Customer acquisition cost: formula and examples

What CAC is and why it matters

Customer Acquisition Cost (CAC) is the dollar amount your business spends to convert one net-new paying customer. It is the cleanest answer to "is our growth engine paying for itself?" — and the first metric a finance partner will challenge when you propose raising the marketing budget. Get the definition right and every downstream metric — payback, LTV/CAC, contribution margin — behaves.

The metric is load-bearing because it sits between two worlds: the paid marketing P&L the CMO defends, and the unit economics the CFO and board read. When CAC moves 20%, somebody asks why — and the analyst who can decompose by channel, attribution window, and cohort is the one whose Slack message gets screenshotted into leadership review.

The CAC formula

The base formula is short enough to fit on a sticky note:

CAC = Total acquisition spend / Number of new paying customers

A concrete example: a B2C subscription app spends $50,000 in March and closes 1,000 net-new paying subscribers. CAC for March is $50. Whether that is good depends on LTV, gross margin, and payback — never on the absolute value alone.

What belongs in the numerator depends on whether you want a "paid CAC" or a "fully loaded CAC". Both are legitimate; they answer different questions.

Paid CAC is what channel managers optimize day-to-day. Fully loaded CAC is what investors and the board want. Report both, label the methodology, and prevent the awkward meeting where two slides disagree by 40%.

Blended CAC vs channel CAC

Blended CAC divides total spend by total new customers, ignoring channel. It is one number, easy to track, and great for high-level dashboards. It also hides almost everything useful. That single number cannot tell you which dollar to cut first.

Channel CAC computes the same ratio per channel: paid search, paid social, influencer, content/SEO, referral, outbound, partnerships. Channel CAC is what you act on. Blended CAC is what you report.

A worked example: a Series B SaaS spends $500,000 in March across channels and adds 1,000 new paid customers — blended CAC of $500.

The blended number says "we acquire customers for $500." The channel split says "paid social is 2.5x more efficient than paid search this month — investigate whether we can shift spend." Without the breakdown, that decision is invisible.

The catch is attribution. Channel CAC is only as honest as the attribution model. Last-touch overcredits retargeting and underdog organic; multi-touch fixes that but adds engineering cost. Pick a model, document it, re-baseline when you change it.

CAC payback period

CAC payback is how many months a customer takes to repay what you spent to acquire them. It is the metric finance teams actually plan against because it converts CAC into a cash-flow question.

CAC payback (months) = CAC / (ARPU × gross margin)

If CAC is $300, monthly ARPU is $50, and gross margin is 80%, then payback is 300 / (50 × 0.8) = 7.5 months. That is healthy for SMB SaaS. The same CAC with 30% gross margin would imply a 20-month payback — fine for an enterprise deal, dangerous for a consumer subscription with monthly churn.

Benchmarks vary, but the rough ranges that show up in board decks:

LTV/CAC ratio

LTV/CAC is the most-quoted unit-economics ratio in venture-backed companies. It compares the lifetime value of a customer to the cost of acquiring them.

The widely-cited rule is LTV/CAC ≥ 3 with payback under 12 months. Treat it as a heuristic, not a covenant. A pre-product-market-fit company will live below 3 for a while and that is fine. A mature company with LTV/CAC of 7 is often signaling it has too much cash sitting unused — the right move there is to spend more, not celebrate the ratio.

When the ratio is below 3, the two levers are obvious in principle and ugly in practice: lower CAC (better conversion, kill bad channels, sharper targeting) or raise LTV (better retention, higher ARPU, more cross-sell). Most companies pull on retention first because acquisition costs are sticky while a churn fix compounds.

SQL recipe: CAC by channel

Monthly CAC-by-channel is the most-asked SQL question in growth analyst interviews. The pattern: join marketing spend by channel against new paying customers attributed to that channel, then divide.

WITH monthly_spend AS (
    SELECT
        channel,
        SUM(amount) AS total_spend
    FROM marketing_costs
    WHERE cost_date BETWEEN '2026-03-01' AND '2026-03-31'
    GROUP BY channel
),
new_customers AS (
    SELECT
        attribution_channel AS channel,
        COUNT(DISTINCT user_id) AS new_users
    FROM users
    WHERE first_paid_at BETWEEN '2026-03-01' AND '2026-03-31'
    GROUP BY attribution_channel
)
SELECT
    s.channel,
    s.total_spend,
    c.new_users,
    ROUND(s.total_spend::numeric / NULLIF(c.new_users, 0), 2) AS cac
FROM monthly_spend s
LEFT JOIN new_customers c ON s.channel = c.channel
ORDER BY cac;

Three things to flag in an interview. First, NULLIF(c.new_users, 0) prevents a divide-by-zero on channels that spent but converted nobody. Second, the denominator filters on first_paid_at, not created_at, because CAC counts paying conversions only. Third, the LEFT JOIN direction matters: starting from spend means channels that wasted money on zero conversions show up, which is exactly what you want to surface.

In BigQuery or Snowflake the cast syntax changes (SAFE_DIVIDE, ::float), but the structure is identical.

