SaaS metrics software
Each month's MRR is last month's, plus new and expansion revenue, minus contraction and churn. CX Cash computes the four movements from your billing system, reconciles them to your books every month, and reads retention, CAC payback and the Rule of 40 off the same bars, with every definition written down.
Private beta. Free during the beta. We set up each account by hand, one at a time.
MRR bridge · June
Sample companyRead off the six bars
Why the bridge
ARR appears nowhere in US GAAP, and no SEC rule defines it. Retention is no better anchored: "Since there is no official US-GAAP definition of churn or retention, different companies use different ways to measure and report these metrics," Christoph Janz of Point Nine Capital wrote in 2017 (Janz, 2017). The filings of listed companies show what follows.
| Company and year | Metric | What the filing shows |
|---|---|---|
| Zoom, fiscal 2022 | Net dollar expansion rate | Reported 129% after "greater than 130%" for the two years before. The figures behind the threshold were 164% and 135%, so a 35-point fall read as a slip of a point. |
| Fastly, 2022 | Net retention | Three net retention rates for one customer base in one year: 122.7%, 119.1% and 110.7%. |
| HubSpot, 2025 | Net revenue retention | A change of method restated 2024 from 102.2% to 101.8%, and 2023 from 103.9% to 103.0%. |
| SentinelOne, June 2023 | ARR | A one-time cut of $27M, about 5% of ARR. "The adjustment to ARR did not impact historical revenue or total bookings." |
| CrowdStrike and Commvault, 2026 | ARR | One keeps an expired contract in ARR while renewal talks continue; the other makes no renewal assumption at all. |
Source: the filings and SEC correspondence linked in each row. The Zoom and Fastly cases are set out on CX Cash's net revenue retention page.
Listed companies have a reviewer. The SEC's 2020 guidance on key metrics expects "A clear definition of the metric and how it is calculated" (SEC, 2020). On CX Cash's count from EDGAR's full-text search, 62 staff letters contain the phrase "annual recurring revenue", 17 of them from 2021, and a search on NRR's names finds 193 staff comment letters to 111 companies between 2014 and 2025. A mention is not always a comment, so both counts are upper bounds. The requests keep returning to the movements: reviewing Instructure's draft IPO filing in 2021, the staff asked for "the contribution of each identified factor" behind its revenue growth, in numbers (SEC staff, 2021).
A private company has no staff reviewer. Its board gets the same protection only when the figure arrives with its bridge: four movements that add up to the change, each with a written rule for what it counts.
One cohort, two outcomes. The waterfall below follows a year of one cohort's ARR. Switch from the healthy preset to the leaky one: starting ARR stays at $1.0M while NRR falls from 111% to 87%, because churn rises from $80,000 to $220,000 and expansion falls from $190,000 to $90,000. CX Cash draws this waterfall for every signup cohort.
The figure on the board slide and the figure in diligence come from the same bridge, under the same rules.
+ New MRR → Acquisition
New MRR is the bar a company pays for directly. Customer acquisition cost (CAC) divides what winning customers cost by the number won, and CAC payback divides CAC by the monthly gross profit each new customer brings: $12,000 recovered at $800 a month takes 15 months.
Built from an ad dashboard, CAC counts media spend and leaves out the salaries, commissions and tools that carry most of a sales-led channel's cost. Blended across channels, it hides the channel that loses money. HubSpot found one when it split the ratio: "When we looked at the math we realized we had a LTV:CAC ratio of 1.5 selling direct, and a LTV:CAC ratio of 5 selling through the channel," in the words of HubSpot's Brad Coffey, quoted by David Skok (Skok, 2013). Margin moves the figure as well: a 12-month payback on revenue becomes 14.1 months on gross profit at Atlassian's 84.8% gross margin and 24.5 months at Twilio's 48.9% (our arithmetic on the latest 10-Ks of Atlassian and Twilio).
| Channel | New customers | New MRR | CAC: paid-only / fully loaded | Payback, fully loaded |
|---|---|---|---|---|
| Paid search | 50 | $25,000 | $1,800 $3,000 | 7.5 mopaid-only: 4.5 |
| Outbound sales | 12 | $24,000 | $833 $20,000 | 12.5 mopaid-only: 0.5 |
| Partners | 10 | $9,000 | $0 $4,500 | 6.3 mopaid-only: 0.0 |
| Blended | 72 | $58,000 | $1,389 $6,042 | 9.4 mopaid-only: 2.2 |
Source: hypothetical company. Paid-only CAC counts media and programme spend; fully loaded adds salaries, commissions, tools and allocated overhead. Payback on an 80% gross margin. Our arithmetic.
