How subscription businesses measure growth: the metrics that matter

When Slack reports quarterly earnings, the headline might focus on revenue or net income. But investors in subscription software care about a different set of numbers: ARR (Annual Recurring Revenue), NRR (Net Retention Rate), and churn. These metrics are non-GAAP, invented by the industry, and often more predictive of future growth than traditional GAAP earnings.

The reason is structural. A subscription business is fundamentally different from a product business. A software license sold once generates revenue once. A SaaS subscription generates revenue every year—potentially forever—as long as the customer doesn't cancel (churn). That creates a predictable revenue stream that's valuable and sustainable, but only if the company can keep customers renewing and expand into their accounts.

The industry developed SaaS metrics to track precisely this: how much recurring revenue the company has, how fast it's growing, how fast customers are leaving, and whether existing customers are expanding. These metrics are now standard in venture capital, and most public SaaS companies disclose them in earnings releases and investor presentations.

Quick definition: ARR = the annualized value of all active subscriptions; MRR = ARR divided by 12; NRR = the percentage of revenue retained from existing customers, including expansion; Churn = the percentage of customers (or revenue) lost to cancellation each period.

Key takeaways

1. ARR is the foundation — It's the total annual value of all active recurring contracts, the single best measure of the size of the subscription base and its trajectory.

2. NRR over 100% is the holy grail — If customers expand faster than they churn, revenue grows from the existing base alone, even without new customer acquisitions. This is a sign of a very healthy business.

3. Churn is the silent killer — Even a company with 20% annual customer churn looks like it's growing if it acquires enough new customers. But churn indicates product-market weakness and unsustainability.

4. MRR is ARR / 12 but tells a monthly story — For high-growth early-stage companies, tracking MRR momentum (month-to-month growth) reveals acceleration or deceleration before quarterly results do.

5. CAC and CAC payback period predict profitability — How much does it cost to acquire a customer (CAC), and how long until that customer's subscription pays back the acquisition cost? If CAC payback is 14 months and customer lifetime is 3 years, that's a good business.

6. Magic number predicts self-sufficiency — Magic number = net new ARR / sales and marketing spend. A magic number above 0.75 suggests the company can profitably grow without raising capital.

Understanding ARR and how it grows

ARR is the lifeblood metric for subscription businesses. It's calculated simply: take every active customer subscription at the end of the period and annualize it. If you have 1,000 customers paying $1,000/year on average, ARR is $1 million.

ARR grows from three sources:

1. New customers (new logo or new seat sales): Each new customer adds to ARR immediately. If 100 new customers each with an annual contract of $10,000 sign in Q1, ARR increases by $1 million.

2. Expansion (upsell and cross-sell): Existing customers upgrade to higher-priced plans, add users, or purchase new modules. This is called "dollar-based net retention" because it measures expansion in dollars, not headcount. If a customer already paying $10,000/year adds a product for $5,000/year, ARR increases by $5,000.

3. Contraction and churn (headwind): Customers downgrade (contraction) or cancel (churn), reducing ARR. If 50 customers paying $10,000/year each cancel, ARR falls by $500,000.

The net of these three—new customers, expansion, minus churn and contraction—is called net new ARR, the truest measure of organic growth.

ARR can be tracked at the company level (total ARR) or at the customer level (ARR per customer or ARPU, average revenue per user). Tracking ARPU helps spot whether expansion is from more customers or from higher spending per customer.

NRR: the metric that predicts everything

Net Retention Rate (NRR) measures what percentage of revenue is retained from existing customers, including expansion. It's calculated as:

NRR = (Beginning ARR - Churn + Expansion) / Beginning ARR

If you start the quarter with $10 million in ARR, lose $1 million to churn, but expand existing customers by $2 million, NRR is ($10M - $1M + $2M) / $10M = 110%.

An NRR above 100% is extraordinary. It means the company's existing customer base is growing faster than it's declining. In other words, revenue growth can come entirely from existing customers; the company doesn't need to acquire a single new customer and would still grow. This is a signal of a very healthy, scalable business. Companies with NRR above 120% are exceedingly rare and valuable.

NRR below 100% is fine—most companies have NRR between 90% and 110%. It simply means existing customers are leaving or contracting faster than they're expanding, so the company must acquire new customers to grow. But NRR below 80% is a major warning flag. It suggests the product isn't sticky, customers are unhappy, and the company is on a treadmill of acquisition-to-replace-churn.

