What Execs Need to Know About Usage-Based Pricing
Updated: Apr 7
Usage based pricing has been generating a lot of buzz lately and with good reason. AWS, Snowflake and Twillo are some of the most well-known companies leveraging usage-based pricing (UBP) and are often held up as examples of how advantageous it can be. So it's no surprise that, due to their very public and successful adoption of UBP more companies, large and small, are starting to experiment with the potential of metered pricing.
In fact, based on OpenView Partners most recent “State of Usage-Based Pricing Report,” culled from nearly 600 participants, 45% of SaaS companies offered a UBP model in 2021.
Unlike legacy subscription models, UBP enables businesses to more quickly and easily adapt to changing customer needs and business requirements. It's appealing to potential customers because it shortens purchasing cycles, eliminates high up-front payments for price-conscious businesses, simplifies procurement and contract amendment; and allows for cost optimization. Conversely, it's appealing to UBP companies because it enables true product-led growth, and increases the monetization opportunity for various products and services. (Think of this analogy: which is going to be better for the diner and the restaurant - a buffet, or an a la carte menu?)
In March 2022, Snowflake adjusted corporate guidance for FY23 revenue growth of “just” 65-67%, which is far below historical growth. Alternately, Twilio saw a stock jump when the company generated 10% more revenue than previously predicted.
Both companies missed their expected guidance. Why? Because forecasting is the Achilles heel of UBP. And since UBP happens AFTER usage is completed, all of the usual subscription billing techniques are rendered useless.
For instance, you know your number of usage transactions for the month runs into thousands, and there can be 100s of thousands of transactions. But, how can you tell in advance if that will be 100,000 transactions, or 1M?
Snowflake's CEO, Frank Slootman commented: “We report revenue on what people are actually consuming during the quarter. We have tons and tons of customers that we have zero history with that we somehow have to project exactly what they're going to do and how they're going to grow.”
Couldn't have stated the problem better myself.
Without better visibility, it’s incredibly difficult to predict revenue with UBP and many companies will project conservatively to be safe. I'd hate to be a CEO who answers to shareholders, or worse yet the CFO that rolls-up to them, when those conservative projections miss the mark.
So how can you solve for this predictability problem?
Companies must monetize the consumption of services/products differently based on type of service. This requires that the product and pricing teams model various monetizable dimensions that are applicable for each service. In addition, Engineering teams must instrument their products to track these dimensions (or any other intermediary dimensions) and then aggregate, transform, and enrich the data to transform it to billable formats.
Finally, to effectively embrace the business benefits of usage based models, it's critical to have a sophisticated rules engine that can process these massive volumes of structured and unstructured (multi-metric and multi-dimension) consumption data, and then identify, transform, cleanse and extract meaningful data for billing purposes (mediation), and determine how much those actions are worth (rating). Using machine learning, these complex processes are simplified and automated and can be scaled across millions of unique users and their transactions. The end result? Consumers can be charged for services the way that makes the most sense for them (think about Liberty Mutual spending millions to advertise “only pay for what you need”) and businesses will be able to meet their revenue goals, while presenting accurate real-time forecasts to their shareholders and stakeholders, including Wall Street.