Why population is not TAM
A market estimate that begins with population must survive several filters before it becomes annual provider revenue. Let A denote affordability, U relevant use and C category choice. These filters are conditional: relevance can depend on affordability, and category choice can depend on both. Let f be annualized purchase frequency and sP the provider's realized share-price outcome.
R = N × Pr(A) × Pr(U | A) × Pr(C | A,U) × f × E[sP | A,U,C]
This is the probability chain rule, not an assumption that the filters are independent. Every conditional term matters. A product can be affordable but irrelevant. Relevant but unnecessary. Needed but substituted. Purchased once every five years rather than annually. And even a dominant provider captures only a fraction of the category.
Consider a high-end device priced at $1,000. Suppose 800 million people could afford it; conditional on affordability, 20% have a meaningful use; conditional on that use, 50% choose the category; replacement occurs every four years; and the provider captures 30% of purchases. Annualized unit demand is:
800M × 0.20 × 0.50 × 0.25 × 0.30 = 6M units/year
Revenue = 6M × $1,000 = $6B/year
The original headline—800 million people can afford a $1,000 product—sounds like an $800 billion opportunity. The annualized provider revenue under these assumptions is $6 billion. The gap is created by relevance, category choice, replacement frequency and share.
This is not a pessimistic model. Changing any assumption can increase the result. Its purpose is to show that a human being cannot be counted as annual revenue merely because a product exists and the person can theoretically pay for it.