Replication Errors and Error Robustness: When Positive Complementarities Are Negative

Abstract

Listen

Translate article

Replicating successful solutions to complex problems is an important strategy for organizational growth and performance improvement, but accurate replication is often difficult, if not impossible. In theoretical accounts, small replication errors are often depicted as valuable sources of variation. Empirical research, in contrast, has failed to identify such positive effects. We extend an existing model of replication to also explain when even small replication errors have negative long- run performance effects. We demonstrate that the error robustness of solutions, driven by the presence of complementarities, is the key structural driver: if there are strong positive complementarities, even small replication errors are negative while with less strong or negative complementarities, small errors are positive. We also revisit how complexity and the quality of the replication template moderate these effects. The results of our experiments with an NK performance landscape model have important implications for both theory and practice.