Measuring What Actually Matters

The Metric That Lets You Lie to Yourself

Signups are the favorite metric of early-stage founders because they go up every time you do anything. You launch a Product Hunt post: signups go up. You run an ad: signups go up. You get a mention from an influencer: signups go up. They are responsive, they are visible, and they feel like evidence that the product is working.

They are not evidence that the product is working. They are evidence that your distribution is working. These are different things.

The question signups cannot answer is: after signing up, did users come back? Did they use the product a second time, a third time, a fourth? Did they find enough value to pay for it? Did they tell someone else to use it?

The metric that answers those questions is retention. Everything else in this lesson — LTV, CAC, the ratio between them — is built on the foundation of retention. If retention is broken, the rest of the math does not matter.

The Three Metrics That Tell You If You Have a Real Business

Retention (or its inverse, churn rate) is the percentage of users who come back in the next period after their first use. For a monthly subscription, it is the percentage of users who are still active 30 days after signup.

The thresholds that matter: above 80% month-one retention, you have evidence the product works. Between 60–80%, you have a leak — users are leaving faster than a strong product would allow, and you need to understand why. Below 60%, you have a crisis that no amount of marketing can fix. You are spending money to fill a bucket that is draining faster than you can pour.

LTV (Lifetime Value) converts retention into dollars. The formula is straightforward: average monthly revenue per user, multiplied by the average number of months before a user churns. If users pay $20 per month and stay for an average of 8 months, LTV is $160.

LTV is the number that tells you how much a customer is worth over their entire relationship with the product. It is what you compare against acquisition cost to know whether growth is sustainable.

CAC (Customer Acquisition Cost) is the total marketing and sales spend divided by the number of new paying users acquired in the same period. If you spent $5,000 on marketing in a month and acquired 100 new paying users, CAC is $50.

What to Measure and When

The metrics you track should match the question you are actually trying to answer. Tracking the wrong metric at the right time is as misleading as not tracking at all.

Week 1: Track activation rate — what percentage of new users take the core product action within their first session? This is the first signal of product-market fit. If users arrive and do not activate, you have an onboarding problem, not a retention problem. Also track completion rate for any onboarding flow.

Month 1: Track day-30 retention directly. Did users come back? Also track support ticket rate — if users are confused, they will either write in or leave silently. High support volume at month one is often a leading indicator of churn.

Month 6: Track net revenue retention — are users on the same plan they signed up on, upgrading to higher tiers, or downgrading? Also track churn reasons. Exit surveys and cancellation interviews are the highest-information input you have at this stage.

When the Math Doesn't Work

The path from a broken LTV/CAC to a healthy one is specific. Not all interventions are equal.

Fix retention first. Every additional month a user stays multiplies LTV. Going from 4-month to 8-month average retention doubles LTV with no other change. This is the highest-leverage action available. The way to improve retention is understanding why users leave — user interviews, churn surveys, session recordings — not by changing marketing.

Fix your price second. Most early AI SaaS products are underpriced. The founder is uncertain about value and sets a price that feels "safe." Safe prices produce thin LTVs that make the math not work at any reasonable CAC. If retention is healthy (users stay) but LTV is low, the price is the problem. Raise it. Test with new cohorts. Most products can tolerate a higher price than the founder believes.

Reduce CAC last. Lowering acquisition cost through better targeting, referral programs, or SEO is meaningful — but only when LTV is healthy. Reducing CAC from $60 to $40 when LTV is $45 still produces a business that loses money per customer. The numerator must be fixed before optimizing the denominator.

Where This Course Ends — And What Comes Next

You now have the full picture of what it takes to build an AI product that people pay for.

You validated the idea before building. You understood the unit economics. You know the five failure modes that kill vibe-coded apps in production. You have the minimum viable architecture that — on top of a managed auth provider — prevents them. You know how to budget for AI maintenance and version your prompts like code. And you know which metrics actually tell you if you have a business.

That is the business and operations layer — the part most AI courses skip entirely.

What comes next is the technical layer: the architecture that scales, the multi-agent patterns for complex workflows, the production deployment that handles real load, the evaluation frameworks that measure AI quality systematically. That is Course 2: AI as SaaS: The Technical Foundation.

If you are building your own AI product right now, indecision.io covers decisions and strategy under uncertainty. For deeper reading on the systems behind production AI products, jeremyknox.ai has the technical architecture content that Course 2 is built on.

The foundation is in place. Now go build something people pay for.