What CX Cash does
CX Cash takes sales and marketing costs from your accounting system by account and department, new MRR from billing, and each new customer's source channel from HubSpot or Salesforce. For every channel and month it reports paid-only CAC, fully loaded CAC (with the overhead share you set once, at setup), payback in months on gross margin, and LTV:CAC with lifetime capped at the horizon you choose. A channel whose fully loaded payback runs past your threshold is flagged in the month it crosses.
The channel budget follows the fully loaded figure, and the board sees the same one.
+ Expansion → Net revenue retention
Across the latest 10-Ks of 17 listed software companies that state a numeric NRR, NRR alone accounts for about 68% of the variation in revenue growth (our arithmetic). Take net expansion out of each company's growth and the remainder, roughly what new customers added, lies between 2 and 10 points for 15 of the 17, with a median of 7.2.
| Company | NRR | Revenue growth | Net expansion | Rest of growth |
|---|---|---|---|---|
| Snowflake | 125% | 29.2% | 25 | 4.2 |
| MongoDB | 121% | 23% | 21 | 2 |
| Datadog | 120% | 28% | 20 | 8 |
| Cloudflare | 120% | 30% | 20 | 10 |
| GitLab | 118% | 26% | 18 | 8 |
| CrowdStrike | 115% | 22% | 15 | 7 |
| Confluent | 114% | 21% | 14 | 7 |
| Workiva | 112.8% | 19.7% | 12.8 | 6.9 |
| Dynatrace | 110% | 19% | 10 | 9 |
| DoubleVerify | 109% | 14% | 9 | 5 |
| Twilio | 108% | 14% | 8 | 6 |
| Okta | 106% | 12% | 6 | 6 |
| Procore | 106% | 15% | 6 | 9 |
| Vertex | 105% | 12.2% | 5 | 7.2 |
| HubSpot | 103.5% | 19% | 3.5 | 15.5 |
| Yext | 97% | 6% | -3 | 9 |
| Weave | 93% | 17% | -7 | 24 |
Source: each company's latest Form 10-K on SEC EDGAR, for example Snowflake, Datadog, CrowdStrike and Weave. MongoDB's and Datadog's NRR are approximate, Cloudflare's is a fourth-quarter figure and Snowflake's growth is product revenue. A convenience sample that defines NRR four different ways; the split is our arithmetic.
Small companies run the other way. In OpenView's 2023 benchmarks, for companies under $1M of ARR, "86% of new-net ARR growth comes from acquisition and only 14% from expansion" (OpenView, 2023). The bar that carries growth changes as a company scales, and a single growth rate does not say when it changed.
Net revenue retention follows one cohort for a year: (starting ARR + expansion − contraction − churn) ÷ starting ARR. Gross revenue retention (GRR) drops the expansion and caps each customer at its starting revenue, so it cannot pass 100%. SaaS Capital made the case for reading the two together in 2016: "Two very different stories emerge between companies that both have 150% Net Retention, if one has strong Gross Retention of 95% and the other only has 60% Gross Retention" (SaaS Capital, 2016).
What CX Cash does
CX Cash builds a cohort for every signup month and quarter and reports NRR and capped GRR side by side, point-in-time and trailing 12 months, with churned customers kept in at zero, the cohort method of the SaaS Metrics Standards Board. Beside each rate sit the share of expansion that came from the five largest accounts and the NRR without them, so one large upsell cannot pass for a healthy base.
The board sees which cohorts grow on their own, and how much of that growth rests on a handful of accounts.
− Contraction and churn → Retention by segment
Contraction is revenue lost from customers who stay; churn is revenue lost from customers who leave. Counted in customers, the loss is logo churn; counted in dollars, it is revenue churn; and the two part ways whenever the customers who leave are smaller or larger than average.
| Cut | What the source found | Source |
|---|---|---|
| Price point | Median monthly logo churn falls from 6.1% for accounts paying under $25 a month to 2.2% above $500. | ChartMogul, 2022, from its billing data |
| Contract size | Median gross revenue retention is about 91% below $250,000 of annual contract value and 95% above it. | SaaS Capital, 2025, survey of private B2B SaaS companies |
| Customers or dollars | A 10% customer churn rate is 3.4% revenue churn when 9 of the 10 lost accounts are small. | Skok, 2013, worked example |
| Denominator | Stripe adds the period's new subscribers to the denominator; ChartMogul removes customers who joined and churned in the same period. | Stripe and ChartMogul documentation |
Source: as linked in each row. The vendors' figures come from their own customers.
A churn rate pulled from a billing export also mixes plans that could have churned this month with plans that could not. Janz again: "By including customers on annual plans who aren't up for renewal in the period you're measuring you're underestimating your true churn rate." And the moment a lapsed contract becomes churn is a policy. CrowdStrike keeps an expired subscription in ARR "if we are actively in discussion with such organization for a new subscription or renewal" (CrowdStrike, 2026).
