For SaaS and Cloud Computing
Table of Contents
What is Usage-Based Pricing?
In a usage-based pricing model the customer is charged based upon their actual consumption of a cloud or SaaS service. If the customer uses the product for one hour they are only billed for a single hour. If the customer doesn’t use the product at all, they will not have to pay any fees. Usage-based pricing is an increasingly popular alternative to the subscription pricing model. With the subscription model the customer is charged a fixed fee per month regardless of whether the customer uses the product each day of the month or not at all.
The charges a customer pays each month in a usage-based pricing model are determined from a simple formula:
Charge = per unit price x quantity of units consumed
Why Usage-Based Pricing?
Usage-based pricing is not a new concept. It has been a popular model throughout the history of computing from the early days of timesharing mainframes to the present day. However, it is currently enjoying a renaissance in the current era of cloud computing and software-as-a-service (SaaS) as it has been one of the key success factors contributing to hyper growth business models from organizations such as Twilio, DataDog, and Snowflake. These organizations have been able to achieve extraordinarily high levels of revenue growth and net revenue retention that have captured the attention of Wall Street analysts and Silicon Valley investors.
While venture capitalists and private equity firms are attracted to the economics of usage-based pricing, the management teams at operating companies like it because of the flexibility and creativity it enables with monetization models. With usage-based models, sales reps can configure contracts, pricing, packaging, and payment terms to best match the budget, cash flow, and risk dynamics of any specific customers. Product managers can select from an almost infinite number of value metrics to perfectly balance the value the customer receives with the price that is paid.
However, usage-based pricing is not the right model for every SaaS and cloud provider. Business, finance, and technology leaders should evaluate the pros and cons of consumption pricing against traditional subscription business models.
The Usage-Based Pricing Guide
Learn about Discounts, Contracts, Billing, Financial Reporting and More
To assist all the employees in SaaS and cloud companies who are trying to understand how usage based pricing works, we developed this guide. It’s written in plain English and offers examples and illustrations to understand each of the key concepts from packaging and presentation to discounting and contracts to billing and investor reporting.
The Unit of Measure for Billing
The crux of any usage-based pricing model is the unit of metric or value metric for the per unit charge. In traditional subscription models, the value metric is typically the number of registered users, but with usage-based pricing SaaS and cloud companies can select from an infinite number of metrics. We group them into four categories to make them easier to understand:
1)Time – The first common metric is charging based upon the length of time the service is used. Think hours, minutes, or even seconds. Many cloud infrastructure companies like AWS, GCP, and Azure price their core compute services based upon duration of use.
2)Transactions – A second common consumption metric is charging based upon the number of transactions processed. Think number of API requests, number of emails sent, or number of user sessions. Each transaction is tracked throughout the billing period and final count is calculated at month-end.
3)Volume – A third set of metrics are volume-oriented. Think megs, gigs, or terabytes of data processed. Cloud infrastructure providers like AWS, GCP, and Azure price storage and networking services based upon the volume of data stored or transmitted.
4)Count – The fourth, and final, set of metrics are count-based. The count could be the number of fleet vehicles measured by a telematics offering or the number of endpoints protected by a cybersecurity service. The count could go up and down throughout the month. Some providers charge based upon the high count while others may bill based upon the end count.
Rates might vary based upon the date/time of use (business vs off-hours), service level (2-hour vs 24-hour turnaround), or even the outcome (success or failure).
Incentivizing Customers to Use More
With usage-based pricing SaaS and cloud providers can generate higher revenues by driving greater consumption. As a result, technology vendors will typically create discount schedules that offer customers a financial incentive to use more of the product. There are four popular discounting models that are most common:
1) Single, Fixed Rate – In the first model, the discount level is fixed independent of volume. For example, the customer might negotiate a 20% discount on an annual contract. The 20% discount is applied whether 1 unit is consumed or 1 trillion units are consumed.
In the other three models the level of discount the customer receives increases in bracketed ranges as they exceed certain usage thresholds.
2) Tiered Model – There is a discount of x% for the first n units, then a higher discount of y% for the next n units, etc. The customer receives 10% discount ($0.90/unit) on the first 100 units. For the next 100 units, the customer receives a 20% discount ($0.80/unit). For the next 100 a 30% discount ($0.70/unit) and so on.
3) Volume Model – A single discount is applied for all units, determined by the total volume used in the period. The customer receives a 10% discount ($0.90/unit) if consuming less than 100 units, a 20% discount ($0.80/unit) for all units if consumption falls between 101-200, a 30% discount ($0.70/unit) for all units if consumption falls between 201-300 and so on.
