Measuring productivity as 'Science, Nature, and Cell (SNC) publications'

From: Staffing labs for optimal productivity

How can this possibly be a metric that people think is acceptable? Even simply number of publications? When will we accept that any reasonable metric of productivity will have to be (a) multidimensional, (b) based on article-level metricsĀ  and not journal-level metrics and (c) those underlying article-level metrics must be based on critical study of the work, not just easily-obtainable but ultimately uninformative features of the work like citation counts.

Finally, we should remember that any such 'productivity metric' will be ultimately be a very rough, error-prone one, as it's often difficult to judge the value of a work until many years after its publication.