The spread in results across companies usually tracks how much they reorganized, not which model they licensed. Brynjolfsson's been making this case since the J-curve work: the productivity shows up after a firm rebuilds the workflow, not when it buys the tool, which is why two shops on identical software land in completely different places.
The spread in results across companies usually tracks how much they reorganized, not which model they licensed. Brynjolfsson's been making this case since the J-curve work: the productivity shows up after a firm rebuilds the workflow, not when it buys the tool, which is why two shops on identical software land in completely different places.