NRR is a leading indicator of a company's profitability and sustainability. A company with high NRR can afford to invest heavily in customer acquisition because existing customers will grow the revenue base. A company with low NRR must be disciplined about customer acquisition costs or risk burning cash forever.

Churn: the erosion measure

Churn rate measures the percentage of customers (or revenue) lost in a period, typically reported as:

• Customer churn: Percentage of customers cancelling (e.g., if you start with 1,000 customers and lose 50, customer churn is 5%)
• Revenue churn (or dollar churn): Percentage of ARR lost (e.g., if you start with $10 million ARR and lose $500,000 to cancellations, revenue churn is 5%)

A 5% monthly churn rate sounds small. Annualized, it's about 50% of the customer base gone per year. But it's easily overlooked in press releases. Always annualize monthly churn to see the true damage.

For SaaS, churn has a massive impact on growth rates. A company with 100% new customer growth but 40% annual churn is barely growing net revenue. A company with 20% new customer growth and 10% churn is more valuable, all else equal.

Churn comes from two sources: involuntary churn (credit card failures, billing issues) and voluntary churn (customers cancel because they don't need the product anymore or switched to a competitor). Involuntary churn is fixable through better billing operations. Voluntary churn is a product-market fit problem and much harder to fix.

Healthy SaaS companies report annual revenue churn below 10%. 5% or less is very good. 15% or higher is a red flag. For enterprise SaaS, churn is often even lower (2–5%) because enterprise contracts are sticky.

The supporting cast: CAC, LTV, and magic number

Customer Acquisition Cost (CAC): The total sales and marketing expense divided by the number of new customers acquired in a period. If a company spends $1 million on sales and marketing and acquires 100 customers, CAC is $10,000. This should be compared to the ARR of those customers. If CAC is $10,000 and the customer's ARR is $5,000, the payback period is two years. That's reasonable; bad is if payback is longer than half the expected customer lifetime.

Lifetime Value (LTV): The total revenue a customer generates over their lifetime, minus the cost of serving them. If a customer stays for 5 years at $10,000/year ARR, and the cost to serve is 30% of revenue, LTV is ($10,000 × 5) × (1 - 0.30) = $35,000. This should be compared to CAC. A rule of thumb: LTV / CAC should be at least 3x. If CAC is $10,000 and LTV is only $25,000, the unit economics are too thin.

Magic Number: (Net new ARR in quarter) / (Sales and marketing spend in prior quarter). A magic number above 0.75 suggests that for every dollar spent on sales and marketing, the company generates 75 cents of incremental annual revenue. That's very efficient. Below 0.5 is a warning that unit economics are poor and growth is uneconomical.

Payback Period: How many months until a customer's cumulative revenue covers CAC. A 12-month payback period means the customer pays back the acquisition cost in the first year. This is a key metric for understanding unit economics and capital efficiency.

Real-world examples

Salesforce and the NRR advantage: Salesforce has reported NRR above 125% for years, meaning its existing customer base expands revenue by 25% annually through upsell and cross-sell. This is why Salesforce can raise prices, invest heavily in new products, and still grow revenue at healthy rates even when new customer acquisition slows. The existing base is a flywheel.

Slack and the churn challenge: Slack's customer churn rate has been reported in the 80–85% range for annual revenue churn (mid-market customers), though lower for enterprise customers. This high churn is offset by aggressive new customer acquisition, but it indicates that many small and mid-market customers don't renew. It's a warning that the product-market fit in that segment is not as strong as in enterprise.

Zoom and the magic number explosion: During the pandemic, Zoom's sales and marketing efficiency (magic number) exceeded 1.0—for every dollar spent on sales and marketing, the company generated more than a dollar in new ARR. This is exceptionally rare and unsustainable, driven by viral adoption. Post-pandemic, Zoom's magic number normalized to 0.5–0.75 as growth moderated.

HubSpot and the expansion story: HubSpot's NRR has been consistently above 110%, driven by customers expanding from marketing software to sales, service, and other modules. This expansion generates recurring revenue from the existing base, making HubSpot's growth more predictable and less dependent on new logos.

Real-world examples and benchmarks

Datadog's NRR magic: Datadog has reported NRR above 130% for years, meaning its customer base expands by 30% annually. A customer might start with monitoring and expand to log management, infrastructure monitoring, and security, creating multi-product expansion. This expansion is the engine of Datadog's growth; the company doesn't need to acquire many new logos to grow revenue because existing customers spend more over time.

