Usage-based pricing is having its moment in the sun. Leading pricing experts have been saying that usage-based pricing is the path forward and that subscription pricing models are oversubscribed. Foremost among these is Kyle Poyar at OpenView, who has recently published a series of posts on Tech Crunch on why product led growth companies need to move to usage-based pricing now. I would add that this is just as important for service-led growth companies.
What is usage-based pricing?
Usage-based pricing is not new. In many ways it is the oldest approach to pricing: “pay for what you use.” Zuora, the company most closely associated with the subscription economy, has known for years that having a usage-based component in the pricing model led to higher growth.
Usage-based pricing simply means that some variables in your pricing model are based on actual use. There are many ways to implement this.
The most basic is transactional pricing, where one only pays when a transaction takes place. There can be a fee per transaction, or the fee may be tied to the size of the transaction. The fees paid when real estate changes hand are an example of this.
Usage-based pricing can also be based on a user taking some action on the system. This could be as simple as making a click or as complex as completing a business process or value path (more on value paths later).
It can even extend to third-party use of your system. This is relevant for solutions that are meant to support your customers. It may not matter what you do, as value is created by having other people take action. An example from advertising is “shares.” The number of times other people share your content (which can be tracked) is a good measure of value in many social media advertising applications.
Why usage-based pricing?
Usage-based pricing leads to higher willingness to pay by reducing the risk to the buyer. In every purchase, there is some form of risk discount. The greater the uncertainty on whether a new application will be used, or if use will create value, the greater the discount. Shifting this risk from buyer to seller leads to higher prices and a higher willingness to pay.
This seems rather obvious, so why is usage-based pricing surging today?
Why usage-based pricing now?
There is a short-term trigger for the shift to usage-based pricing supported by a long-term trend.
The trigger has been the impact of the Covid-19 pandemic. In March of 2020, one year ago, we published an article (“Pricing under uncertainty and the need for usage-based pricing”) in which we predicted that smart companies would respond to the pandemic by introducing usage-based pricing.
Now that the end of the pandemic is in sight and the new normal is establishing itself, usage-based pricing is establishing itself as part of standard business practice. SaaS companies that do not include a usage-based component to their pricing are being questioned. Within a few years we predict that virtually all companies will have a usage variable in their pricing models.
The pandemic was the trigger, but the move to usage-based pricing is supported by larger trends.
- Availability of data
- Better prediction
- Causal analysis
Availability of data
Cloud software, mobile applications and the Internet of Things (IoT) have given us a world of ambient data. There is a growing amount of data available to inform tracking of usage and even of outcomes.
Usage-based pricing depends on good prediction. One of the most common reasons not to adopt usage-based pricing is predictability. The current generation of deep learning Artificial Intelligence (AI) has radically reduced the cost of prediction, as explained by Ajay Agrawal, Joshua Gans and Avi Goldfarb in their excellent book, Prediction Machines: The Simple Economics of Artificial Intelligence. As it becomes easier to make predictions about usage, usage-based pricing becomes more attractive.
Usage-based pricing is a step on the path to outcome-based pricing (see “The Ends Game – How Smart Companies Stop Selling Products and Start Delivering Value” by Marco Bertini and Oded Koenigsberg). To cross the chasm from usage-based pricing to outcomes (or results) based pricing, we need to go beyond prediction to causal analysis. Fortunately, this is developing almost as fast as prediction. Innovations from Judea Pearl and his colleagues are already transforming how causality is measured and allocated in healthcare (this is the discipline of Health Economics and Outcomes Research or HEOR) and these approaches to evidence-based healthcare will spread across B2B over the next few years.
Prediction + Casualty + Understanding of Value → Outcomes Based Pricing
Why not usage-based pricing?
When you introduce the idea of usage-based pricing inside your company and with customers you are likely to encounter the following objections:
- We don’t know how people use our products – OK, solve that first, but you had better solve it or you will lose the innovation game.
- Usage does not correlate with value or outcomes – Maybe, but see the note on value paths.
- We have no control over if or how our products will be used – Yes, but that is why there is a risk discount on your price, and user experience design is advancing fast enough that it should be possible to design software people are willing to use, if it provides value that is.
- It is too difficult to predict contract value – Historically, this has been the best reason not to adopt usage-based pricing, but (i) prediction engines will solve for this and (ii) contracts can be designed that protect both sides against this.
These are all legitimate objections (at least 2 through 4 are) and designing software, user experience, prediction engines and pricing models to address them are the key to usage-based pricing. Read through these again. You want to know how people use your products, to deliver products that are creating value, that people are eager to use and where usage can be predicted.
Not all usage matters
Modern software platforms can collect a lot of data. Every click is logged; in some cases, every trace of the finger, swipe of the thumb and track of the mouse are picked up along with location data, who else is on the system and on and on.
Some of this data is just noise. Collecting it may be in violation of privacy rules or data ownership contracts.
Be selective about the data you collect and the data you use for usage-based pricing.
Usage-based pricing is not all or nothing
Usage-based pricing does not mean only usage-based pricing. Zuora has found companies where usage-based pricing accounts for about 20% of total revenues. Complete reliance on usage-based pricing can actually cause growth to stall out. Begin your usage-based pricing by layering it in, and not with an abrupt switch. The first versions of value-based pricing should be skewed towards the buyer to give them an incentive to change.
Begin by keeping it simple. Pick one usage metric that:
- Is easy to measure
- Tracks value
- Can be predicted from other usage data and/or other firmographics
Over time, layer in additional usage-based metrics or combine two or three metrics into a composite metric (composite metrics are often easier to predict and better track value).
In many cases, one has to complete a set of actions, a business process so to speak, in order to create value. These processes, that end with the clear creation of value, are called value paths.
A value path is a series of actions on the system, sometimes taken by more than one party, that when complete generate economic, emotional or community value.
In the future, usage-based pricing and value-based pricing will converge on shared value paths.
Usage-Based Pricing 1-2-3
- List the usage metrics that you can:
- List the value metrics (the unit of consumption that track how the user gets value)
- Look at the intersection set, the usage variables that are also value metrics
- Make sure that variables from the intersection set are used in your pricing mode
Not everyone will be ready to jump to usage-based pricing. But you need to begin preparations for this journey. You can do this by:
- Instrumenting your software so that it gathers the usage metrics that track value
- Adding in third party integrations to deepen the data set
- Set up a prediction machine (your product engineers can do this easily using a combination of Keras and Tensorflow) to predict usage
- Develop some trial usage-based pricing models, predict their performance, and then track what actually happens (and use this to improve your prediction models)
The future belongs to the companies that can take on the most risk because they are best able to manage it. Usage-based pricing is an important step along this path.