The argument usually starts in a quarterly business review. RevOps presents a slide saying the company leaked a large share of its revenue last year. The CFO asks how that figure reconciles to the general ledger, and nobody can answer, because the number contains customers who churned, deals that slipped, a sales team that spends most of its week on admin, and an invoice that went out without the contracted price uplift. Some of those are revenue. One of them is cost. Two of them may be the same dollar counted twice. Without a shared revenue leakage definition, the leakage figure turns into a debate about vocabulary instead of a decision about what to fix.
If churn sits inside the leakage number, the company may buy a billing recovery tool for a problem that lives in customer success. If inefficiency sits inside it, the figure swells with hours that never touched a customer, and finance discounts the whole slide. Our guide to the seven places revenue disappears maps the lifecycle; this guide gives you three working definitions, a test that sorts any loss into one of them, and the rules that keep them from overlapping.
How we evaluated: what makes a definition finance will accept
VANDFORT publishes this guide, and we sell a diagnostic that puts a dollar figure on revenue leaks, so we have an interest in leakage being taken seriously. That is exactly why we drew the definitions narrowly: a leak number that finance cannot audit helps nobody, including us. We take no referral fees from any vendor or publisher named here.
The published definitions show how wide the gap is. MGI Research, in a December 2025 research note, defines revenue leakage as "the variance between contractually obligated revenue and actual recognized revenue." Clari's 2024 Revenue Leak Report defines it as "revenue that is lost to breaks in your revenue process," and KPMG's 2025 study of 286 technology, media and telecom leaders includes missed cross-sell, onboarding delays and ineffective renewals. Both views are defensible. They simply cannot share one number, which is why the same Clari survey found RevOps leaders estimating 26 percent of revenue lost while CROs estimated 16 percent. These are self-reported estimates from large companies in a vendor-sponsored survey, not measured SaaS losses.
We judged each term against five criteria. A definition that fails any of them will not survive a CFO's review.
| Criterion | The question it answers | Why finance cares |
|---|---|---|
| Traceable | Can you point at the record where the loss happened? | Without a record, the number is an opinion |
| Mutually exclusive | Does each dollar land in exactly one bucket? | Overlapping buckets double count and inflate the total |
| Right unit | Is it measured in revenue dollars or in cost dollars and hours? | Revenue and cost sit on different lines of the income statement |
| Reconcilable | Does it tie to a number finance already reports, such as GRR or billings? | A figure that ties out gets used; one that does not gets discounted |
| Owned | Does it name the team that fixes it? | A loss with no owner stays a loss |
The options compared: leakage, inefficiency, churn and normal loss
The three terms are not competing names for one problem. They measure different things in different units, and a fourth category, normal loss, catches what none of them should hold. The table is the one-page glossary we suggest finance and RevOps sign together.
| Term | What it measures | Unit | The test | Usual owner | Where it reconciles |
|---|---|---|---|---|---|
| Revenue leakage | Revenue due or within reach, lost to a process break | Revenue dollars | Would a working process have kept this dollar without the buyer deciding differently? | RevOps with finance | Bookings to billings; pipeline and renewal records |
| Inefficiency | Cost or time spent per revenue dollar beyond what the process needs | Cost dollars, hours | Did the revenue arrive, but cost more than it should have? | Function leaders, RevOps | Operating expense, CAC, time studies |
| Churn | Recurring revenue that ended or shrank, for any reason | ARR or MRR, percent | Did a customer's recurring revenue stop or contract? | Customer success, finance | GRR and NRR |
| Normal loss | Revenue lost to a buyer's informed decision | Revenue dollars | Did the buyer choose a competitor, no budget or no purchase, with full information? | Sales leadership | Win/loss analysis |
Revenue leakage: two tiers, reported separately
The cleanest way to end the narrow-versus-wide argument is to stop forcing one figure. Split leakage into two tiers. Contractual leakage follows MGI's definition: money a signed contract says you are owed that never reached an invoice or the bank, such as an uncharged price escalator, unbilled usage or a ramp that never stepped up. It ties to bookings and billings, so finance can audit it line by line. Process leakage is the wider RevOps view: demand or renewals lost because a step failed, such as an inbound lead nobody answered or a renewal with no owner activity. It cannot tie to the ledger, because the revenue never existed there, so it is presented as an estimate with its method shown. Our leak ledger framework covers how to price each process leak in a way that survives questions.
