Predicting Revenue in Usage-based Pricing

0:00I happily accept some of the risks of a consumption -based model because I think that the benefits far exceed the cost. These providers need to be held accountable to continuously delivering value. It is not okay to simply sell a deal, walk away for 11 months, and then one month before the renewal is set to go, then you re -engage and say, hey, how was the last 11 months? If we would have just asked that question, they would have told us. but instead we put it down on our chart as a trend that would endure for the next year and we called it ARR and that's a mistake. I think actually what you're going to see is more hybrid pricing models.

0:40It involves also telling them proactively how to spend less on your company by implementing some best practices that will reduce their consumption. There is no shortcut to creating long -term successful businesses. Pricing is hard, which is why so many companies have defaulted to standard pricing models like subscription. And that should come to no surprise, because predictable revenue is the linchpin of any company's planning, execution, and ultimately evaluation. But, it also happens to be one of the most difficult things to nail about implementing another pricing model. That is usage -based pricing, which is what we're here to talk about today.

1:21Because once you've established the right processes, Oregon compensation structures, and tech stack to operationalize it, your revenue can actually become more predictable with usage base pricing than it might be with traditional SaaS over time. So, today you'll hear from A16Z Growth Partner Mark Reegan as he sits down in Travis Furber, VP of Strategy at 5 Tran, and Dan Burrell, head of sales at Alchemy. Who both have implemented and embraced, you guessed it, usage -based pricing. So today, these are your guidance on best practices, the very real ups and downs of usage -based pricing, key metrics to hone in on both short -term and long -term planning, and ultimately why it's so important to orient your business around the value that a customer is getting.

2:06The first way to hear is Mark, then Travis, then Dan. Oh, and for more content just like this on usage -based pricing, make sure to check out a16z .com slash pricing dash packaging. Enjoy! As a reminder, the content here is for informational purposes only. Should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16z fund. Please note that A16z and its affiliates may also maintain investments in the company's disgust in this podcast. For more details including a link to our investments, please see a16c .com slash Dispouchers.

2:51What I'd love to hear from you guys first is your perspective on why usage -based pricing has become so popular in the industry and why is it working so well at each of your respective companies? I think it comes down to usage -based pricing allows customers to pay for what they use. It helps tie value directly to the product. And from a customer standpoint, it's also an easy way to help customers come in and experience your product without making a big commitment to that. So it's really helpful in landing customers and bringing them in. And I think over time, with that, you can build stronger relationships with those customers.

3:31And you can see that growth or time. There's maybe one other thing I'll add to that, which is UCBase pricing forces you as a company to think about the customer all the time. Every part of the organization has to be thinking about the customer. And a booking is a model where you come in and you get them a subscription. The sales person is like, great, did my job. I'll see you in 10 months where we're getting ready to start talking about that renewal. And your engineering team is somewhat tied to them, but you don't get the immediate feedback with the customer. You don't get the immediate, are they being successful?

4:02How they adopted the product and the way that you thought. And with usage -based pricing, you can see that. You get the telemetry data. You get the information right away and you can focus the entire organization to make sure that those customers are successful. And I think the founder or an individual who wants to build a company that drives great value with customers and great relationships with that, usage -based pricing is a good mechanism for aligning all the organization around the success of the customer because revenue you was directly tied to the success of that customer? I think you nailed that.

4:33I've been in Silicon Valley venture -backed companies for 12 -ish years now. What's really interesting when I think about this question is that I think back to my very first sales training ever out of college where I had not yet moved to Silicon Valley, I was in a Fortune 100 company and I was given conventional sales training. In that training, I was taught how to gate information and access to information about our products and services behind all this process that we were supposed to put these prospects in potential customers through. And instantly, that never sat right with me. That never felt like a reasonable trade.

