While this is true in many institutions there are now useful proxies, specifically a Google Scholar profile or an individual h-index which can be calculated wit

While this is true in many institutions there are now useful proxies, specifically a Google Scholar profile or an individual h-index which can be calculated without publication impact factor. You are correct that it will take a long time to update documentation and policies that reference impact scores, but informally, at least in Machine Learning, people know who the contributors are with high actual impact. In addition because a lot of impactful work is coming out of industry and there is rapid turnover between academia and industry the culture as a whole is evolving.