timelets

Два года назад вышла статья*, где на примере одного шведского сталеплавильного завода было показан систематический рост производительности, который не зависил от капиталовложений и инноваций.
Hengel Spiegel Productivity growth

В условиях стабильных внешних условий производительность росла "сама по себе" темпом примерно 2% в год (общий рост примерно в 2 раза за десять лет).

Finding the optimal balance between the various trade-o§s requires a lengthy trail and error process that can be very slow, given that there are many variables that may affect the various trade-offs. Moreover, these variables differ across melt shops, even if they are all using the same equipment. The implication is that learning how to use the melt shop optimally, or ìtweaking the production process is a slow process that involves numerous trade-offs and can take a long time.
Beyond their theoretical and empirical relevance, our findings imply that learning by doing is not simply a function of cumulative output and is not guaranteed automatically. Rather, it is the result of an active experimentation process.

Эволюция - страшная сила.

* Hendel, Igal; Spiegel, Yossi (2013) : Small steps for workers, a giant leap for productivity, CSIO Working Paper, No. 0117. http://hdl.handle.net/10419/92483