In the fourth model, the discount is not expressed on a per unit basis. Instead a discounted price is applied for all usage within a certain bracketed range.
4) Stair Step Model – Prices increase in steps as usage grows from one bracketed range to the next. The customer pays $150 per month if their usage is between 0-100 units, $300 per month if their usage is between 101-200 units, and so on.
Prepaid, Monthly Minimums, Spend Commits
As consumption grows, most customers prefer to have the cost savings and predictable expenses that come with a long-term contract even if means losing the flexibility to cancel at any time. As a result, most customers with a meaningful amount of usage will prefer to switch to an annual contract model. The four most popular contract models for usage-based pricing are:
1)Prepaid Usage – The customer purchases a fixed number of units purchased in exchange for discount. For example, the customer purchases 1000 units to use over a 12-month period at a discount of 25% off list prices. The customer draws down against the prepaid balance each month until all of the units are consumed or the prepaid units expire.
2) Monthly Minimums – The customer agrees to pay for a certain minimum quantity of usage per month in exchange for a discount. If the usage exceeds the minimum amount, the customer pays additional usage fees on a per unit basis. For example, the customer might agree to a minimum of 1,000 units per month in exchange for a 20% discount and a per unit of rate of $0.10 for each unit above 1,000.
3) Spend Commitments – The customer commits to spend a creating dollar amount per year (or per quarter) in exchange for a discount. If the actual spend does not exceed the commitment then the customer will pay a “true up” at the end of the period to make up the difference. For example, the customer might agree to spend $1M over 12 months in exchange for a 30% discount off all services.
4) Tiered Subscriptions – The customer purchases a usage-centric subscription package that entitles them to consume a certain amount of the product each month for a fixed fee. For example, the customer might choose from the pro package which includes up to 10,000 units per month for $1K, the business package which includes up to 50,000 units per month for $5K or the enterprise package which includes up to 250,000 units per month for $10K.
Bundles of Entitlements, Features, and Allowances
Many SaaS and cloud providers bundle their features, entitlements, and usage allowances into three tiers, each with a fixed price point. These packages make it easier for customers to select a bundle of services that is most appropriate for their use cases. There are three popular strategies for packaging services with a usage-based pricing model:
1) Feature-defined tiers – Different levels of functionality are offered at increasing per-unit price points. For example, you might offer three tiers of service – good, better, and best. A customer might pay $0.10 per unit for the good feature set, $0.20 per unit for the better feature set, or $0.30 for the best feature set.
2) Usage-defined tiers – Different packages with increasing value are offered at increasing price points, but the key differentiator between the tiers is the volume consumed. For example, you might have three tiers of service – big, bigger, and biggest. A customer might pay $100 per month for consumption up to 5000 units, $200 per month for consumption between 5000 and 10,000 units, $300 per month for 10,000 to 20,000 units, etc.
3) Technology-defined tiers – Packaging is based upon tech specs. Pre-configured bundles are designed to meet certain performance goals and offered at a range of increasing price points. For example, you might have a set of six “t-shirt sizes” – XS, S, M, L, XL, XXL. A customer might pay $0.10 per hour for the S and $0.60 per hour for XXL.
Grids, Tables, Sliding Scales
Usage-based pricing can be complex and intimidating to buyers. If the customer cannot understand how they will be billed for consuming the service, they might delay a purchase or select an alternative vendor with a simpler pricing model. As a result, SaaS and cloud providers invest time and energy to ensure they present their pricing in a format that is easy to understand. Three popular presentation formats include:
1) Grid format – Popular with companies that offer a variety of different products. The user can view each of the company’s offerings in a grid that displays the name of the product, the value metric used for billng, and the list price per unit. For example, a generative AI service might display the cost-per-word for marketers to generate a blog post, a Google Ad headline, Amazon product description, FAQ list, press release or Facebook post in a 3 x 2 grid.
2) List format – The most popular presentation model. Lists can be a simple two-column rate card format with the name of the product and the unit price. Alternatively, a list might have multiple columns providing additional details or parameters about the product features. For example, a cloud compute product might have a product name as well as details about the CPU, memory, and disk
3) Sliding scale – Is a user-friendly format that enables customers to quickly see exactly how much they would pay per month for a given consumption quantity. The user can move a slider to the left or right to reflect the forecasted usage volume and the pricing will be dynamically updated on the web page. For example, if pricing for a marketing service was based upon the number of emails transmitted, the slider would enable customers to quickly estimate costs ranging from 1 to 1M+ messages per month.
