CAC, LTV and Payback Period: A Founder's Guide With Formulas

CAC (customer acquisition cost) is what you spend to win one new customer. LTV (lifetime value) is the gross profit that customer generates before they churn. Payback period is how many months of gross profit it takes to earn back the CAC. The core formulas: CAC = sales and marketing cost / new customers; LTV = (monthly revenue per customer x gross margin) / monthly churn; payback = CAC / (monthly revenue per customer x gross margin). For most early-stage startups, payback is the most reliable of the three.

Most guides stop at the formulas. This one walks through a full worked example, explains why blended CAC beats paid CAC for decision-making, shows where LTV quietly lies to young companies, and gives payback targets by business model.

The three formulas on one page

ARPA means average revenue per account, per month. If you sell annual contracts, divide the annual contract value by 12 first.

Two notes before the example. First, every one of these uses gross margin, not revenue. A dollar of revenue from an AI product that spends 40 cents on inference isn't worth the same as a dollar from a lean SaaS app. Second, make sure the time periods match. Monthly CAC with annual churn will give you nonsense.

A worked example from start to finish

Say you run an AI meeting-notes tool for sales teams. These figures are an illustration, not a real company.

• Price: $40 per seat per month, average account has 5 seats, so ARPA is $200 per month
• Cost of revenue (inference, hosting, support): $60 per account per month
• Gross margin: (200 minus 60) / 200 = 70%
• Monthly gross profit per account: $200 x 70% = $140
• Monthly logo churn: 3%
• Last quarter's sales and marketing cost: $90,000 (one marketer, tools, ads, a contractor writer, 30% of the founder's time)
• New customers last quarter: 45

Now the maths.

CAC = $90,000 / 45 = $2,000.

Payback = $2,000 / $140 = 14.3 months.

LTV = $140 / 0.03 = $4,667.

LTV:CAC = $4,667 / $2,000 = 2.3.

So what does that tell you? A payback of 14 months for an SMB product is on the long side. LTV:CAC of 2.3 is under the 3:1 rule of thumb that many investors use. The business isn't broken, but there's work to do. The fastest fixes would be lifting ARPA (more seats per account, an annual plan discount that pulls cash forward), trimming inference cost to push gross margin above 75%, or cutting the channel that's dragging CAC up.

You can run your own numbers in the marketing budget calculator, which estimates CAC and payback from your stage, spend and channel mix.

Blended CAC vs paid CAC: which one to trust

Founders often quote "CAC" and mean different things. There are three versions worth knowing.

Paid CAC flatters you if you leave out the people running the ads. It also flatters you if your attribution model hands paid search credit for customers who were going to buy anyway after reading about you.

Blended CAC is the honest version. It's what growth actually costs the company. It includes salaries, contractors, tools, events, PR and the founder's selling time if the founder is the main closer.

A useful habit: track both. If paid CAC looks great but blended CAC keeps rising, you're probably over-crediting paid and under-investing in whatever is actually creating demand. In my experience that hidden demand driver is often earned media, word of mouth or a community, because those don't show up cleanly in last-click attribution.

What to include in CAC

• Paid media spend across every platform
• Salaries and benefits for marketing and sales staff (pro-rated if they do other work)
• Agency, freelancer and contractor fees, including PR
• Marketing and sales software (CRM, email, analytics, enrichment)
• Events, sponsorships and travel for selling
• Content production (writers, designers, video)
• A fair share of founder time if the founder sells or markets

Leave out product development and customer success for existing customers. Those belong elsewhere.

Why LTV lies to young startups

LTV is the metric I'd trust least before Series A. The formula is fine. The inputs are the problem.

Churn is unstable early

LTV divides by churn. If you have 40 customers and lost one last month, your churn is 2.5% and your LTV looks wonderful. Lose three the next month and it collapses. Small denominators swing hard, so a single month of churn data can triple or halve your LTV.

