A startup raises $50M at $400M valuation with "150% ARR growth." Three years later it's closing because each customer it acquires loses money for two years before starting to recoup the cost. The company grew toward a bigger financial hole. The error that destroyed hundreds of millions: confusing growth with scalability. The difference between the two lives in unit economics.
Unit economics is the question: "does this business make money PER customer, after covering the cost of acquiring them?". If the answer is yes with good margin, scaling generates more cash. If the answer is no, scaling generates more loss — speed aggravates the problem instead of solving it.
Although it sounds trivial, most teams confuse unit economics with adjacent metrics: gross margin (doesn't include acquisition cost), CAC payback (doesn't capture churn), LTV/CAC (misleading without time context). Fluency in the four key variables — ARPU, gross margin, CAC, retention — separates the financial executive from the operator.
Let's go.
SaaS company: charges $100/month per customer. Gross margin 80%. CAC $1,200. Monthly churn 3%. Is LTV/CAC healthy?
In plain language
Before the mechanics, the four basic questions.
Why do this at all?
Because GROWTH without positive unit economics is a trap. It's exactly what destroyed dozens of 2020-2022 startups that grew 100%+ ARR while burning cash at unsustainable rates. The CFO who looks at unit economics BEFORE approving more sales and marketing spend distinguishes companies that scale healthily from those that accelerate toward collapse. And the operator who understands unit economics knows WHERE to invest the next dollar to move the business in the right direction.
Who uses them?
CFO obviously. CEO at data-driven companies. Director of revenue / sales / marketing — because their decisions directly affect CAC and retention. Investors (VC, PE) use them as filter #1 in due diligence — before growth or TAM. M&A buyer analyzes them by cohort to value. Any operator approving acquisition budget should read them fluently.
When are they calculated and reviewed?
Monthly by cohort of customers acquired. Quarterly with trend analysis (is CAC rising? retention falling?). Before every big marketing spend or sales hiring decision. In M&A due diligence, the buyer builds unit economics by segment and cohort, comparing before and after strategic changes. And critically: when the business "changes" (new product, new market, new segment), recalculate from scratch because unit economics do not transfer automatically.
What if we ignore them?
Three traps: (1) You scale acquisition spend because "growth looks good" without noticing CAC payback rose from 8 to 18 months. At 18 months, each new customer is a cash drain for a year and a half. (2) You calculate LTV with optimistic assumptions (tenure of 5 years when data shows 3) — LTV/CAC ratio looks 4x when real is 2.4x. You make decisions on fictitious numbers. (3) You apply one model's formula to another — using SaaS unit economics for a services business where retention doesn't apply the same. Any of the three leads to over-investment in a business not yet ready to scale, and under-utilization of others that could absorb more capital.
Three models, three readings
Unit economics look different by model. Applying the same assumptions to all three is fundamental error.
B2B SaaS. Average customer $100/month, gross margin 75-80% (high — software scales without proportional marginal cost). Typical CAC $600-1,500 depending on channel. Churn 2-4%/month for SMB, 0.5-1.5% for enterprise. Target payback: <12 months. Target LTV/CAC: 3-5x. Mechanics: retention is EVERYTHING. Raising margin from 75 to 80% is hard; cutting churn from 4% to 2% doubles LTV. Where investment moves the needle most: customer success and product that reduces churn, not more sales heads.
E-commerce. Effective ARPU $40-80/month (combination of frequency × average ticket). Gross margin 30-45% (physical, with cogs). CAC $20-80 depending on channel. Retention: more complex behavior — not binary churn, it's "repeat-purchase frequency" that decays over time. Typical cohorts lose 50% of revenue contribution month 1 → month 6. Target payback: <6 months. Target LTV/CAC: 2.5-4x. Mechanics: winning is repetition. Highest-leverage investment: post-purchase experience, loyalty programs, efficient retargeting — NOT more acquisition spend.
Services. Average customer $5-50K/month depending on ticket. Gross margin 30-55% (people-heavy). High CAC ($5-30K — long sales cycle). Retention varies a lot: project contracts (12-24 months fixed) or retainers (continuous but terminable). Target payback: <12 months. Target LTV/CAC: 2-3x (lower acceptable because ticket is high and margins contribute in absolute). Mechanics: the question is not retention, it's SCOPE EXPANSION — how much the contract grows within each customer. Customer starting at $5K/month and growing to $20K/month in year 2 is worth much more than customer staying at $5K.
The visual below lets you play with the three models and see how payback, LTV, and cohort contribution change with each input movement.
Three models, live
Three model archetypes (SaaS / e-commerce / services). Each with a reasonable default. You move ARPU, gross margin, CAC, churn — and watch payback, LTV/CAC, and the cohort contribution curve over 36 months.
The critical experiment: load SaaS default. Raise churn from 3% to 6%. Watch LTV/CAC fall from ~4x to ~2x. Payback more than doubles. The SAME unit economics that looked winning are now a "grow into profitability" trap. That fragility to churn is the sensitivity #1 a CFO must test BEFORE approving more growth spend.
Interactive visual
Unit economics — three models, one lens
Same framework, three business models (SaaS / e-commerce / services). Move CAC, margin, and retention. Watch payback go from 6 months (healthy) to 36 months (does not scale). You'll learn the question is not "are we profitable?" — it's "do we scale without destroying cash?".