High Water Mark, Percentile Billing
The amount that the customer consumed during any given month is not necessarily the amount of usage that the customer will be billed for. A customer may have consumed 10,00 units, but only be billed for 8,000. SaaS and cloud providers use a variety of strategies for calculating the “billable usage quantity.” Some approaches are designed to more fairly bill the customer and others are intended to boost revenues. Two of the more popular models include:
1) High Water Mark – Consumption is metered throughout the month and the charges are based upon the maximum usage in the billing period. For example, suppose you offered an cybersecurity service that billed based upon the number of endpoints connected and managed. On the first day of the month you had 200 devices connected. On the last day of the month you had 210 devices connected. But on day 10, you had 225 devices connected. The high water mark would be 225 and that would be the “billable usage quantity” used for invoicing.
2) Percentile Billing – Sometimes referred to as burstable billing is designed to remove spikes or anomolies in consumption during the billing period. For example, suppose you offered a networking service that charges based upon the of terabytes of data passing through it. However, you don’t want to penalize the customers for occasional bursts in traffic so you decide to bill based upon the 95th percentile. During the month 100 measurements are taken and the top 5 peak measurements would be removed from the usage data submitted for billing.
Metering, Rating, Invoice Generation
The billing process can be the most complex aspect of a usage-based pricing model, particularly when prepaid units, monthly minimums, spend commitments contracts are involved or non-billable usage categories such as free credits or rollover units can apply. Determining the appropriate price and billable usage quantity often require multiple steps and mathematical operations. There are four key steps to the usage-based billing process:
1) Consumption Metering – Native app telemetry is used to collect how much of each service is being consumed by each customer. Depending upon the value metric, the metering might require continuous monitoring of events throughout the billing period, or it might be as simple as measuring the total consumption at the end of the month.
2)Data Mediation – The raw, metered data must be converted into the input format required by the billing system. The consumption data must be sorted by product and mapped to each individual customer. Duplicate records will need to be consolidated and non-billable events such as a failed API call, will need to be eliminated.
3) Usage Rating – The consumption data for each customer is analyzed in the context of the account’s specific discount schedule (volume and tiered discounts) to determine the per-unit price. If the customer has a monthly minimum commitment, the billing system will need to determine if the actual consumption was above or below the threshold. If above the commitment, there may be a need to calculate overage fees. If the customer has prepaid usage, the amount consumed during the billing period will need to be deducted from the prior month’s balance. If the prepaid units have been depleted additional charges to replenish the balance may be required.
4) Invoice Presentation and Delivery – The usage-based charges calculated by the rating engine will need to be combined with other recurring fees, one-time charges, and sales taxes to arrive at the total amount due. The invoice will be rendered into a PDF human-readable format and then routed to the customer via email, a supplier portal, or an EDI transmission to the accounts payable application.
ARR, RPO, Net Retention
Investors want to compare the financial performance of companies with usage-based pricing to the broader community of SaaS and cloud vendors. As a result, technology providers with usage-based pricing are asked to report on many of the same recurring revenue metrics that companies with traditional subscription pricing models report on. Four of the most common investor metrics tracked by companies with usage-based pricing models are:
1) Annualized Recurring Revenue (ARR) – Calculated differently for usage-based pricing than for traditional subscription models. For consumption-based pricing, ARR is calculated based upon the annualized run rate of trailing GAAP revenues. For example, a cloud or SaaS provider might use last month’s recognized revenue x 12 or last quarter’s revenue x 4.
2) Remaining Performance Obligations (RPO) – Large customers will sign annual or multi-year contracts with monthly minimum commitments or prepaid usage credits. Investors want to understand what the dollar value of the backlog is for these contracts. In other words – how much revenue has been contractually committed to, but not yet recognized because performance obligations have not been satisfied?
3) Net Revenue Retention (NRR) – Investors want to understand how cohorts of customers are growing over time. Net retention is one of the most popular ways to measure true growth in a recurring revenue model. Some of the most successful usage-based pricing companies have boasted net retention figures in the 130s and 140s, significantly higher than traditional subscription business models.
4) Large Customer Expansion – As SaaS and cloud providers grow to $50M, $100M and beyond, the ability to land-and-expand enterprise accounts becomes a key focal point for investors. Reporting on upsell and cross-sell activities to current accounts is as important, if not more important, than new customer acquisition. Investors want to see how many large customers are using 2, 3, or 4 different product offerings and how many accounts are generating $100K or $1M in annual revenues.