Keeping the tiers apart is what earns trust. A CFO will usually accept a contractual figure at face value and a process figure as a range, but not a blend of the two.
Inefficiency: a cost, not a lost dollar
Inefficiency is real and often large. Salesforce's State of Sales research, based on 7,775 sales professionals surveyed in 2022, found that reps spend just 28 percent of their week actually selling. That gap costs money. But it is an expense problem: the company paid for hours that produced too little, and the revenue that did close still arrived.
The common error is converting those hours into "lost revenue" by multiplying them by quota or average deal size. The arithmetic assumes every freed hour would have produced a closed deal, which no finance team will accept, and it is the fastest way to get an entire leak analysis set aside. The rule we suggest: count inefficiency in cost dollars, and count its consequence as leakage only when you can trace it to a record. The admin hours are inefficiency. The lead that waited two days because of them is leakage.
Churn: an outcome, of which some is leakage
Churn is the most measured of the three and the most often mislabeled. SaaS Capital's 2025 survey of more than 1,000 private B2B SaaS companies reported a median gross revenue retention of 91 percent and a median net revenue retention of 101 percent. Its definitions are a good shared standard: NRR is this year's recurring revenue from last year's customers, divided by last year's recurring revenue, including upsells, cross-sells and price increases; GRR uses the same formula without them and cannot exceed 100 percent.
Churn measures what left, not why. It includes deliberate choices that no process could have changed, such as an acquisition, a budget cut or a product that no longer fits. It also includes losses that a working process would have prevented. Involuntary churn is the clearest case: Recurly's analysis of about 25 million recurring transactions, using 2015 data, found roughly one-fifth of churn in its sample was involuntary, as reported by MarketingCharts. That secondary account and old sample are not a current SaaS benchmark; the category still matters. KPMG's finding that 27 percent of TMT leaders cited churn among the top areas of revenue or margin leakage does not establish that all churn is preventable. Only the preventable part belongs there. We cover how to catch that part early in our post on building an NRR early-warning system and our renewal leakage guide.
Where the three overlap, and how to stop double counting
Overlap is where most leak numbers break. Three rules keep each dollar in one place. First, treat churn as a metric and leakage as a classification: report churn in full through GRR, then tag the preventable portion as "of which leakage" rather than adding it again. Second, inefficiency never enters a revenue figure; it travels on its own line, in cost dollars. Third, classify each lost dollar at the record where it was lost, once. A renewal that lapsed because nobody worked it is churn in the retention report and process leakage in the leak report, but it is the same dollar and is counted once in any total.
Expansion needs one extra word of care. An account that outgrew its plan and was never offered the next tier did not churn and was not billed incorrectly. If the usage data showed the need and no one acted, it is process leakage. If no signal existed, it is simply growth you did not capture, and it belongs in planning rather than in the leak figure.
How to choose: a five-step process for one agreed number
The steps below are a suggested starting point, not a benchmark. They are built so that RevOps does the sorting and finance can check every line.
Sign the glossary first. Put the four definitions from the table on one page and have the CFO and the RevOps lead agree to them before any number is calculated. Arguments about definitions are cheap before the data is pulled and expensive afterward.
List last quarter's losses, one row per record. Each row is a deal, invoice, subscription or renewal, with its dollar value and the system it lives in. If a loss cannot be tied to a record, it goes on a separate list of hypotheses.
Run every row through the test in order. Did the revenue arrive? If yes, the only question is cost, so it is inefficiency or nothing. Did an informed buyer decide? If yes, it is normal loss or voluntary churn. Was the money contractually due? If yes, it is contractual leakage. Did a traceable process break cause it? If yes, it is process leakage. Anything left is unclassified, not leakage.