What CX Cash does
CX Cash keeps contraction and churn as separate bars and reports logo churn, gross revenue churn and capped GRR for any cut: plan, tier, deal-size band, signup quarter, region or account owner. Annual plans enter the churn denominator only in their renewal month. When a lapsed contract counts as churn (on notice, at contract end, or after a grace period you set) is a written setting, printed under every churn figure.
A leaking segment shows up in the month it starts, in customers and in dollars.
= Ending MRR → Efficiency
Growth alone says nothing about what it cost. The Rule of 40 adds revenue growth to profit margin, the burn multiple divides net burn by net new ARR, the magic number divides net new ARR by sales and marketing spend, and revenue per employee divides ARR by headcount. Each one sets a cost from the books against a line of the bridge.
revenue growth % + profit margin %
Bridge: growth Books: profit margin
The margin was never fixed. Datadog's 2025 10-K scores 28.0 on EBITDA margin and 54.4 on free cash flow margin (Datadog, 2026); across eight listed software companies' latest 10-Ks, none passes on EBITDA and seven pass on free cash flow (our arithmetic).
net burn ÷ net new ARR
Bridge: net new ARR Books: net burn
David Sacks's measure of "how much is the startup burning in order to generate each incremental dollar of ARR?" (Sacks, 2020). Median 1.0 under $5M of ARR (Benchmarkit, 2026).
net new ARR ÷ prior period's sales and marketing
Bridge: net new ARR Books: sales and marketing
Lars Leckie's 2008 line: "if you are below 0.75 then step back and look at your business, if you are above 0.75 then start pouring on the gas for growth" (Leckie, 2008).
ARR ÷ full-time employees
Bridge: ending ARR Books: headcount
Median $141,125 for private SaaS companies in 2026, from a survey of more than 1,000 (SaaS Capital, 2026).
Score it. Drag growth and margin below. A company growing 30% at a −5% margin scores 25; one growing 15% at a 25% margin scores 40. CX Cash charts the two parts separately every month, so the board can see which of them moved the score.
What CX Cash does
CX Cash computes all four monthly on trailing 12 months. The Rule of 40 names its margin (EBITDA, operating or free cash flow) beside the score; the burn multiple takes net burn from your books and net new ARR from the bridge; the magic number uses the net-new-ARR form, with Scale Venture Partners' GAAP-revenue form one switch away; revenue per employee takes headcount from payroll or a sheet.
Growth, retention and efficiency read off one reconciled bridge, so when two ratios disagree, the reason is in the business.
Private beta. Free during the beta.
Under every bar → The books
MRR comes from billing and revenue from the books, and the two should agree. Twelve months of MRR should add up to the recurring revenue recognised over the same months, give or take the timing of starts within each month. Public filings allow a yearly version: recurring revenue divided by the average of opening and closing ARR should come out close to 1.
Source: each company's Form 10-K for fiscal 2026 (CrowdStrike, Samsara, SentinelOne, Zscaler, Dynatrace): subscription revenue for CrowdStrike and Dynatrace, total revenue for the others. The ratios are our arithmetic.
Growth that bunches late in the year explains a ratio a little under 1, and a definition that leaves usage out explains one a little over. A ratio well away from 1 with neither explanation is a question for the CFO before it becomes one for the auditor.
What CX Cash does
Each month CX Cash compares recurring revenue from the bridge with the revenue accounts in QuickBooks Online, Xero or NetSuite and lists every difference by cause: one-off charges, credits and refunds, currency, and mid-month starts. A difference with no cause is flagged for review before the month's pack goes out.
From a bar to the books · sample data
Any figure on any screen opens onto the customers behind it, each customer onto its subscription events and invoices, and each invoice onto the entry in your books. Access is read-only: CX Cash cannot post an entry or change a subscription.
A figure questioned in diligence has its answer a few clicks away, down to the invoice.
The method
The SEC asks listed companies for the definition and the calculation of every key metric. CX Cash gives a private company's board the same, on screen and in the pack, with each definition published on a CX Cash metric page or in the SaaS metrics hub.
