Freemium Economics

Seufert, a user acquisition and analytics specialist, wrote this as the quantitative treatment of freemium businesses, focused on the modelling rather than on the design.

The central object is lifetime value, and Seufert treats its estimation as the discipline the whole business depends on. He works through the methods: cohort-based retention curve fitting, where retention at defined days is projected forward using a decay function; monetisation curves estimated separately from retention; and the combination into a lifetime value estimate with confidence bounds. He is explicit about why this matters: user acquisition spending is committed before the revenue arrives, so an overestimate of lifetime value produces spending that never recovers, and the error compounds with scale.

The user acquisition material covers the mechanics of paid channels, the measurement of channel quality by the lifetime value of users acquired rather than by cost per install, the fact that scaling a channel raises its cost as the marginal user becomes less valuable, and the modelling of the point at which spending stops being profitable.

The product analytics material covers instrumentation, the design of the data model before the product ships rather than afterward, cohort analysis, funnel measurement, and the statistical treatment of experiments including sample size requirements and the multiple comparison problem when many tests run simultaneously.

He also covers the distribution of revenue across users, which is extremely skewed, and the implications for what the product should optimise.

The book is quantitative and assumes comfort with modelling. It was published in 2013 and the measurement environment has changed fundamentally: platform privacy changes removed the device-level attribution the acquisition modelling depended on, which makes parts of the method unusable as described. The lifetime value framework and the analytics discipline transfer regardless.