Common pitfalls

The most frequent mistake is mixing organic and paid customers into a single CAC. If 40% of March signups came from organic search and word-of-mouth, dividing total spend by total signups produces an artificially low CAC that makes paid channels look better than they are. The fix is to split the denominator: paid CAC uses only attributed paid acquisitions, blended CAC uses everything but is labeled clearly so finance does not double-count.

A second pitfall is including reactivated customers in the new-customer count. A churned subscriber who comes back six months later is a win-back, not an acquisition. Win-back CAC is usually much lower because the customer already knows the product. Mixing the populations muddies channel comparisons and leads to over-investing in retargeting at the expense of net-new acquisition.

A third trap is leaving out non-media costs. The marketing manager's salary, the agency retainer, the attribution tool, the $5 free-trial credit — all of these are acquisition costs in a fully loaded view. Excluding them generates a flattering paid CAC that finance will eventually correct. Publish both versions from day one with a clear definition.

The fourth pitfall is attribution-window mismatch. A user clicks a paid ad in March and converts to paid in April. If you charge March spend against April conversions, March looks great and April looks terrible. The fix is a fixed attribution window (e.g., 30 days from first touch) applied consistently — the exact rule matters less than picking one and never changing it mid-quarter.

A fifth pitfall, expensive when it bites: counting customers who never produced revenue. Trials that ended at day 14, payment failures, fraudulent signups — none of these are paying customers. The denominator should be first successful charge > 0, not "created an account".

When CAC starts rising

CAC creep is the slow kind of metric drift. By the time the board notices, you have usually spent a quarter overpaying for growth. The usual suspects:

• Market saturation. The cheap audience has been mined; each new customer requires reaching deeper at higher CPMs.
• Auction competition. More advertisers bidding on the same intent keywords. Q4 is the canonical example — CPM can double in two weeks.
• Seasonality. Hiring season, back-to-school, holiday — demand spikes and inventory shifts make CAC volatile in predictable ways.
• Funnel decay. Traffic stays flat, but conversion drops because a competitor relaunched, your page got stale, or a tracking pixel broke.
• Channel mix shift. A new channel ramps with higher CAC but better LTV. The blended number rises while unit economics improve — rising CAC as good news.

Not every CAC increase is bad. If LTV rises faster — because retention improved, ARPU climbed, or you launched a higher-tier plan — the ratio gets better even as headline CAC grows. Always frame CAC in the context of LTV and payback, never CAC in isolation.

Interview questions on CAC

Five questions show up repeatedly in growth analyst, product analyst, and finance/strategy interviews:

1. What is CAC and how do you compute it? Give the formula. Specify what is in the numerator (paid vs fully loaded). Mention the denominator is new paying customers, not signups.
2. How are CAC and LTV related? Through the LTV/CAC ratio. Explain why ≥ 3 is a heuristic, not a hard rule, and why payback matters alongside the ratio.
3. CAC jumped 30% last quarter. How would you investigate? Split by channel, check CPM and conversion separately, look at seasonality, validate attribution, check channel mix shift.
4. Write SQL to compute CAC by channel for a month. The recipe above. Bonus points for NULLIF and using first_paid_at not created_at.
5. Blended vs channel CAC? Blended averages everything; channel-level shows efficiency per source. Decisions get made at the channel level; blended is reporting.

To drill metrics like this with worked solutions, NAILDD ships SQL and analytics question sets covering exactly this pattern.

Related reading

• CAC explained simply
• How to calculate CAC in SQL
• How to calculate CAC by channel in SQL
• How to calculate CAC payback in SQL
• How to calculate LTV in SQL
• How to calculate CPA in SQL

FAQ

How is CAC different from CPA?

CPA (cost per acquisition) measures the cost of a target action — a signup, an install, a lead form. CAC measures the cost of a paying customer. Because not every signup converts to paid, CPA is almost always lower than CAC, sometimes by an order of magnitude. In interviews, conflating the two is a quick way to lose credibility — exactly the kind of detail that separates a senior analyst from a junior one.

What CAC is considered "good"?

There is no universal answer because CAC scales with LTV and gross margin. A consumer mobile app might have a healthy CAC of $15; a mid-market SaaS company might be fine with $8,000. The honest interview answer: evaluate CAC against LTV and payback — LTV/CAC ≥ 3 with payback under 12 months is the standard healthy zone, and the absolute value depends on the business model.

Should salaries be included in CAC?

If you want a fully loaded CAC, yes — marketing salaries, sales comp on new logos, agency retainers, and tooling all belong. In practice teams report two versions: paid CAC (media only) for channel optimization and fully loaded CAC for board reporting. Document which version any given chart represents to avoid the marketing-vs-finance dispute over a 40% gap.

How do I lower CAC?

Four levers, in roughly decreasing order of impact. First, improve funnel conversion — same traffic, more paid customers; usually the highest-ROI work. Second, lean into organic and product-led growth — SEO, referrals, in-product invites. Third, sharpen targeting on paid channels. Fourth, kill the worst-performing channels — many teams discover 20% of spend produces less than 5% of conversions.

How often should CAC be recalculated?

Weekly for paid channels, monthly for finance, quarterly for board decks. Common cadence: a weekly paid CAC by channel plus a monthly blended and fully loaded CAC.