5:14I felt like we should be freely giving more access to information. And I moved to Silicon Valley and I loved the updated philosophy that I saw. At the time we were all talking about the consumerization of IT, of enterprise infrastructure. And making sure that the tools that employees want to use inside a company to do great work matched the great experience of consumer tools that were on the market that were available for people. And there was this discussion of shadow IT and people bringing in their own technologies from outside that weren't necessarily sanctioned, because these tools are so much better to use.

5:50And the key to that whole motion and what has driven this is the consumer preference and the consumer in this case could be the enterprise customer, the enterprise knowledge worker. They deserve great tools, not only do they deserve access to information about the tools that they might buy, we've now taken it and progressed it over the last ten years to include consumption of that tool. So as part of this consumption based pricing model, of course we have different tiers and access to different tiers for these solutions, including free tiers, which the internet has helped democratize access to information about all these solutions.

6:25And of course you can actually go use them, which I think is a huge benefit for customers and I think is the right expectation for the industry to have. It's going to help ensure that companies are able to find the right tool to meet their business cases and their actual needs. So I love that we've done that. Now, I may be cheating ahead a little bit on additional questions, but what I think this naturally goes is if we're giving out all this access to use our tools, with very generous free tiers and we're spending money to provide that solution for free, what are we then trading in exchange for those scale customers and long -term relationships and commitments, and how do we fuse product -led growth with the appropriate level of sales -led growth, and there's actually a ton of exciting stuff in that category where my perspective is you can do that in a really healthy way to manage your company's resources effectively.

7:20But to go back to the original intent of the question, why does this make sense? Why is this so popular? Why is this not going away? Because it's the right thing to do for customers. They know it deep in their heart that they should be able to use these tools. they should be able to have complete access to information about these tools and these providers need to be held accountable to continuously delivering value. It is not okay to simply sell a deal, walk away for 11 months, and then one month before the renewal is set to go, then you re -engage and say, hey, how is the last 11 months? Hopefully you're ready to renew and expand.

7:54That's not an okay motion, and that's not the way to maximize business value for anybody. I love both your perspectives on that and it almost seems overwhelmingly positive, right? I think we all know the living in the reality of this, especially when you're going through that arc as these growth stage companies, there are a ton of challenges with this model in practicality, right? And I know you guys have lived through it and I'm really interested in what your key observations have been around those challenges. And just as importantly, what you've seen in the organization you worked in to try to mitigate those, try to overcome those, and to really be able to operate this model at scale.

8:34Oh, man, there's a lot there. Still less unpack on that one. So yeah, a little bit of background, five -trend, when it was started, it was a book, In Space Business. We had a variety of connectors. We priced those connectors in different groups, and so congratulations, you go by your connector, go forth, talk to you in a year. And then we switched to a usage space pricing model using something called monthly active approach and that switch we had an established sort of cells culture that was this bookings based business model that switched over to usage based and that was a hard switch. Making sure that you had all the systems in place and everything else that comes along with that.

9:12And I think there's a couple of different ways to think about the challenges of usage based pressing. So one is on the systems and internal systems and processes to be able to manage that. It's a bigger investment. It is much easier to run a bookings based business from like a just a planning, comp, your internal systems. All of that is way easier, way simpler. And therefore you have to make a lot of investments into your operations teams, into the systems, until the data models that you have to run, all the telemetry data from your product. You have to have that information. You need that information to help drive that.

9:48So that's one challenge. The other side of this is that it does introduce a lot more variability, particularly when you have fewer customers. It can introduce a lot more variability into your revenue because customers can change their usage. And you have less commitment from a customer. You can see customers come and go and move up and down. So you always have to focus on value. At 5 .3 we've seen a couple of different drivers of that. There are multiple drivers on predictability. So we try to give a lot of flexibility with customers so that they can match what they need in their value to what the product is offering But in doing that flexibility that can drive a lot of variability in what customers actually are using and how they're changing that so they can Optimize a lot for the needs of their business and that can drive some unpredictability in the revenue And there's this other factor that we see which is just General macroeconomic things like just things that happen in the world so as we move data like there's stuff that can be sometimes outside of the control of the customer that can have impact their usage.