Twilio's high CAC, solid unit economics: Twilio pays approximately $1.50–2.00 in sales and marketing costs for every dollar of quarterly revenue generated. This seems high until you remember that customer lifetime is often 3–5+ years. A customer acquired for $50,000 CAC (Twilio's CAC) and spending $10,000 annual ARR will have lifetime value of $30,000–50,000, yielding a positive LTV/CAC ratio of 0.6–1.0x (below the ideal 3x, but acceptable for a growth company capturing a large market).

Atlassian's magic number decline and profitability: Atlassian's magic number fell from above 0.75 to 0.5x as the company scaled and new logo growth slowed. This prompted the company to shift focus from growth at all costs to profitability, turning off aggressive acquisition spending and optimizing sales efficiency. Magic number can decline as companies mature, but the decline should come with improving operating margins.

Stripe's opaque metrics: Stripe, a private company, discloses minimal information to the public. But venture investors have noted that Stripe's customer acquisition is highly efficient (magic number above 1.0x) because customers sign up through product experience rather than sales teams. This makes Stripe's unit economics exceptional and partially explains its $95 billion+ valuation.

Churn in practice: Slack vs. Zoom: Slack and Zoom both experienced high growth, but Slack's annual revenue churn for smaller customers ran 40–50%, while Zoom's was much lower (10–15%). Zoom's lower churn meant customers were more sticky and valuable long-term. Slack's higher churn meant the company had to constantly acquire new customers to grow, a more expensive model. Both have succeeded, but the path is different.

How to use SaaS metrics for valuation

The traditional valuation approach for SaaS companies is to multiply ARR by a multiple (often 5–20x depending on growth rate and churn). The reason is that ARR represents the annual run rate of a predictable, recurring revenue stream. A company with $100 million ARR growing 40% annually is worth more (higher multiple) than a company with $100 million ARR growing 10% annually, all else equal.

The multiple also reflects NRR and churn. A company with NRR above 120% and churn below 5% can justify a 15–20x multiple because growth is sustainable without expensive acquisition. A company with NRR below 90% and churn above 20% might trade at 5–8x because revenue stability is questionable.

Here's a simplified framework:

ARR Growth	NRR	Churn	Valuation Multiple
50%+	>120%	<5%	15–20x
30–50%	110–120%	5–10%	10–15x
15–30%	100–110%	10–15%	5–10x
<15%	<100%	>15%	2–5x

NRR and churn matter more than headline growth because they indicate sustainability. A company with low churn and high NRR can cut customer acquisition spending and still grow. A company with high churn must spend aggressively to maintain growth, which eventually becomes uneconomical.

Magic number and the efficiency frontier

The magic number is the most forward-looking metric for SaaS profitability potential. Here's why:

Magic number > 0.75: Efficient growth. For every dollar spent on sales and marketing, the company generates 75 cents of annual recurring revenue. This is sustainable because the customer lifetime value (assuming 3–5 year lifetime) far exceeds CAC.

Magic number 0.5–0.75: Moderate efficiency. Still profitable eventually, but requires sustained growth to justify the spend. As growth slows, the company will need to reduce sales and marketing spend or accept lower profitability.

Magic number < 0.5: Inefficient growth. The company is spending heavily but the return in ARR is low. This is only justifiable if the company is building market share in a winner-take-most market or if churn is falling dramatically.

Many high-growth SaaS companies sacrifice magic number for growth, running at 0.3–0.5x magic number to expand. This is fine as long as management has a path to improving efficiency later. When the company's growth slows and it's still stuck at 0.3x magic number, that's a red flag.

CAC payback and the cash flow timeline

CAC payback period is how many months until a customer's contribution margin covers the customer acquisition cost:

CAC payback = CAC / (Monthly ARR × Gross Margin)

If a customer is acquired for $10,000 (CAC) and has $2,000/month ARR with 70% gross margin, the monthly contribution is $1,400. Payback is $10,000 / $1,400 = 7.1 months.

A CAC payback of 12 months or less is very good. A payback of 14–18 months is acceptable for high-growth companies. Above 18 months, the math becomes challenging unless NRR is very high.

The reason payback matters is cash flow. If a company is acquiring customers at a 12-month payback, it needs enough cash on hand or access to credit to fund acquisition until those customers generate cash. This is why many SaaS companies have enormous gross margins (60–80%) but still raise capital: they're managing cash flow timing, not profitability.

