Keyword density is dead as a ranking factor and alive as a briefing habit, which is where it still does damage.
Density targets in briefs produce worse content
A brief specifying a term eight times gets a writer optimising for the count rather than the reader. The result reads as written for a checker, because it was, and it performs worse than the same writer working without the constraint.
What to specify instead
The question the page answers, who is asking it, what it must cover to be complete, and what the reader should be able to do afterwards. Coverage of a topic emerges from that naturally, and it is what the brief should be judged against.
Tools that score density measure the wrong thing
Content optimisation tools reporting term frequency against competitors describe correlation, not cause. Writing to hit those scores produces text that resembles ranking pages without being useful, which is a strategy with a known expiry date.
Where editorial judgement belongs
Review drafts by reading them as the target reader. If a term appears awkwardly often, that is the finding — no measurement required. Density was always a proxy for coverage, and reading is a better instrument than counting.