SaaS Metrics Explained: MRR, Churn, CAC, LTV and What to Track First
SaaS metrics explained in plain English: MRR, ARR, churn, net revenue retention, CAC, LTV and payback, with formulas, worked examples and what to track first.
Mythex Team · · 7 min read
SaaS metrics are the handful of numbers that tell you whether a subscription business is healthy: how much recurring revenue you have (MRR), how fast it leaks away (churn), what it costs to win a customer (CAC), and how much a customer is worth over time (LTV). Start by tracking MRR and churn every month; add CAC, LTV and payback once you spend money on acquisition. The formulas are simple. The hard part is defining them consistently and not fooling yourself.
All numbers in this guide are made-up examples to show the arithmetic. They are not benchmarks.
Why these metrics matter
A subscription business earns money slowly. A customer who pays monthly might take many months to repay what you spent to find them. That makes it easy to feel busy while the business shrinks. The metrics below answer four questions:
| Question | Metric |
|---|---|
| How big is the business right now? | MRR, ARR |
| Is it keeping what it has? | Churn, net revenue retention |
| What does growth cost? | CAC, payback period |
| Is a customer worth more than they cost? | LTV, LTV:CAC ratio |
MRR and ARR
Monthly recurring revenue (MRR) is the total subscription revenue you expect every month, normalised to a monthly figure.
- Monthly plans count at their monthly price.
- Annual plans are divided by 12.
- One-off fees, setup charges and refunds are left out.
- Discounts count at the discounted price the customer actually pays.
Annual recurring revenue (ARR) is MRR × 12. It is mostly used once a business sells annual contracts.
Example (made-up numbers): you have 40 customers on a $20/month plan and 5 customers on a $480/year plan.
- Monthly customers: 40 × $20 = $800
- Annual customers: 5 × ($480 ÷ 12) = $200
- MRR = $1,000, ARR = $12,000
MRR movements
Total MRR hides what happened underneath. Break the change each month into parts:
- New MRR — from brand-new customers
- Expansion MRR — existing customers upgrading or adding seats
- Contraction MRR — existing customers downgrading
- Churned MRR — customers who cancelled
Net new MRR = new + expansion − contraction − churned.
Example: new $300, expansion $50, contraction $20, churned $80. Net new MRR = $250. A month where total MRR grows can still hide rising churn, which is why the breakdown matters.
Churn
Churn measures what you lose. There are two kinds, and you want both.
Customer (logo) churn rate = customers lost in the period ÷ customers at the start of the period.
Revenue churn rate = MRR lost to cancellations and downgrades ÷ MRR at the start of the period.
Example: you start the month with 50 customers and $1,000 MRR. Three customers paying $20 each cancel and one downgrades from $40 to $20.
- Customer churn = 3 ÷ 50 = 6%
- Revenue churn = ($60 + $20) ÷ $1,000 = 8%
Two rules keep churn honest:
- Only count customers who were there at the start. Customers who sign up and cancel within the same month distort the rate if you mix them in; track them separately as early churn.
- Decide when a cancellation counts. When they click cancel, or when their paid period ends? Either is fine; be consistent.
Monthly churn compounds. A monthly rate that sounds small removes a large share of customers over a year, so don't multiply by 12 to get annual churn. To estimate how many of today's customers remain after 12 months at a steady monthly churn rate, use (1 − monthly churn)^12. At 6% monthly churn that's 0.94^12 ≈ 0.48 — roughly half left.
For what to do about it, see how to reduce churn.
Net revenue retention (NRR)
NRR asks: of the revenue I had from a group of customers a year (or month) ago, how much do I have from those same customers now, including upgrades?
NRR = (starting MRR + expansion − contraction − churned) ÷ starting MRR
Example: a group of customers paid $1,000 MRR in January. Twelve months later, the ones who stayed pay $900 and upgrades add $150. NRR = $1,050 ÷ $1,000 = 105%. Above 100% means existing customers grow faster than they leave, so you would grow even with no new sign-ups.
Gross revenue retention (GRR) is the same calculation without expansion, and can never exceed 100%. It shows the pure leak.
Customer acquisition cost (CAC)
CAC = total sales and marketing spend in a period ÷ new customers won in that period.
Example: you spend $600 on ads and $400 on a freelance writer in a month and win 20 customers. CAC = $1,000 ÷ 20 = $50.
The main arguments are about what to include: your own time, tools, salaries. There's a full walkthrough, including blended vs paid CAC, in how to calculate customer acquisition cost.
ARPA and gross margin
Two supporting numbers feed the formulas below.