Payback
8.0 months
LTV
$2.5K
LTV / CAC
4.17x
Healthy
Cohort contribution (36 months)
Monthly ARPU
$100/mo
Gross margin
75%
CAC (customer acquisition cost)
$600
Monthly churn
3.0%/mo
What you are seeing
Three critical lessons: (1) Payback is common sense; LTV/CAC is the math. 6-month payback with LTV/CAC = 4x is ideal. 24-month payback with LTV/CAC = 1.5x is "we'll grow into profitability" trap that rarely materializes. (2) Each model has a distinct profile. SaaS wins with retention (low churn × high margin × long tenure). E-commerce wins with repetition (low CAC × decent margin × frequency). Services wins with scope expansion inside each customer. Applying one's formula to another = error. (3) The critical question to the CFO is not "what LTV/CAC do we have today?" — it's "how does LTV/CAC change if CAC rises 30% or retention drops 20%?". Companies that scale have unit economics that SURVIVE moderate stress on any input.
The critical reading of the visual: there are THREE indicators to look at together, not isolated. Payback says "how long it takes to recover the cost of acquiring a customer." LTV/CAC says "how many times the acquisition cost we will recover over the customer's lifetime." Cohort contribution says "how the real trajectory looks" — if the curve rises linearly, the business scales healthily; if it flattens or decays after months 12-18, there is structural deterioration.
Second reading: levers are not symmetric. Lowering CAC 20% is very different from raising retention 20% (which doubles LTV). Raising gross margin 5pp is very different from raising ARPU 5%. When planning improvement initiatives, the CFO must prioritize by the LEVER with greatest unit impact, not the easiest to communicate. For SaaS, retention > ARPU > margin > CAC in typical impact order.
And critical for mid-market: most established companies (not startups) have VERY stable unit economics, but few MEASURE them explicitly. They spend years deciding sales or marketing spend without knowing CAC payback by channel. The new CFO's first exercise: build unit economics by acquisition channel (e.g. inbound vs outbound vs partners vs paid). Almost always reveals that one channel is 5x more efficient than another, and budget decisions don't reflect that difference.
The mechanics: how to calculate and read unit economics
- Define ARPU correctly for your model. For SaaS: average monthly recurring revenue per customer. For e-commerce: average monthly revenue in acquisition cohort (frequency × ticket). For services: contract revenue divided by months of duration. Incorrect ARPU definition is where almost every calculation error begins.
- Calculate fully-loaded CAC, not just paid media. CAC = ALL spending on acquiring customers (paid media + sales salaries + commissions + tools + portion of branding marketing attributable) divided by new customers. NOT just "ad spend / acquisitions" — that underestimates CAC typically 40-60%.
- Measure retention by COHORT, NOT global average. January 2023 cohort (which has 24 months of data) tells you REAL retention. Global average mixes mature cohorts with new ones and hides deterioration. Cohort retention curve is the first tool a sophisticated VC asks for in due diligence.
- LTV = Monthly contribution × average tenure. Do NOT use revenue × tenure — use contribution (revenue × gross margin). It's the most common error. LTV inflated by revenue is 30% higher than real LTV. Decisions made on inflated LTV are systematically over-investment in growth.
- Payback < 12 months for B2B SaaS; < 6 months for e-commerce; < 12-18 months for services. Longer = the business doesn't generate cash to sustain growth without external financing. That's not prohibitive (startups in growth phase accept 18-24 payback with funding) but MUST be conscious and planned, not discovered.
- LTV/CAC > 3x for "healthy, scale with confidence." Between 1.5-3x = works but fragile; below 1.5x = doesn't scale. These bands assume typical context — companies with better cross-sell capability can tolerate LTV/CAC 2x.
- Recalculate unit economics every time the model changes. New segment, new product, new geo, new acquisition channel → different unit economics. Don't assume they transfer. The most expensive error is assuming "the model works in LATAM because it works in US."
- Build unit economics BY CHANNEL, not just aggregate. Aggregate hides which channel is subsidizing the other. Almost always, inbounds are 3-5x more efficient than outbound, but budget is allocated as if they were equal.
- ARPU. What you charge per customer per period. Lever: pricing power, customer mix, purchase frequency. Raising ARPU 10% usually requires model change (move to higher segment, add premium SKU).
- Gross margin. What's left of revenue after direct cogs. Lever: economies of scale in infrastructure, operational efficiency, reducing direct costs. Raising margin 5pp is hard but permanent.
- CAC. Total cost of acquiring a customer. Lever: channel mix (more inbound, less outbound), funnel efficiency, brand awareness reducing paid need. CAC tends to RISE with scale — easiest customers to acquire are first.
- Retention. How long the customer stays and grows. Lever: product, customer success, contracts. Retention is the most underestimated variable — raising retention 20% typically DOUBLES LTV. It is the lever where the marginal dollar goes furthest.
- LTV / CAC. Combination of the four. Synthesizes model health in a single number, but requires all four to be well defined. Without all four correct, LTV/CAC is misleading.
- Payback. How long the customer takes to pay back their own acquisition. It's the time metric (vs LTV/CAC which is magnitude). Companies with good LTV/CAC but long payback need more financing to grow; good LTV/CAC + short payback = self-funding growth.
Adversarial check
Adversarial check
1.Your marketing director proposes increasing paid spend 50% to grow acquisitions 30%. What do you ask BEFORE approving?
2.Why is LTV calculated with revenue (not gross contribution) misleading?
3.SaaS company with LTV/CAC of 4x but payback of 22 months. Is it healthy?
Exit checklist
Suggested re-review: quarterly with growth/marketing review. Annually with any material model change (new segment, new geo, new product). When a new acquisition channel activates, measure that cohort's unit economics from month one — don't wait 12 months to discover if it works.
Optional
Go deeper
Sources and books to dig into the original material