Snowflake on the advantages of consumption-based pricing

Founded in 2012, Snowflake delivers the Data Cloud, a single data experience that spans multiple clouds and geographies. The company provides storage and analytics services through a single, global platform to more than 7,000 customers as of October 2022.

We spoke with Michael Scarpelli, Snowflake chief financial officer, about managing financial data, its partner ecosystem, and how its integration with Stripe provides near real-time data analysis.

Products used

    Data Pipeline
North America

You’ve led finance teams at several of the biggest cloud-computing companies. How did your journey take you to Snowflake?

It’s actually pretty simple – I’d been CFO for Frank Slootman, the chairman and CEO of Snowflake, at two prior companies that we’d taken public. I was going to retire, but Frank approached me and asked if I would join Snowflake with him. I really didn't know anything about Snowflake other than seeing their billboard on Highway 101 in San Francisco, but as I dug more into the product, I realised what a massive market opportunity there was. And it’s turned out to be a great ride the last three years.

One thing that makes Snowflake unique is its consumption pricing – customers only pay for what they use. I imagine that can make financial planning more difficult.

We have a consumption model which we use to try to understand how much our customers will consume. That was probably the biggest challenge at Snowflake, coming on board here, really getting comfortable around that internal forecast.

But even more importantly, one of the biggest challenges is that, in a model where you book a deal and you know you're gonna get the revenue, it's very easy to compensate your salespeople. But in a consumption model, you want to make sure there's an incentive for the reps to be involved in getting the customers to use the product. And so we had to change our compensation model so it wasn’t just paying on the booking. Instead, a big portion was paid on the actual consumption. And that was a big shift, to get the sales people in that mindset.

How do you do that? How do you convince sales that encouraging usage is part of their job?

We don't have a separate customer success organisation to communicate with customers post-sale. Everyone in the company owns customer success, and salespeople are involved because their comp is tied to revenue. And we convinced them of the benefits of spending more time with their customers – the more time you spend with the customer, the more opportunities you will recognise for them to consume, which will lead quicker to additional sales with those customers.

How would you describe your leadership approach in situations like this?

I'm extremely direct, but I don't micromanage. I believe in going directly to the source if I see something unusual. Rather than asking my sales leader to follow up, I call the individual sales representative directly to work through a problem. Companies also need employees to take risks. If I have to micromanage them, it removes the opportunity to take that risk and learn from their mistakes.

Do you use Snowflake’s data analytics platforms to manage your own data about how Snowflake is doing?

We pull all of our data – from contracts to customers to metering – into Snowhouse, our globally distributed data platform. We then look at a customer’s historical consumption patterns to predict their revenue for the next year, and this data gets refreshed daily. I can see how every single customer is performing relative to our forecast. I can also review daily revenue, comparing it week over week or year over year. I can look at all that, slice and dice it, do whatever type of graph I want to do. The visuals are really good.

It sounds like you’re saying to potential customers that even their CFO can use Snowflake’s tools.

Absolutely. Anyone in the company.

Is it hard for customers to understand the difference between billing models? Is it hard for Snowflake to sell the consumption-based model when others are offering Software-as-a-Service (SaaS)? Have you ever thought about going over to SaaS?

Well, in a model like Snowflake where we use a lot of compute and storage from vendors, the only way we can ensure we're covering our costs is to have our pricing to our customers tied into that actual usage. But it’s useful for the customer too. We sell one product, unlike a SaaS company where billing is tied to the number of users on a system. Under the SaaS model, a customer has to go through the procurement cycle again if it wants to grow or if a new feature is released. With Snowflake, that step is avoided, making it easier to roll out new features, which often drive consumption.

What would you recommend to a CFO wanting to transition to consumption-based pricing?

The first is determining your unit of consumption. It has to be one that is easy for the customer to understand. The second is going all in on the consumption model rather than doing a hybrid, because in a hybrid your users end up arbitraging what’s in their best interest, in a way that makes it challenging to predict your revenue. And then you need to show customers that every year they can do more with the same price as the year prior.

How does Snowflake think about its partner ecosystem, particularly when it comes to choosing partners to integrate with?

We evaluate partnerships in two ways. One is whether they drive Snowflake consumption. The other is whether they align with our customers. Stripe is a prime example – customers wanted to integrate their own Stripe data into our platform, and Stripe and Snowflake have over a thousand mutual customers. Traditionally, to get data from one application into Snowflake, companies will have to use various ETL tools or build custom pipelines. But now we’re building standard pipelines to seamlessly integrate sales data into Snowflake to run analytics.

Are there pain points that Stripe and Snowflake are solving together?

The biggest pain point that we hear from customers is getting data from one place to the other so you can see it in real time. But with our standard pipelines, it’s seamless. Stripe and Snowflake working together means that you can do real-time analysis of your data. That can help with fraud detection, customer purchasing patterns, whatever you as the customer want to do.

You can also integrate that with other data sources, right?

Absolutely. You could add your Salesforce data, your CRM data, so you see both your sales and your marketing data, so you can tie purchasing back to targeted campaigns.

Finally, how do you see the partnership with Stripe helping joint customers navigate the uncertainty of the current economic situation?

By getting near real-time data from Stripe, customers can analyse and understand their spending – having a joint Stripe and Snowflake dashboard can show you economic indicators or payroll data by region. You can do that analysis to help you understand where your weaknesses are. And that makes you better prepared for any economic slowdown.

Always know what you pay

Integrated per-transaction pricing with no hidden fees

Start your integration

Get up and running with Stripe in as little as 10 minutes