Reconcile before you present. Churn should tie to the GRR finance already reports. Contractual leakage should tie to a bookings-to-billings comparison. Process leakage goes in as an estimate with its method stated next to it.
Review the ambiguous rows together. Once a month, finance and RevOps look at the rows the test could not settle and update the glossary with the decision.
Once the terms are fixed, mapping each leak type to the system that closes it is the logic behind our Nine Systems Framework. Preventable churn points toward the Churn Signal Watchtower and Renewal Radar; stalled pipeline toward the Pipeline Hygiene Sentinel. Presenting the agreed figure to a board is its own discipline, covered in our guide to board-ready revenue reporting.
Where VANDFORT fits, and where it doesn't
You run a B2B SaaS company between roughly $3M and $30M in ARR, and finance and RevOps already disagree about how much is leaking. You want a quick read on which of the four domains is weakest before anyone builds a spreadsheet, which is what the free GTM Health Score gives you. If you then need each leak priced from your own records, using definitions finance has signed, the Revenue Leak Report is a three-week, read-only diagnostic. Any build after that ships one of the nine systems at a time, tested on around 20 of your own past cases and released only if it reaches 85 percent.
If the open question is accounting policy, such as how a contract modification is treated under ASC 606, your controller and auditor should decide it, not a revenue operations firm. If your leakage is concentrated in invoicing for complex usage-based pricing, a billing or revenue recognition specialist fits better, because we do not run billing. If all you need is a shared glossary, the table on this page and a half-day working session between finance and RevOps will get you there without outside help. And if you have little CRM or billing history, there are not yet enough records to sort.
Frequently asked questions
What is the definition of revenue leakage?
Revenue leakage is revenue a company was contractually owed, or could reasonably have won from a ready buyer, that it lost because a process or system failed. The finance-grade version is narrower: the gap between what contracts say you should bill and what you actually invoice and collect. The test in both cases is that a fixable break, not a buyer's informed decision, caused the loss.
Is churn the same as revenue leakage?
No. Churn measures all recurring revenue that ended or shrank, for any reason, including deliberate decisions by customers to leave. Revenue leakage covers only losses caused by a preventable failure. Some churn is leakage, such as a renewal nobody worked or a failed payment nobody recovered, but most churn analyses should report the full figure through gross revenue retention and tag the preventable part separately.
Is involuntary churn considered revenue leakage?
Usually yes, when the loss was recoverable. Involuntary churn happens when a customer who intended to stay is lost to a failed card, an expired payment method or dunning emails sent to the wrong person. Because no buyer decided to leave, the loss passes the leakage test. Count it once: as churn in your retention metrics, and as the "of which leakage" portion in the leak report.
What is the difference between revenue leakage and inefficiency?
Revenue leakage is lost revenue; inefficiency is excess cost. If the deal closed but took twice the hours it should have, that is inefficiency, measured in cost dollars or time. If the deal was lost because a step failed, that is leakage, measured in revenue. Converting wasted hours into lost revenue by multiplying by quota inflates the figure and rarely survives a finance review.
How do you get finance and RevOps to agree on a revenue leakage number?
Agree on the definitions before calculating anything. Sign a one-page glossary that separates leakage, inefficiency, churn and normal loss, then sort last quarter's losses record by record using one test. Reconcile churn to reported GRR and contractual leakage to billings, and present process leakage as an estimate with its method shown. Review disputed rows together each month.
Sources: KPMG, RevOps Redefined: A growth playbook for TMT (n = 286, fielded late 2024), 2025. Clari and Vanson Bourne, Revenue Leak Report (n = 420), 2024. SaaS Capital, 2025 B2B SaaS Retention Benchmarks. MGI Research, Revenue Leakage Series Part 2, December 2025. Salesforce, State of Sales (n = 7,775), 2022. Recurly churn research (2015 data), via MarketingCharts, 2016. Sources are linked beside their claims; accessed October 2026.
Last reviewed: October 2026. Vendor features and pricing change; check each vendor's site before buying.