10:46So for example, we see a lot of our retail customers around November and December timeframe, huge spikes in usage because that's when all the POS systems are going. So when all their cells are happening, so you get these big spikes in usage. If you have a diversified customer base, you can sometimes nell out those spikes through broader industry diversification or through understanding and planning for those spikes. But you have to have some history, some data history to understand that. And And then the final one is, depending on what your product is, so again, Fytrend to kind of the unique thing is we're interacting with other products.

11:16So we pull data out of what's happening from other users or other applications, and then we move that data over. And so those applications make changes. And that impacts our product. So we have HubSpot, I think, a few years ago, made a change to their API. And it forced a complete resync for all of our customers, which, like, huge spike in usage across the board for all of our customers that we're using the HubSpot Connection. It's one of those things where you have to be on top of that all the time. You have to be watching the interactions and stay on top of those things so you can kind of protect the customers from these unnatural spikes that can happen.

11:50That means like more investment in your product and your engineering teams. So I think if you want to know like the complications of usage -based pricing and going from booking space, we're simpler to plan, simpler to use, the sales people understand it, quite frankly the procurement people understand it better too. They're like, I know what I'm buying. I know what this is going to cost me. I can predict this in my budget to a, hey, I'm not really sure how much I'm going to use. I'm not really sure what this is going to do for my budget over time. I'm not sure if I have control over that. So a lot of education that has to happen.

12:19Well, Travis, there's something else that I hear in that too, which I feel like our industry sort of loses sight of this a little bit in this discussion often, which is the idea that whether I choose to be a bookingspace business or a usage -based business, that somehow is the sole determining factor as to whether or not we're going to be a good company or a bad company. This is just for strategy. This is a strategy to unlock growth. I personally think that it's a very good strategy to unlock growth, but it also comes with costs and we can talk about those and we can talk about mitigation for those and you should be eyes wide open.

12:53But there is no shortcut to creating long -term successful businesses. Fundamentally, at the heart of all of this and you alluded to this Travis and I totally agree, your products and services have to deliver immense amounts of value to your customers. Plain and simple. And that is regardless of what growth strategy you choose or what sales model you choose to have. These software products are never done being built. Absolutely never done being built. And so consumption actually that growth strategy lines up beautifully with that because as long as you're continuing to build these software products and you're adding that flexibility, that feature set, that next generation of innovation, then you're able to command good margins, make customers wildly successful, and they're excited to come back and spend more and more and more on that consumption every period because they know they're getting more value than they're paying you, and that's how you build a great business.

13:44So I do think people lose sight of that. So you need to be tightly aligned with your product organization and thinking about that product roadmap, because that's going to be a much bigger determining factor. Now I think part of this question too is what are some of those costs. We should recognize that in good economic times, a consumption -based model can be a big accelerant because there is less friction to customers being able to use more consume more and therefore your company getting to make more revenue when they do that. In tougher economic times where you've got percentages of your business and your revenue that is tied to full consumption where there are no bookings commitments in place, obviously that represents and that's going to also be a less friction place for those businesses who are your customers to save money by pulling back their consumption on your service and lopping off use cases or shutting down one department's use of that solution.

14:39And we've seen that in the last couple of years. There was a ton written about that and a ton of analysis. For me personally, when I think about building a great business, first and foremost, I want that amazing product roadmap where we are so confident in the value that our solution provides. Secondarily, I happily accept some of the risks of a consumption -based model, because I think that the benefits far exceed the costs. Even knowing that there will be tough times ahead, and our customers may as a result of a need to save money and extend runway or drive more profitability, they may reduce consumption on us.

15:15I'm willing to accept that and deal with that turbulence and do right by our customers in those moments, because I actually think those moments, even though they don't feel good because maybe revenue is pulling back on our side. Those are incredible opportunities for us to build long -term trust and long -term relationships with those customers. They will remember how we treated them when they needed our help. And that will factor into their decision when times are good again, and they're ripping and they're investing in growth. They'll remember which providers stuck by them, took good care of them, and recognized that they were at a tough moment.

