After You Announce: How to Win With Usage-Based Pricing
Usage-based pricing (UBP) is about change. Once you have launched your pricing model, a lot will change, including the initial usage-based pricing model itself. Crafting the best messages for your customers is only one part of a successful transition. In order to win, you have to be prepared to learn and adapt on the fly based on any knowledge you gain along the way.
If you want to simply set it and forget it, then usage-based pricing is not for you. Once you move to UBP, you’ll always be thinking about the consumption and the price, how the two things shape each other, and where you can iterate to improve. But that is the nature of the beast, as UBP helps you to adjust:
- Customer experience;
- Customer value; and
- How you respond to changing conditions.
What do you need to learn?
When you move to usage–based pricing, you need to learn how to:
- Connect customer experience to price to understand how much you’ll get in revenue;
- Understand the relationship between usage and value; and
- Predict usage with increasing accuracy.
These are all underlying principles of product-led growth and among the reasons why usage-based pricing and product-led growth are increasingly found together.
Connect customer experience and price
Get your customer experience team involved in pricing. They have tools and ways of thought that can be used to improve your current pricing models, such as the customer journey map. You can layer price and value onto the customer journey map so that it represents the customer value journey. Make sure to track use and the impact of use on revenue as part of the customer journey design. Customer success platforms such as Pendo can be used to define the value paths that are used as usage-based pricing metrics.
Know how the usage-to-value correlation is changing
Usage is basically a proxy for value and how your customers get value from using your product and solution. This is not static though, as it also changes over time. Develop a formal model where you can track changes in the relationship between value metrics, usage metrics, and pricing metrics.
Value models are typically developed by value engineers, either internally or at consulting firms, and can be kept as spreadsheets or in customer value management platforms.
Regularly ask yourself:
- Are my usage metrics tracking value?
- Are my pricing metrics tracking value?
- Am I capturing the right data from my application to understand value?
How to predict usage
One of the most important capabilities to develop in usage-based pricing is the ability to predict usage. This is also referred to as Predictive e or predictive engagement. Predicting usage is important to both buyer and seller. Buyers need to know how much spending they are committing to, and sellers need to be able to predict revenue.
Modern machine learning makes this possible. Back in 2018, in their book Prediction Machines, Ajay Agrawal, Joshua Gans, and Avi Goldfarb asked, “What happens when prediction becomes cheap?”
Four years later, we have the answer. One option is that usage-based pricing becomes compelling for both users and businesses. Companies adopting usage-based pricing also need to invest in usage prediction. They can do this by developing applications that gather the appropriate data needed to make these prediction engines spit out useful information.
Unlocking usage-based pricing with machine learning
The initial usage-based pricing model will never be perfect—no matter how much research is done or how smart the designers are (and there are some very smart people working in this field). In general, people are always learning with UBP, and quickly at that. UBP means that your billing changes with use, and this means learning is critical to success.
Three common challenges for usage-based pricing also occur with machine learning. Understanding these ‘traps’ is important to designing and then evolving value-based pricing.
The below figure is from Philipp Koehn’s excellent book Neural Machine Translation:
All three of these challenges (too high learning rate, bad initialization, and local optimum) apply to usage-based pricing.
In order to overcome these:
- Companies need to adjust usage-based pricing in a sensible and measured way and not jump all over the place.
- The initial model needs to be reasonable and a good first approximation (otherwise you will never get to a good pricing design, you will have to toss the whole thing out and start again).
- You have to be careful not to get caught in a local optima (a solution that is better than other similar solutions but that is not the best solution) and thereby miss the larger opportunity.
Based on my own informal study of SaaS pricing models over the past decade, I estimate that almost all good pricing models–including those that I have had a hand in–are caught in local optima. Finding ways to escape this trap is an important part of any good pricing strategy.
Make sure UBP encourages positive changes over time
There’s no doubt that a customer’s usage can change over time–and the introduction of usage-based pricing will change usage. This can be a negative feedback loop, where UBP discourages use, or a positive feedback loop where usage is encouraged. If it isn’t obvious, it is important to design usage-based pricing to encourage use.
But how? Use has to correlate with more value. In an ideal world, the value provided to the customer grows faster than the value to the seller captured through pricing.
With UBP, the design goal is to have both price and value grow over time with value growing faster than price (as seen with the left and middle figures above). This is the only sustainable value-based pricing approach.
Often it takes time for value to catch up with and overtake the price. Time to value is always important in pricing, but the importance is more apparent in usage-based pricing. A lack of value can discourage use, triggering the negative feedback loop.
One strategy to overcome this is by only having usage-based pricing kick in after the value is greater than the price being paid.
When price grows faster than value, the account will eventually churn. In this case one should either reduce the rate at which use drives price or abandon usage-based pricing altogether.
Time series analysis
Being able to understand any patterns in use over time, also known as time series analysis, is another key capability for usage-based pricing companies. Time series analysis is used to tease out causal patterns. Ideally, your value and pricing software gathers the required data and analyzes trends. The most important trends are:
- How well use predicts value
- How price and value change together over time
- Changes in slope (inflection points) where the rate of change is speeding up or slowing down
- Intersection points, such as when value rises above or falls below price
Staying on top of tools and trends for usage-based pricing
As the UBP landscape changes, make sure to keep up with new frameworks, tools and best practices.
Good frameworks help with design, planning and execution. Tools are needed to track trends in usage, value and price and to understand Predictive e.
Any best practices should be shared so that we can all get better at this critical capability, and grow the overall value of B2B software and data ecosystems. The more value created for customers the more value we will all be able to capture. And any surplus value gets invested in continued innovation.
This is the third in a three-part blog series on launching usage-based pricing. The three installments build on each other.
Does automating outbound sales efforts really work? It can and it does, as shown in this post by Thena. They use automated outbound to 10x their efforts. Here’s how they do it.
How do you find and hire a sales leader who can thrive in today’s rocky selling environment. Expert Amy Volas lays it out here.
B2B SaaS companies need to not only learn how to “do more with less” but also “do different with what we have.” This three-part framework can help.