If you sell metered APIs, cloud infrastructure, or any product with usage-based pricing, the standard LTV formula can be dangerously misleading. It assumes a stable average revenue per account and treats churn as the main leakage. But in a metered billing world, revenue often expands and contracts month to month, and contraction risk—not just churn—can destroy lifetime value long before a customer cancels. In this article, you’ll learn how to calculate LTV for metered billing with contraction risk in a way that reflects the real economics of usage-based software.
Why the Standard LTV Formula Breaks Under Metered Billing
The textbook LTV formula is simple:
LTV = (ARPA × Gross Margin) / Churn Rate
That formula works reasonably well for flat-fee subscriptions. Revenue is predictable, churn is the main reason a customer stops generating value, and ARPA is a representative number for the customer base. But usage-based pricing changes the dynamics. Customers pay for what they consume, so contract value is not fixed at signup. A customer who appears healthy on an annual contract can quietly halve its usage over three months. Standard LTV calculations do not capture that behavior until it is too late.
There are four reasons the standard formula fails in a metered billing environment:
- Revenue is volatile. Usage can spike during the onboarding phase and then fall as customers optimize their workflows or move only a small percentage of their workload.
- Contraction is not the same as churn. A customer can remain active, renew, and still become dramatically less valuable to you over time.
- Cohort averages are misleading. A high-usage customer and a shrinking customer can produce the same average revenue, but their future value is completely different.
- Expansion and contraction compound. Usage-based startups often report high net revenue retention, but that aggregate hides cohorts where expansion offsets severe contraction.
The Hidden Variable: Contraction Risk
Contraction risk is the probability that a customer will use less of your product over time. It comes from many places: seasonal demand, engineering teams scaling back after the initial integration, budget pressure, pricing model changes, or simply a new technical leader who standardizes on a different internal tool.
When you ignore contraction risk, very different customer segments can appear to have the same LTV. Consider two customers:
- Customer A consumes 100 units each month consistently.
- Customer B consumes 150 units in the first quarter, then drops to 30 units by month six.
Both may have similar first-year ARPA, but Customer B’s second-year revenue is a fraction of Customer A’s. The standard LTV formula treats them as identical. A metered billing LTV model separates the two because it treats the revenue stream as a probability distribution of future usage, not a flat line.
A Better LTV Calculation for Usage-Based Pricing
To move beyond the standard model, you need an LTV calculation that accounts for variable revenue, contraction probability, and expansion potential. The goal is no longer “average customer lifetime.” It is the expected net present value of a customer’s future usage revenue, adjusted for how that usage is likely to change over time.
Here is a practical framework that starts simple and layers in risk assumptions as you validate them.
Step 1: Model Expected Usage per Customer
Start with a baseline usage projection, not an average invoice. Segment your customers by product, onboarding size, and measurable usage patterns. For each segment, calculate the expected usage revenue for each month or quarter after signup. This can be as simple as a forward-looking curve derived from cohort data.
If you are a cloud provider, you might see that customers start with high, spiky usage in month one, decline sharply in month two, and then gradually ramp as they move more workloads into production. That shape is your expected usage baseline. Without it, any LTV calculation is just an extrapolation of an average that may not exist in any actual customer.
Step 2: Estimate Contraction Probability and Magnitude
Next, estimate the probability of a contraction event and the size of that contraction. For usage-based pricing, this means asking questions like: what percentage of customers reduce their usage by more than 20% in a given month? What is the average magnitude of that reduction? What does the usage revenue curve look like after the reduction?
Use historical usage data to find the answer. A useful metric is usage-based gross revenue retention, or usage GRR. It tracks the sum of customer usage revenue from a cohort at the start of a period compared with the same cohort later, excluding any new expansion revenue. If a cohort starts at $1,000,000 in monthly usage revenue and drops to $700,000 six months later without any customers canceling, contraction risk has quietly eaten 30% of that cohort’s value.
Step 3: Discount Cash Flows and Account for Expansion
Once you have usage projections, contraction probabilities, and expansion assumptions, you can write a more honest LTV formula:
LTV = Σ (Expected Usage Revenue_t × Gross Margin × Survival Probability_t × (1 - Contraction_t)) / (1 + Discount Rate)^t
Where Survival Probability_t is the probability that the customer remains active through period t, and Contraction_t is the expected percentage reduction in usage revenue at period t. If you have expansion, include an expansion factor for the segment. This formula does not pretend revenue is constant. It values each future cash flow according to how likely it is to exist and how large it is likely to be.
What the New Formula Looks Like in Practice
Consider a customer segment with an initial monthly usage revenue of $1,000. The gross margin is 80%, and the monthly churn rate is 4%. Under the standard formula, LTV would be:
LTV = ($1,000 × 0.80) / 0.04 = $20,000
But that calculation assumes revenue will stay at $1,000 per month for as long as the customer remains active. Now suppose historical data shows that usage declines by 5% per month. In month one, expected usage revenue is $1,000 × 0.95 = $950. With a 4% monthly churn rate, the probability the customer still exists and generates that revenue is 0.96. After applying an 80% gross margin and a 1% monthly discount rate, the expected contribution from month one is roughly:
$950 × 0.80 × 0.96 ÷ 1.01 ≈ $722
By month 12, expected usage revenue has been compounded down by both contraction and churn. The resulting LTV is substantially lower than $20,000. The difference between the standard number and the contraction-adjusted number is the true cost of contraction risk.
This is not just a theoretical exercise. For a usage-based startup, misstating LTV can lead to overly aggressive acquisition spend, poor segmentation, and pricing decisions that look profitable but are not. A customer acquired because it first consumes a lot may become unprofitable once usage normalizes to a lower sustainable level.
How to Operationalize This LTV Model in Your Metrics Stack
You do not need to wait until your finance team hands you a perfect model. Start with a spreadsheet and a cohort usage table, then move into a metrics warehouse as you refine the assumptions. The key data points you need are:
- Monthly usage revenue by customer, not just monthly recurring revenue.
- Cohort-based gross retention curves that separate churn from contraction.
- Expansion events such as new seats, new features, or natural usage growth.
- Invoice-level revenue, because metered pricing can include minimums, credits, discounts, and overage tiers.
From these inputs, build a simple forecast of expected usage revenue over the next 12 to 24 months for each active cohort. Then discount the cash flows and aggregate. It is not enough to calculate LTV once at acquisition. Recalculate it monthly as usage data accumulates, and use the new number to guide customer health monitoring, expansion plays, and product roadmap decisions.
Usage-based pricing rewards companies that deeply understand the movement of customer value. The standard LTV formula was built for a simpler world where revenue was fixed and churn was the only leak. In metered billing, contraction risk deserves equal weight. By modeling expected usage, contraction, and survival probability together, you can turn LTV from an optimistic historical artifact into a forward-looking measure that helps you prioritize product investments, pricing changes, and customer health interventions.
