"Being too busy to train is the moral equivalent of being too hungry to eat." The perception that training takes too much time is the biggest inhibitor to doing it — but no other investment does more to improve productivity.

The right question to ask a manager isn't 'have you ever fired anyone?' but: when you fired them, how did you know with certainty that the employee both understood the expectations of the job and was missing them? Without training, you have no basis for performance management, and firing becomes sloppy and inconsistent.

At Netscape, Horowitz wrote a short document — Good Product Manager/Bad Product Manager — to set basic expectations for his team. Performance instantly improved, previously 'hopeless' PMs became effective, and it soon became the highest-performing team in the company. The lesson: even a simple written articulation of expectations is high-leverage training.

Reading every Netscape exit interview revealed two non-economic reasons people quit: (1) they hated their manager — specifically the lack of guidance, career development, and feedback; and (2) they weren't learning anything. A serious training program attacks both directly.

Beautiful product architectures turn into Frankenstein because companies hire engineers fast and skip training. Untrained engineers replicate existing facilities in ad hoc ways, causing UX inconsistencies and performance problems. 'And you thought training was expensive.'

Enforce functional training by withholding new headcount requisitions until a manager has built a training program for the role. Enforce management training by having the CEO personally teach the course on management expectations — because they are, after all, your expectations.

Track the metric startups usually miss: not candidates screened or offers made, but how many fully productive employees you've added. Recruiting stats without a productivity denominator hide the fact that hiring investment is being wasted.