Average revenue per account (ARPA) = MRR ÷ number of paying customers. With $1,000 MRR and 50 customers, ARPA = $20.
Gross margin = (revenue − cost of serving customers) ÷ revenue. Cost of serving includes hosting, databases, paid APIs such as AI models, payment fees and support. If serving costs $250 of every $1,000, gross margin is 75%.
Customer lifetime value (LTV)
LTV estimates how much gross profit a customer brings before they leave.
LTV = ARPA × gross margin ÷ monthly customer churn rate
Example: ARPA $20, gross margin 75%, monthly churn 5%. LTV = $20 × 0.75 ÷ 0.05 = $300.
Be careful with LTV:
- It is sensitive to churn. Halve the churn rate and LTV doubles. A few good months can make LTV look far better than it is.
- With under a year of data, you are extrapolating. Present it as a range.
- Use gross profit, not revenue, or you overstate what a customer is worth — especially with AI features that cost money on every use.
LTV:CAC ratio and payback period
LTV:CAC compares what a customer is worth to what they cost to win. In the examples above, $300 ÷ $50 = 6. A ratio below 1 means you lose money on every customer you acquire. Rules of thumb for a "good" ratio circulate widely, but they depend on the business; treat your own trend over time as more useful than any single target.
CAC payback period = CAC ÷ (ARPA × gross margin), in months.
Example: $50 ÷ ($20 × 0.75) = $50 ÷ $15 ≈ 3.3 months.
Payback matters more than LTV for a small company because it is about cash. If payback is 18 months, you need 18 months of money in the bank for every customer you add before they pay you back. Short payback lets you reinvest sooner, which matters a lot if you are bootstrapping.
Activation and engagement
Revenue metrics lag. By the time churn shows up, the problem started weeks earlier. Two leading indicators help:
- Activation rate — the share of new sign-ups who reach the first meaningful result (sent their first invoice, published their first page, invited a teammate). Define the action that predicts staying, then measure it.
- Active usage — how many paying customers used the product in the last week or month. A paying customer who stopped logging in is a cancellation that hasn't happened yet.
Which metrics to track at each stage
| Stage | Track | Why |
|---|---|---|
| First paying customers | Number of customers, MRR, why people cancel (in their words) | Too few customers for rates to mean much |
| Some traction | MRR movements, customer and revenue churn, activation rate | Find the leaks while they're cheap to fix |
| Spending on acquisition | CAC by channel, payback period, LTV | Know which channels pay for themselves |
| Growing a team | NRR, GRR, gross margin, cohort retention | Show the business can grow from existing customers |
Common mistakes
- Counting annual payments as one month's MRR. A $480 annual plan is $40 of MRR, not $480.
- Mixing free and paid users in churn or ARPA.
- Ignoring failed payments. Cards expire; involuntary churn is still churn.
- Changing definitions silently. Write down how each metric is calculated and keep it.
- Averaging everything. Look at cohorts — customers grouped by the month they signed up — to see whether newer customers stay longer than older ones.
- Using other companies' benchmarks as targets. Published figures vary by market, price point and definition. Compare yourself with last quarter.
Where to get the numbers
If you bill through Stripe or a similar provider, the billing dashboard is the source of truth for subscriptions, and most providers show MRR and churn in some form. Product usage comes from your own database and analytics. A simple internal dashboard that pulls both into one page saves a lot of spreadsheet work.
If your app is built with Mythex, you can ask the agent to build that admin page in the same project: a protected route that reads your subscriptions and usage tables and charts MRR, churn and activation. The dashboard guide covers the layout, and adding analytics covers tracking events like activation. Billing itself runs on your own Stripe account — the Stripe recipe shows how the agent wires it up.
Questions
What is the most important SaaS metric?
Early on, the most useful single number is usually monthly recurring revenue together with churn, because together they tell you whether the business is growing or leaking. Once you spend money on acquisition, add customer acquisition cost and payback period.
What is the difference between MRR and revenue?
MRR counts only the recurring subscription amount normalised to one month. One-off fees, setup charges and usage overages that don't repeat are left out, and an annual plan is divided by 12.
How do I calculate LTV for a new SaaS with little data?
A simple estimate is average revenue per account per month, multiplied by gross margin, divided by monthly churn rate. With only a few months of data, treat the result as a rough range, not a fact, and recalculate as you learn more.
Should I track logo churn or revenue churn?
Both. Logo churn counts customers lost; revenue churn counts money lost. If big customers stay and small ones leave, logo churn looks worse than revenue churn, and the reverse can also happen.