Part of the debate about the ‘productivity puzzle’ concerns the potential mismeasurement of digital activities. Specific measurement adjustments explored in previous research appear not to make a quantitatively large difference to real GDP or productivity growth estimates. However, although these potential adjustments may be small individually, taken together they could be wide in scope and quantitatively significant. This paper sets out a taxonomy of the range of potential measurement artefacts arising from digital innovations. It also specifically considers digitally-enabled substitutions in activity across the production boundary. I argue that these, along with other substitutions occurring within the production boundary, go beyond the effects of digital considered in earlier research; and may be making a meaningful contribution to the productivity puzzle as measured on existing statistical definitions.