April 28, 2021 - Many companies are undergoing digital transformation and converting their traditional on-prem solutions to an as-a-service product (SaaS, PaaS, IaaS) delivered through the cloud. These products are typically delivered through a subscription model and results in a very attractive recurring revenue stream for the companies. Different XaaS (used to encompass all as-a-service products) pricing models abound and the more sophisticated companies with more complex products usually choose a combination of approaches.
In subscription businesses, it is critical to be thoughtful about pricing and have a monetisation strategy from the start because the customer is constantly assessing the value of the service with each periodic payment.
In fact, data from Profitwell shows that the right monetisation strategies can be 4x more effective than acquisition efforts and 2x more effective than retention efforts in improving the bottom-line of your product.
A cost-plus strategy might work for some industries but not for XaaS given that the marginal cost can be very low. I have also seen teams decide on a price based solely on what the competitor is charging, which does not extract all the value on the table especially if you have a differentiated product. The strategy that derives most value for XaaS (and indeed all innovative products) is a value-based one. It requires a deep understanding of what the customer values and their willingness-to-pay (WTP).
In Part I, I will touch on the different pricing models employed by most XaaS companies, with a focus on usage-based pricing which is seen as the next evolution in XaaS. In Part II, I will be sharing the importance of incorporating value and WTP discovery early in a product innovation process and a framework for doing so.
XaaS companies, particularly those with a more complex suite of products typically employ a pricing model that involves a combination of the following:
1. Per user: Typically a flat rate with revenue scaling linearly. The advantage is that it is simple for the customer to grasp. However, it tends to be seen as an additional marginal cost for every additional user which limits adoption. Customers also typically try to “share accounts”. More importantly, charging per user is less useful as more layers of the tech stack get automated.
2. Freemium: Can be very powerful when the TAM is huge and the strategy is to “land and expand” quickly. The intention is to provide enough value for free to attract people into the ecosystem and then provide an on-ramp to a paid product, with ostensibly more features and support etc. Due to the virality of a freemium model, it can also have very low customer acquisition costs (typically 50% lower) which leads to stellar unit economics.
Companies that have mastered this well include Zoom (free 40mins), Slack (communicate with co-workers for free) and Calendly (schedule unlimited events for free). Successfully executing on a freemium model requires a very thorough understanding of each user segment, what they value and what their WTP is for that value. It also requires a very large TAM (typical free-to-paid conversion rate is 2–5%) with a large enough product scope such that both the free and paid versions can be deemed valuable. Oji Udezue who used to head up product in Atlassian and was VP of Product at Calendly shares that one tip is to find features in the usage data that is used by the 50th percentile of users (thereby embraced by a broad segment of the user base) and consider setting it free. Since individual end-users might only be able to make small purchases, freemium could also be used as a lead generation tool, where companies that have a large base of free users could be cross-sold a more expensive enterprise plan with more features.
In today’s environment where there can be multiple companies competing for the same slice of the market and it is easy to copycat features, customer WTP could decline quickly with time and so there also needs to be this laser focus on customer WTP deltas should freemium be an arrow in your quiver.
3. Pricing pan flute: Tier the offering into a “basic-premium-pro” type pricing grid, with the price increasing as the “flute” of features get longer with each tier. Once again, this model is predicated upon knowing customer segments, values and WTP intricately well.
4. Usage-based: There are many advantages to this model. It greatly reduces friction to adoption, allows for the seeding of more use cases within a company with the optionality of growing revenue as usage scales (potentially hitting unexpected home runs). TAM is further broadened with increased accessibility to more cost-conscious customer segments e.g. SMEs. Customers find this model attractive as they can control costs more effectively during a downturn. You also end up with lower churn as flat-rate subscriptions are easier to rationalise away when cost cutting.
Usage-based pricing models have indeed corresponded to greater revenue growth, margins and valuation multiples.