15:49and they needed some forgiveness or some help or some actual assistance saving money with best practices that enabled them to lower consumption of your service. And that's a separate big topic of the role of account management and customer success. But hopefully that addressed the question. Definitely. And Dan, I'll go right back to you, getting into a bit of the operational nitty -gritty of this. I'm particularly interested in your perspective as a sales leader when it comes to forecasting the business, right? and just living in the presence of a quarter or a couple quarters ahead of you. How have you learned to confidently forecast the business?

16:26You're growing quickly, but you have all these challenges of just not a heck of a lot of data in the rearview mirror. You don't have a perfect signal detection and hitting indicators. So how are you working through that? How have you learned to become confident in your forecasting? I really appreciate that question. My answer may surprise you slightly, because the key to good forecasting, even in a consumption based business, is a very healthy bookings element. And so the foundation of the relationship with our customers may still be entirely consumption based. That's how we have the conversation.

16:58That's how we eat their usage. That's how we talk about their usage. That's how we forecast their usage purely in the form of what they're going to consume. And as Travis said earlier, they're going to pay for what they use. That's the objective. Now, that being said, I think it's still totally fair and reasonable that my business values predictability, like you just talked about. I've got a job to do, which is to forecast accurately. We all know why those forecasts are so important, that enables us to make healthy forward -looking decisions about the business, how we're going to invest. What teams we need?

17:29There's a ton that requires a great forecasting methodology. Therefore, because I'm going to get a bunch of business value from a healthy forecast, I can return value to my customers who are willing to make commitments to us. And that's a super fair exchange of value. And it's on this beautiful continuum. The more flexibility that my customer requires, the more fair it is for them to pay a premium for the consumption that they're going to use. The more they're willing to commit to me and my team and my company, which enables me to be better at forecasting, the more I'm happy to return discounts and commercial incentives to them and we'll execute that on a bookings contract.

18:08So this is part of the motion that you want to breed in the sales team which is that you're continuously selling you're continuously taking care of them you're continuously monitoring their use case you're continuously forecasting with all of your customers. It is expected in any organization that I'm running that if you're taking care of a customer. You are continuously not only monitoring their use case in the telemetry that Travis was talking about, which is very important that you give yourself team and your customer success team the monitoring capabilities to understand in very granular detail how their customers are consuming products from a growth telemetry perspective.

18:48But you also expect that those folks are deeply understanding the dynamics within the customer business. What is causing that growth. It's not enough to just know what the growth rate is. I want my team to explain to me why that growth rate is. Is it because they're aggressively expanding into a new market? Is it because they just acquired another company and now we've combined two teams uses? We actually have to know why because that is the key to good forecasting. I can't tell you the number of times that I've seen this issue of massively over forecasting a given customer's usage because the team didn't understand that the behavior that customer was engaging in was a one -time thing.

19:30It was only ever gonna last for one quarter. And if we would have just asked that question, they would have told us. But instead we put it down on our chart as a trend that would endure for the next year and we called it ARR. And that's a mistake. So there's a whole lot. I could probably go on for another hour about what drives good forecasting. It's a combination of instilling in your team great discoveries, still, Scott, and an expectation that they're doing ongoing discovery to always know the business drivers behind the usage trends. You can't just know the trends. It's arming them with great telemetry tools, monitoring the I solutions to track it at a very granular level so they can get specific.

20:06And it is offering your customers fair contracts and discounts in exchange for commitments, which are really valuable for your business because you value the ability to forecast, you value certainty. and you're happy, in my opinion, to give discounts to customers who can sign up for that level of commitment of minimum amounts of consumption. The point of my own bookings is spot on. We have a couple different components to our business. We have this great big self -service group of customers that come in, use the product, never talk to a salesperson. They just go, they're like, pay as you go. But then there's this other portion of the business which makes commitments to hey, I'm going to buy upfront this much for the year and in exchange for that You built in discounts as you use more we've Disgusted come in and play for that and That bookings helps drive predictability for a portion of the business and forecasting for like how are we doing?