Churn of 1% implies a 100-month life

Most early companies haven't existed for 100 months. When the formula says customers stay eight years, it's extrapolating far past your evidence. A sensible fix is to cap customer lifetime at 36 months for LTV purposes until you have real cohort data that says otherwise.

Expansion revenue can hide churn

If your best accounts grow while many small accounts churn, average ARPA rises and LTV looks healthy. It's worth splitting LTV by segment so you can see which customers are actually profitable.

Gross margin on AI products moves

Inference prices change, usage patterns shift and a power user can cost five times what a light user costs. Recalculate gross margin quarterly, not once.

A practical early-stage approach: report LTV, but make decisions on payback period and cohort retention. Payback uses fewer assumptions and you can verify it within a year.

Payback period targets by business model

These are typical ranges founders and investors use as rough guides. They aren't hard rules, and they shift with interest rates and how much capital is available.

Enterprise gets more room because contracts are larger, churn is lower and expansion is common. Consumer gets less room because churn is high and nobody signs a three-year deal for a note-taking app.

Cohort payback: the version investors trust most

Average payback hides timing. A stronger version tracks each monthly signup cohort and plots cumulative gross profit against the CAC you spent to acquire that cohort. The month the line crosses CAC is that cohort's real payback. If newer cohorts cross sooner than older ones, your efficiency is improving even if the blended average hasn't moved yet. Two or three cohorts plotted this way will say more in a fundraise than any single LTV:CAC ratio.

If you sell usage-based, calculate payback on the first 90 days of actual usage revenue rather than list price. Many usage-based customers ramp slowly, and payback based on their eventual spend will mislead you.

How to improve each number

Every lever below moves at least one of the three metrics. Pick the one closest to your biggest problem.

Lowering CAC

• Fix activation before buying more traffic, so every signup has a better chance of paying
• Cut the worst-performing channel by blended cost, not by paid CAC alone
• Invest in compounding channels (SEO, earned media, community) that lower blended CAC over time
• Shorten the sales cycle with better qualification, so reps spend time on deals that close
• Reuse content across channels so one piece of work feeds email, social and sales

Raising LTV

• Push gross margin up by renegotiating inference or hosting costs
• Add annual plans with a modest discount to cut early churn and pull cash forward
• Build an expansion path (seats, usage tiers, add-ons)
• Find the activation event that predicts retention and design onboarding around it

Shortening payback

• Raise prices. It's the single fastest payback lever and most early founders underprice
• Take annual payment upfront, which makes cash payback near-instant even if accounting payback doesn't change
• Focus spend on segments with higher ARPA and lower churn

If you're stuck on pricing, AI product pricing strategy for usage-based models covers the trade-offs.

Where PR and brand fit in the CAC maths

I spend my working life on earned media, so I'll be direct about the measurement problem. PR rarely gets last-click credit. A founder reads a CoinDesk or Forbes piece, forgets about it, sees a LinkedIn post a week later, then Googles the brand name and signs up. Last-click says "organic search". What actually happened is more tangled.

That's why blended CAC matters. If you add a PR program and blended CAC falls over the following two quarters while paid spend stays flat, that's a signal worth taking seriously. Pair it with branded search impressions and direct traffic, which tend to rise when coverage lands. The 12-number marketing dashboard shows how to track those alongside CAC.

A quick self-check before your next board meeting

• CAC is blended and includes people, tools and agencies
• Gross margin is recalculated this quarter and includes inference costs
• LTV uses a capped lifetime (36 months or less) unless you have older cohorts
• Payback is shown by segment, not only as one average
• Churn is reported as a three-month rolling average to smooth small-number noise
• You can explain the single biggest lever you'll pull next quarter and why

For a ready-made model, the marketing budget template has tabs for spend by channel, blended CAC and payback, so you can plug in your own figures instead of starting from a blank sheet.

Investors will forgive a long payback period if you can show it shrinking. They won't forgive a number you can't explain.

Want a second pair of eyes on your unit economics before a raise? Book a 30-minute teardown.