Businesses it is best suited for
Usage-based pricing is very suitable for products that are very sticky once adopted, with usage increasing over time. An example will be SaaS tools that are embedded in a customer’s workflow. OpenView shares that Twilio’s communication API is used by developers to quickly prototype ideas and develop applications. Developers pay once the app is live with customers and do not want to undo it. In fact, the more usage pricing can be incorporated as a COGS (for e.g. a payment SaaS product that charges per transaction), the more resilient the revenue is as companies tend to niggle subscription line items in Opex while seeing COGS items as part and parcel of doing business.
Usage-based is a less effective pricing model when customers feel the marginal cost of consumption. For e.g. Hired found that recruiters only sent them a fraction of their open roles when pricing was on a “pay per hire” basis. It is also less effective when there is a lot of seasonality or variability in the customer’s business and revenue does not scale well over time.
In fact, usage-based pricing is not a new concept. It has been used to great effect in industries outside of software. Michelin is a great example. In 2001, they invented new tires that lasted longer than their competition, but it was difficult to charge premium pricing to extremely price sensitive trucking companies. Longer lasting tires would ironically also cannibalise their future sales. They then decided to charge fleets by the mileage driven on their tires instead of by the number of tires bought; a pricing model which back then was considered disruptive. They had figured out what was being valued by the customer (miles driven) and used it as a sales growth lever. By 2011, they had the greatest profits in the industry; 3x that of Goodyear’s.
Today, we see elements of usage-based pricing models in industries like car insurance and telco services. With the advent of IoT devices as well as more companies that can help cloud companies meter usage across all kinds of metrics, it is becoming easier to track usage and bill on it. The strength of this pricing model in creaming off value as businesses scale behooves us to consider imputing an element of it in our monetisation strategy if possible.
What is a good metric to use?
The success of a usage-based model lies on finding the right metric to meter. This value metric should:
1) Align with the value the customer receives (e.g. a per square foot pricing for a smart building management platform aligns well with how lessors make money)
2) Be easy for the customer to understand (e.g. GB of data processed)
3) Be flexible where customer can choose and pay for their exact scope of usage
4) Scales over time as the customer uses more of the product
5) And obviously, possible to meter
Companies could charge solely based on a usage metric. However, it is more prevalent for companies (particularly those that have a more complex product suite) to incorporate a usage element in their pricing plan, e.g. Salesforce has subscription tiers that scale pricing with usage volume. Even the likes of Uber combine both dynamic pricing with usage-based rates that scales with variables like time and distance.
Zoura, which is a platform for subscription businesses, has done some research on 900+ companies to show that there is an optimal proportion of revenue that usage should account for. Activity on their platform showed that companies that had usage make up between 1–25% of their overall revenue grew revenue 1.5x more than those that did not have a usage element, and 1.2x more than those with more than 25% of revenue coming from usage. It is thus important to strike a right balance between usage-driven growth and sustainable recurring revenue from subscription as we roll out XaaS products.
Finally, usage-based pricing is not a decision taken within a silo. It involves cross-functional teams being willing to overhaul certain processes or the way things have been done.
1) Finance would need to be prepared for a different revenue trajectory at the start and potentially different revenue recognition policies.
2) There needs to be systems in place to accurately meter usage, give customers the ability to access usage in real-time, and provide for transparent billing.
3) One key pushback from customer’s finance departments could be the unpredictability of costs under a usage-based model. There needs to be a plan to tackle overages with an emphasis on providing flexibility to give customers a peace of mind. Some companies have allowed for rollover of unused credits, or early warning to give customers option to upgrade, or to have a different pricing structure for overages.
4) Sales and marketing would need to be motivated and compensated differently. Instead of one upfront commission, they might need to be paid as usage scales to incentivise continual involvement with the customer to drive usage.
5) As more customer data is procured, data analytics should be deployed to further refine the pricing strategy.
When it is all said and done, the success of the monetisation model chosen ultimately rests on how well we know our customer segments, what they value and their WTP. A framework to establish that will be covered in Part II.