Most retention reporting produces two numbers, gross and net, with no visible path from one month's revenue to the next. When the numbers are questioned, nobody can reconstruct them.
A waterfall fixes that by forcing the arithmetic into the open.
The layout
At ORM we call it Gross and Net Retention, computed monthly, with these components:
- Beginning ARR - Churned customer ARR (contraction) - Churned product ARR (contraction) - Product decrease ARR (contraction) - New customer ARR (expansion) - New product ARR (expansion) - Increased product ARR (expansion) - Ending ARR
Beginning ARR equals ending ARR from the prior month. Gross revenue retention and net revenue retention both sit on the same chart, derived from the components rather than reported alongside them.
This reconciling waterfall by month is magic, and the reason is the constraint in that one line: because each month opens where the last one closed, the chart has to balance. Any movement you failed to categorize shows up immediately as a chart that does not tie.
Why three contraction lines instead of one
Lumping contraction into a single churn figure destroys the diagnosis, because the three components have different causes and different owners.
Churned customer ARR is a customer leaving entirely. That is a relationship and value-delivery failure. Churned product ARR is a customer staying while dropping a product. That is a product-specific failure, and it is invisible if you only track logo churn, because the customer is still there. Product decrease ARR is a customer keeping the product at a lower quantity or tier. That is usually a usage, budget or pricing signal rather than a satisfaction one.A team reading a single churn number cannot tell those apart, and they call for entirely different responses. The same logic applies on the expansion side, where new customer, new product and increased product represent three distinct motions owned by different teams.
| Component | Type | Usually owned by |
|---|---|---|
| Churned customer ARR | Contraction | Customer success and product |
| Churned product ARR | Contraction | Product and adoption |
| Product decrease ARR | Contraction | Account management and pricing |
| New customer ARR | Expansion | New business sales |
| New product ARR | Expansion | Cross-sell |
| Increased product ARR | Expansion | Account management |
Monthly, not quarterly
Monthly resolution matters more than it looks. Quarterly retention reporting hides the timing of movements inside the quarter and makes it much harder to connect a contraction to whatever caused it.
It also interacts with renewal timing. Renewals cluster around renewal dates rather than distributing evenly, so a quarterly view can show a healthy quarter that contained a very bad month. Renewals behave differently from new business in general, which is why they should be forecast differently, covered in forecasting renewals versus new business.
Reading retention off the chart
With the components in place, both retention rates are direct calculations rather than separately maintained figures.
Gross revenue retention takes beginning ARR less all three contraction components, over beginning ARR. It cannot exceed 100 percent, and it measures how much of what you had you kept.
Net revenue retention adds the expansion components back in. It can exceed 100 percent, and it measures whether the existing base grows on its own.
The value of deriving them from a reconciling chart rather than computing them separately is that a mistake becomes visible. Two independently maintained retention numbers can both be wrong indefinitely. A waterfall that does not tie announces the problem the month it appears.
Building one
Start with a single month and get it to reconcile before extending the series. The first attempt almost never balances, and the difference is the interesting part: it is usually a movement type nobody had a category for, such as a customer who downgraded and expanded in the same month, or a contract restructure that looked like churn plus new business.
Deciding how to classify those edge cases consistently is most of the work, and it is worth doing once, carefully, because consistency is what makes the resulting series usable. For the underlying definitions see net revenue retention and gross revenue retention.
Frequently Asked Questions
What is an ARR waterfall?
A monthly reconciliation that starts with beginning ARR, itemizes every contraction and expansion component, and ends with ending ARR, where each month's ending ARR becomes the next month's beginning ARR. Gross and net revenue retention are calculated from the same chart.
What components belong in the waterfall?
Beginning ARR, then contraction as churned customer ARR, churned product ARR and product decrease ARR, then expansion as new customer ARR, new product ARR and increased product ARR, ending with ending ARR.
Why does the reconciliation matter?
Because beginning ARR equals the prior month's ending ARR, the chart must balance. If it does not, a movement has been missed or double counted, and the retention rates calculated from it are wrong in a way no summary metric would reveal.
Why split contraction into three components?
Because they have different causes and different owners. A customer leaving entirely is a relationship failure, a customer dropping one product is a product failure invisible in logo churn, and a customer reducing quantity is usually a usage or budget signal.
Why monthly rather than quarterly?
Because quarterly reporting hides the timing of movements inside the quarter and makes it much harder to connect a contraction to its cause. Renewals also cluster around renewal dates, so a healthy quarter can contain a very bad month.
See how ORM turns these insights into action
ORM builds custom revenue forecast models for B2B SaaS companies. Not dashboards. Prescriptive analytics that tell you what to do next.
Schedule a Demo