Interpreting diffusion curves in sociotechnical transitions: how to use S-curves without fooling yourself

S-curves are widely used to describe and forecast technological change, especially in debates on energy transitions, emerging technologies, and industrial transformation. Yet they are also widely misused. This perspective argues that the main problem usually does not lie in the curve itself, but in how it is interpreted. It develops a practical framework for reading diffusion curves in sociotechnical transitions, organised around six recurring sources of error: level, stage, object, purpose, interpretation, and horizon. The article shows why cumulative stocks, annual additions, market shares, and substitution processes should not be treated as interchangeable; why formative-phase technologies are especially vulnerable to overconfident curve fitting; and why levels, logs, and growth rates imply different notions of speed and acceleration. Drawing on classic diffusion theory, forecasting research, and examples from solar, wind, and other technologies, it argues that S-curves are best treated as empirical regularities rather than causal mechanisms or universal laws. Used carefully, they provide disciplined summaries, useful benchmarks, and clearer interpretations of transition dynamics. Used carelessly, they can turn noisy, path-dependent, and politically contested processes into stories of inevitability.