| Figure | How CX Cash computes it | Notes |
|---|---|---|
| MRR and its movements | starting MRR + new + expansion − contraction − churned = ending MRR | Annual plans spread by month; one-off charges and taxes left out. |
| ARR | MRR × 12 | Committed revenue and usage reported as two lines. |
| Net revenue retention | (starting ARR + expansion − contraction − churn) ÷ starting ARR | Same cohort 12 months apart; churned customers kept at zero; point-in-time beside trailing. |
| Gross revenue retention | (starting ARR − contraction − churn) ÷ starting ARR | Each customer capped at its starting ARR; read beside NRR, as the net revenue retention page explains. |
| Logo churn | customers lost ÷ customers at start | Customers who joined and left in the period excluded; annual plans counted in their renewal month. |
| CAC payback | CAC ÷ (new MRR per customer × gross margin) | Fully loaded and paid-only, by channel. |
| LTV:CAC | (ARPA × gross margin ÷ revenue churn) ÷ CAC | Lifetime capped at the horizon set at setup. |
| Rule of 40 | revenue growth % + profit margin % | Margin type named beside the score. |
| Burn multiple | net burn ÷ net new ARR | Net burn from the books. See burn rate and the burn multiple. |
| Magic number | net new ARR ÷ prior period's sales and marketing | Annualised quarterly; Scale Venture Partners' GAAP-revenue form available. |
| Revenue per employee | ARR ÷ full-time employees | Headcount at period end, contractors excluded unless you include them. |
Source: formulas as defined on the linked CX Cash pages. The settings shown are defaults; each can be changed at setup, and every change is dated.
Every benchmark in CX Cash shows its publisher, sample and date, the way this page cites them: SaaS Capital's 2025 private medians of 101% NRR and 91% GRR (SaaS Capital, 2025), Benchmarkit's 2026 private median CAC payback of 16 months, from 198 companies (Benchmarkit, 2026), and a public median NRR of 110% on 25 September 2026 (Clouded Judgement, 2026). Your figure sits beside the benchmark for your stage and contract size.
The pack goes to the board and investors as PDF, Excel or a read-only link: the bridge on page one, then retention, acquisition and efficiency, then the definitions. When a definition changes, the pack prints both versions for the overlapping periods, as HubSpot did when it changed its NRR method. It sits alongside the monthly investor update and answers the questions that update raises.
Take the free dashboard today, with every formula written beside its result; request access for the version that fills itself in each month.
Built for
Not the right fit if
Free template
Twelve months of inputs (MRR movements, customers, sales and marketing spend, margins) and a dashboard that computes ending MRR, ARR, monthly and 12-month growth, monthly net and gross MRR retention, CAC, logo churn and the Rule of 40, with each formula written beside its result.
The preview is the file itself: live formulas and worked sample numbers, yours to keep whether or not you use CX Cash.
Private beta · Free during the beta
Leave your work email and you join the queue. We set up each account by hand, one at a time. We email you when your setup slot opens. Free during the beta.
It is in private beta. Request access on this page and you join the queue. We set up each account by hand, one at a time, and email you when your slot opens.
CX Cash is free during the private beta. Pricing for after the beta has not been published yet.
Software that computes a subscription company's operating metrics, such as MRR, ARR, churn, net and gross revenue retention and CAC, from its billing and accounting data, so they are calculated the same way every month instead of rebuilt in a spreadsheet.
A reconciliation of one period's MRR to the next: starting MRR, plus new and expansion MRR, minus contraction and churned MRR, equals ending MRR. Growth, net and gross retention, the quick ratio, the burn multiple and the magic number are all computed from its lines.
Product analytics tools such as Mixpanel or Amplitude track what users do inside an app: events, funnels and feature use. A SaaS metrics platform such as CX Cash tracks the financial side: revenue, retention, churn and acquisition cost. Many SaaS companies run one of each.
The MRR or ARR bridge, net and gross revenue retention, logo churn, CAC payback, gross margin, the burn multiple and runway, and the Rule of 40 once profit means something. Each figure should come with its definition and the period it covers.
Yes. MRR comes from your billing system and recomputes as subscriptions start, change and cancel, with annual plans spread across their months and one-off charges left out. ARR is MRR × 12, with committed revenue and usage reported as two lines.
Yes. Each month CX Cash compares recurring revenue from the bridge with the revenue in your accounting system and lists each difference by cause. Connections are read-only, and the books also supply the costs behind CAC, the burn multiple and the Rule of 40.
Both follow the revenue of last year's customers. Gross revenue retention subtracts contraction and churn and caps each customer at its starting revenue, so it cannot exceed 100%. Net revenue retention also adds expansion, so it can. A wide gap means expansion is covering for losses.
Yes. Any figure opens onto the customers behind it, each customer onto its subscription events and invoices, and each invoice onto the entry in your books.
Company figures come from the filings and SEC correspondence above, and benchmarks from the surveys and vendor data named beside them, with their limits. The counts of SEC staff letters, the 17-company split, the payback and Rule of 40 recalculations and the five-company revenue check are CX Cash's own arithmetic on the filings linked beside them. The sample company in the bridge and the channel table are hypothetical; figures without a source beside them are our arithmetic.
Burn Rate Counts Months, and the Burn Multiple Prices ARR
Bookings vs Billings vs Revenue, Linked by Two Balances
How to Read a SaaS P&L From Revenue to Adjusted EBITDA
