you can see the explosive growth that they've experienced in the past 18 months which should not surprise anyone that follows their stock contacts have grown by more than 4X to 850 million and the number of emails sent has grown by more than 36x to a whoppin 27.8 billion emails there are a number of technical benefits that Trulia has realized by running on Snowflake and you can hear that directly from them in their recorded session today what I really want to focus on is how snowflakes usage-based model complements this pricing strategy first is understand how to align the snowflake cost to ensure that margins aren't inadvertently impacted twilio knows that each contact load and segmentation as well as each email sent correlates to the amount of compute required to process and store those events while not the full cogs of running their service this provides excellent predictability of the snowflake costs as they understand what will be required to support the customer plans that they have sold is also important to understand that snowflake's elastic scalability and usage-based pricing allows twilio to Scale based on their clients usage patterns for example they don't have to invest in excessive infrastructure to handle peak seasonality of Black Friday which would then sit idle during much of the Year additionally they can segment different size clients to the appropriately sized infrastructure to ensure that they're always meeting the slas adding to this is the idea that trulia's investment in Snowflake grows as their usage increases they are using the exact same product as they initially invested in years ago but have only been required to increase their commitment to snowflake as their customers adopted more of their products and that usage grew lastly is the fact that not all twilio products run on Snowflake today and they continued to innovate and develop new functionality as well as they look to expand their snowflake footprint with these additional offerings they have a low cost of development because they don't have to invest in additional licenses or infrastructure builds developers can access compute and storage on an as needed basis pain per second when they want to run tests against snowflake now I'd like to transition to talk about some additional Revenue models that our customers are deploying as a byproduct of having their applications built on snowflake these are all models that are unique to snowflake because of our data sharing capability what this allows customers to do is eliminate the typical process of sending and receiving data through apis blob storage dumps and SFTP all these methods require additional Technologies processes and people to support even when they're working there are limits on how much data can be moved and at what frequency instead with snowflakes multi-tenant approach a provider can simply Grant another snowflake customer access to view and query the provider's data directly obviously with all the security controls you'd expect this data can be joined to other internal data that the customer has for further analytic use cases and because the data is not copied or moved it is always fresh and up to date so how do app Builders monetize on this first there's our data marketplace where providers can advertise and list public data sets through the nearly 5 000 snowflake customers this could be Benchmark data anonymized data from the application Etc factset is a great example of a company that has a multitude of listings on the marketplace today additionally we have seen that SAS applications are becoming the new data silos and companies are trying to figure out how to consolidate this data internally using a direct share which can also be listed in the marketplace providers can give their customers direct access to their unique data from the application braze is an example of a customer that has built both direct and Marketplace listings in either example snowflake has a monetization feature that lets the provider charge per query based on fixed monthly usage or a combination of the two additionally providers can contract directly with their customers for flat rate pricing a huge plus is there's no additional cost to the provider for sharing data it is a simple Grant to open access the last example is what we call a reader account we often hear from application providers that their customers love the product but they'd like to bring their own bi tool analyze the data a little differently or even bring their own ML and AI models to the data if the End customer is not on snowflake the provider can create an additional snowflake account share the client's specific data to that account and provide full access for the customer to query that data and do what they want there are additional costs associated with the reader account because every time that in customer runs a query it consumes compute resources this is individually tracked by Snowflake and providers are charging directly for that usage send a flat rate with limits that snowflake monitors or a combination of the two seismic is a good example of a client using direct data sharing and reader accounts so what are the key takeaways first as Frank showed investors are rewarding these models because of the growth in net revenue retention rate that these models Drive second selecting the key value metric or metrics is the key to making this work when done properly it creates a natural alignment between the vendor and customer that contribute to these sticky relationships lastly snowflake is uniquely positioned to support these go to market models because of our single platform approach usage-based pricing and data sharing thanks everyone for tuning in to this session I wish you the best of luck as you continue to navigate and evolve your pricing strategies