7 ideas
Individual skills are programmatic sequences of behavior executed largely without conscious deliberation, relying on tacit knowledge that the performer often cannot fully articulate. Just as a skilled person cannot write down everything they know, organizations carry capabilities that cannot be reduced to explicit rules or blueprints.
Much individual and organizational capability is tacit. People and firms can perform a skill well without being able to state the rules they follow, much as a cyclist cannot write out the physics of balance. Because such knowledge cannot be fully written down, imitating another firm is costly and imperfect, and capabilities spread slowly even when competitors can see them.
Firms do not continuously optimize. They look for new techniques mainly when performance falls below an aspiration level. Search is local: firms look near their current routines first and try incremental modifications, so where a firm can go next depends on where it already is.
Industries change through three coupled processes. Search produces variation in routines. Market competition acts as selection, because profitable firms grow by investing while unprofitable ones shrink or exit. Retention works through the persistence of routines inside firms and through other firms imitating successful practices, which shifts the industry's mix of behaviors over time without any firm optimizing.
Organizational routines are the regular, predictable patterns of behavior through which a firm produces, prices, hires, and invests. Routines persist over time and are passed on when firms expand or get imitated, so they carry the firm's heritable traits the way genes do in biology. Differences in routines explain why firms facing the same prices behave differently.
The authors built a simulation in which firms follow satisficing search and imitation rules, with no maximization and no equilibrium. It produced macroeconomic time series of output, capital, and wages that fit Solow's historical US data about as well as neoclassical growth accounting. Fitting the aggregate data therefore does not validate the assumption that firms maximize, since an evolutionary mechanism generates the same patterns.
A firm remembers how to do things by exercising its routines, not by storing knowledge in documents. Each member knows their own part and responds to signals from others. Routines also act as a truce that contains conflict among members, which is why changing a routine can reopen disputes and makes firms resist change.