A repost and preservation of Constance Crozier's 2020 blog post explaining why forecasting S-curves is difficult while living through them. Using animated examples, the post demonstrates that even with idealized data, the best-fit S-curve estimate remains wildly inaccurate during the exponential and linear growth phases — only converging on the correct curve once the data begins to level off. The framing note observes how early-phase exponential hype tends to fade once invisible asymptotes are reached, and that domain-specific contextual knowledge is needed to make reliable S-curve predictions.
Sort: