Welcome! This week we’re talking with Tomasz Tunguz, Managing Director at Redpoint Ventures. If you’re not familiar with his writing at TomTunguz.Com, you should be! His blog is mandatory reading for anyone interested in data-driven insights on Startupland, from investment trends to operational best practices to book recommendations. A long-time FOZ (“Friend of Zuora”), we called up Tom to ask him for his thoughts on pricing models.

Tien: Welcome, Tom! Lots of people tend to think of subscription and usage pricing as an “either/or” issue, but I tend to think of it as “both/and.” In fact, our own research shows that subscription companies that take advantage of both tiered and usage pricing tend to do better than companies that pick just one approach. You talk with companies all the time. What are you seeing out there?

Tomasz: I think it really depends on whether you’re selling an application or an infrastructure or middleware product. At the application level, we’re finding that buyers are looking first and foremost for predictability in their expenses. Most of the people we talk to, particularly CFOs, want to know exactly how much they’re spending each month. That’s why seat or tier models tend to dominate in applications, because they’re inherently more predictable. It’s harder to be predictable with usage, unless you have a nice longitudinal data set that allows you to be more predictable.

There are exceptions. Mailchimp comes to mind. They’ll put you in a different pricing bracket depending on how many emails you send, which certainly qualifies as usage-based pricing. But those buckets are very simple and visible. If you’re approaching a limit you should be able to predict when that next pricing tier is coming. You’ve got a total number of subscribers within a given range, and so you have a fairly simple staircase function.

Tien: It’s funny. I was talking to an executive at a cell phone company recently, and he said that in their industry, no one is interested in usage anymore. The rating engine doesn’t even matter. Some of our younger readers might not remember this, but calls used to be “rated” based on the length and the distance involved so that calls between states were higher than calls with a metro area. And today that’s all gone. You’d rather just buy a cell phone with an unlimited calling plan, rather than buy minutes or phone cards. And on the consumer side in general, you’re not seeing Spotify charge by the song, or Netflix charge by the show! Same with the cable companies and ISPs. People want flat plans. So it really depends. You get the sense that it’s a pendulum swing. Tell us about infrastructure.

Tomasz: Sure. So you’ve got companies like Stripe and Twilio and Snowflake that are all building various types of infrastructure and middleware: payment, SMS, database. And they’re following the lead of the big cloud infrastructure companies like AWS and Azure, because that market is more accustomed to usage-based pricing. Again, there are always exceptions.

And so the interesting question is, are those usage-based dynamics going to come to the application world? Slack, for example, is a well-known example of a company that won’t charge you for the users that have signed up, but aren’t actively using the product. And by doing that they nullify the issue of unpredictability in your expenses, because you know there’s always going to be a maximum spend. Another example in the application tier would be Expensify, where if you don’t file an expense report in a given month, you don’t pay.

But I will say this — We’re not hearing from our portfolio companies who sell SaaS that they want to move to activity-based pricing. They’re not moving in that direction, and their customers aren’t asking for it, at least not yet. We are, however, hearing from our billing vendors that usage is the future! So it’s still very much a debate.

Tien: It’s really about this tension between wanting predictability, but also not wanting to overpay. You talk to all sorts of companies about this issue, and you’re telling me it really depends on the product.

Right now the infrastructure companies are appealing to the aversion to overpayment, which is great if you’re a startup or a small company, and you really need to watch your spending. But at some point, as you mature, you’re going to want to know what your monthly AWS spend is going to be. That’s going to be a big number. You’re going to want to flip over to the predictability side.

That’s why AWS incentivizes reserved capacity. They tell their customers that they can save 75% over equivalent on-demand capacity if they pay up-front. The bigger the up-front payment, the bigger the discount. The idea is to minimize risk and manage budgets more efficiently. Lots of companies start off with pure usage then eventually tilt to reserved pricing, or some combination of both.

Tomasz: That’s why it’s so important to maintain flexibility, and understand that your needs change over time as you and your customers mature.

Tien: Exactly. Our Subscribed Institute put out a benchmark report on this issue, and they found that companies that have usage-based pricing making up between 1-25% of their overall revenue grew 1.5 times faster than companies with no usage-based pricing at all. But guess what? They also grew faster than companies with more than 25% of revenue coming from usage. So there’s a sweet spot for every company, and it’s going to change over time. Thanks for the chat, Tomasz!

Tomasz: Thank you, Tien.

Disclosure: These opinions expressed are mine, not those of the company. The companies mentioned in this newsletter are not necessarily Zuora customers.