So, for example, if you get a new versions of excel, which improves your productivity $100/year, but excel costs $100 this year to upgrade, you haven't gained anything in year 1. That will be a multiplier, but, if the upgrades cost a lot up front (e.g changing technology stack causes initial decrease in productivity and obsolescence of some internal IP + change costs, training, etc) and the productivity increase will be small then it may not be obvious that there is an economic benefit for quite some time.
Of course, a lot of companies think this way and failure to just ditch a lot of old tech years ago, ends up massively holding back productivity relative to peers that did make the upgrades. As a result the company sees low per capita productivity and decides not to raise wages.
Note: this is theory I do not have actual academic studies in practice that can demonstrate this explicitly in a controlled environment, but anecdotal experience certainly makes me suspect that this is what is going on.
edit: grammar
Yes, but this is a tautology (it will always be true that making a capital expenditure of $x/employee/year to augment surplus value $x/employee/year will break even). Productivity itself is scalar (units produced/time), so to arrive at a dollar amount of $100 additional revenue generated per employee would suggest an incredibly small increase in units produced over 1 year.