Churn measures the percentage of *existing* customers lost over a specific period, regardless of how many new customers were acquired. This strict definition isolates retention issues from acquisition success, providing a clear and un-muddled health metric for the customer base.

A sale is just the first step. The true measure of product-market fit is high retention, specifically when the product becomes so integrated into a customer's workflow that the idea of canceling their subscription would be bizarre and disruptive. Founders should be designing for this "weird to cancel" status.

CoStar Suite has achieved a status akin to the Bloomberg Terminal in finance. It is the indispensable industry standard with immense pricing power and high switching costs. This dominance means customers often have a love-hate relationship with the service, viewing it as a necessary evil.

The true indicator of Product-Market Fit isn't how fast you can sign up new users, but how effectively you can retain them. High growth with high churn is a false signal that leads to a plateau, not compounding growth.

Reacting to churn is a losing battle. The secret is to identify the characteristics of your best customers—those who stay and are happy to pay. Then, channel all marketing and sales resources into acquiring more customers that fit this 'stayer' profile, effectively designing churn out of your funnel.

Every business has a growth ceiling where new customer acquisition is completely offset by churn. No matter how many new customers you add per month, your business will stop growing once churn equals acquisition. Plugging this 'leaky bucket' is more valuable than pouring more water in.

While individual AI companies see slightly lower retention than SaaS, Stripe's data reveals customers often churn from one provider directly to a competitor, and sometimes switch back. This indicates the problem being solved is highly valued, and the churn reflects a rapidly evolving, competitive market, not a lack of product-market fit for the category itself.

Churn is a lagging indicator. It's the delayed consequence of past product roadmap decisions and a failure to stay aligned with customer needs. By the time a customer leaves, the strategic misstep has already occurred, making churn analysis a post-mortem on old strategy, not a real-time event.

Industrial tech tools build a deep moat through stickiness. Once integrated, they become the trusted system of record not just for the company, but for its partners and government customers. This ecosystem dependency, like Palantir's, makes them nearly impossible to replace, leading to near-zero churn.