20:58So when we think about long -term planning, I think snowflake is famous for their RPO remaining performance obligation Like how much of people booked how much we got to how much is left and that that because a big thing of monitoring, giving that information to your customer success teams to help make sure customers are getting what they say they want. They've made a commitment to you and they've given you an indication of what is valuable to them and what their level are and you can say, are we getting there? I think the mechanics of actually building the protectability, what kind of systems you have to have in place, how do you do that?

21:27We've gone through many iterations of this and this has been an evolution over several years. We had to make major investments in our infrastructure from an analytics standpoint. So, FIJTN moves a bunch of data. So, we have a fairly large analytics team and we've built a predictive model that says, okay, based on what usage is looked like, because we've got this cohort of customers that haven't made bookings. So, based on their past usage, where are they going on an account level basis? When did they join us? So, what kind of usage curve are they on? Based on historical data that we've looked at, and we're like, okay, cool customers that are about this size in this region, they perform on this kind of growth curve.

22:05If you have enough customers, those averages will work out and you can see that as we apply those curves to these customers that come in. And so you layer those cohorts together and that gives you a predictability about what's going on from a revenue standpoint. And that's all the data science side. You can take into account what plan tier are they on? We've got five or six different plan tiers. What's their discount for each individual customer? You have to layer all that stuff in so that you can build a more accurate view of their performance over time and then take into account historical turn rates.

22:35Turn for us isn't a customer has left us. Turn for us can be, hey, I turned off a use case. So I've reduced that thing. So you want to look and take that to account. But that's our data science model. But the data science model only sees historical data and telemetry data from customers. They don't have that piece of data I'm talking about, which is the customer discovery piece. the sales insight side of this, which is the second part. The sales team has the insights on, are they gonna add another use case? Are they gonna turn a table off? They have the insights that the data team can't have. They don't know what the customers are gonna do, because they're not having the conversations with the customers.

23:10So you give the sales team, here's the predictive revenue for your customers. And then the sales team, you go like, well, actually, I know they're gonna add a new use case and it's gonna come online in the next two months. and I know that that's going to be worth this amount of money. But that then gives you the insights to then modify your data science model. And that gives you a little bit more confidence. And I can tell you in the beginning, your sales team will get a broth. Like, their predictions will be way off. And particularly, our model is the more that you use, the cheaper your usage is.

23:45And therefore, unless you're a savant, and you can do multi -variable calculus in your head, you have to have these tools in place to do that. And so you run this rigorous process where data science model comes in, sales modifies it based on their knowledge of the customer and what's moving up or down. And then they've got the tools enabled to predict or to size those different opportunities. I love everything you went through there Travis. What I'm really curious about is how that extends when you need to do annual planning and you're thinking about your longer term investments. investments, obviously that still requires you to forecast revenue going forward.

24:19How has that par laid into longer term planning, accuracy, or what else do you need to do in addition to those key concepts to do that well? Yeah, your capacity model is a very interesting thing that you have to build. And we shifted to more of a demand driven capacity model so we can look at historical demand. What have we been seeing? and then how does that demand and translate into dollars for us? And then what ramp do customers go on when they come in? So you build these like waterfall ramps. The basic outline still is the same though. It's just you have more assumptions that can go into that.

24:55We do a three year long term plan, which is as much for general growth rates, like what's the macrocon we kind of look like? It's a road map for a product standpoint. It's more of a here's where we think we're going to get to. It's not a this is the prediction. and this is like really honed in. On the annual plan, it's much more detailed. And that's where we try to hone in and we work on these other assumptions. Because you're assuming things like turn within the product, not just turn of customers. You assume things of like expansion of growth rates over time, how those been going, what have you seen in the past?