Research has successfully conceptualized digital innovation (DI) to cope with its abstract and complex nature. However, scholars are lacking an adequate measure to support empirical understanding. We establish a new text-based measure for DI by applying an unsupervised machine learning algorithm to 10-K reports of S&P 500 firms. For the first time, our measure captures both DI creation activities and DI outcomes. It correlates strongly with patent-based DI activities of firms that have digital patents and also robustly captures DI activities of firms that do not have digital patents. 326 out of 721 firms in our sample have zero digital patents between 1997 and 2019. We use our novel measure to provide evidence of the positive relationship between DI and firm performance across industries. Our study makes an important methodological contribution to DI literature by establishing a novel measure that captures all facets of DI in mature firms.
Zimmermann, Thomas; Kalscheuer, Ulrich; and Fischer-Kreer, Denise, "A text-based measure for digital innovation - uncovering digital innovation and its impact on firm performance" (2023). ECIS 2023 Research-in-Progress Papers. 26.