FAQ

What's a good NRR for a SaaS company?

Above 100% is healthy. 110%+ is very good. 120%+ is exceptional. Below 90% is a warning that churn is exceeding expansion; the company will struggle to grow without heavy acquisition spending. Enterprise SaaS (Salesforce, Workday) often has NRR above 120% because customers are deeply integrated and have high switching costs. SMB SaaS (Slack) has lower NRR but higher volumes.

How is ARR different from Annual Recurring Revenue reported by the company?

They should be the same. Some companies report ARR as of a point in time (end of quarter), while others report run-rate ARR (extrapolating the end-of-quarter base to a full year). Always check which definition the company uses.

Is churn rate annual or monthly?

Always ask. If a company reports 5% churn, is that monthly (about 50% annually) or annual (5% per year)? The difference is huge. Good companies report both monthly churn and annualized churn to avoid confusion.

Why do some companies not disclose churn?

Because it's negative. A company with 20%+ annual churn will often omit it or bury it in the MD&A. If a company doesn't disclose churn and the business is SaaS, assume it's bad and ask management directly.

How do I calculate magic number from a press release?

Net new ARR (current quarter minus prior quarter) divided by sales and marketing spend in the prior quarter. If a company doesn't disclose it, you'll need to reverse it from the MD&A discussion of sales efficiency.

Why does magic number sometimes exceed 1.0?

Because a company has very high unit economics. A viral or self-serve product might acquire customers for minimal CAC (mostly product development and some marketing) but high ARR. Or a company might be pulling forward multi-year bookings into the current quarter. Above 1.0 is unsustainable long-term but signals exceptional unit economics in the short term.

Related concepts

• Customer concentration: What percentage of ARR comes from the top 10 customers? High concentration is a risk; loss of one customer materially impacts growth.
• Dollar-based net retention vs. logo-based net retention: Dollar NRR measures revenue expansion; logo NRR measures customer retention headcount. Both matter, but dollar NRR is more important for investor analysis.
• Rule of 40: An investor rule of thumb suggesting a SaaS company's growth rate plus FCF margin should exceed 40%. A company at 30% growth with 15% FCF margin meets the rule; 20% growth with 25% margin also works.
• Payback period and CAC ratios: Related measures of unit economics, often used interchangeably with magic number.

Common mistakes

1. Confusing ARR with bookings. ARR is the annualized value of active subscriptions today. Bookings is the total value of all deals signed. ARR is the recurring base; bookings is the total sales. A company can have high bookings but low ARR growth if bookings are long-term.

2. Overlooking customer concentration. A company with 30% NRR sounds good until you learn that 40% of ARR comes from three customers. Loss of one customer would be catastrophic. Always check the disclosure of customer concentration.

3. Assuming CAC is constant. As a company scales, CAC often increases because it's exhausting low-cost channels first. A company that acquired customers cheaply early on might face much higher CAC later, shrinking margins and requiring price increases.

4. Confusing annual churn with monthly. Monthly churn can be manipulated (if many customers cancel on the same date, monthly looks worse than annual). Always ask for both and understand the timing of customer cancellations.

5. Using historical magic number to forecast profitability. Magic number changes with scale. An early-stage company at 1.0x might have a magic number of 0.6x as growth accelerates and CAC increases. Use magic number as a current-period signal, not a permanent fixture.

Summary

SaaS metrics—ARR, NRR, churn, CAC, and magic number—are the language of subscription business analysis. They strip away the accounting gimmicks and reveal whether the underlying business is healthy and sustainable. A company with growing ARR, NRR above 110%, churn below 10%, and magic number above 0.75 is a high-quality business. A company with all of those inverted is burning cash to grow and has structural problems.

The best SaaS companies combine strong growth (ARR growth 30%+) with exceptional unit economics (magic number 0.75+) and customer stickiness (NRR 120%+, churn 5%). These companies can achieve profitability while still investing in growth; they're the ones that compound wealth for investors.

When analyzing a SaaS company, start with these metrics. They tell you more about future earnings quality than GAAP revenue or EBITDA ever will.

Next

Same-store sales and comparable sales

---

SaaS companies with NRR exceeding 130%, annual revenue churn below 3%, and CAC payback below 12 months represent the highest-quality subscription models and are capable of self-funded profitability at scale.