The short answer to this is, in my experience, there is no IDEAL %, and the video above, published many years later after I first collected these opinions (2008), would go to confirm many of the opinions on this page.
There is no one-size-fits-all optimal ?keyword density? percentage anybody has ever demonstrated had direct positive ranking improvement in a public arena.
I certainly do not believe there is a particular percent of keywords in words of text to get a page to number 1 in Google. While the key to success in many niches is often simple SEO, search engines are not that easy to fool in 2016.
There may not be a perfect % for you to aim for ? but I do think you run the risk of tripping keyword penalty filters if you, for instance, were to keyword stuff a page and every element on it with your focus terms.
I write natural page copy which is always focused on the key phrases and related key phrases. I never calculate density in order to identify the best % ? there are way too many other things to work on. I have looked at this, a long time ago.
If it looks natural, it?s ok with me. Normally I will try and get related terms in the page, and if I have 5 paragraphs, I might have the keywords I am focused on in just a few elements and on the page text.
I think an optimal % is a bit of a myth although there are some who disagree. I do think if you can pick up page level penalties for keyword stuffing, there must be an amount or ratio that is used to penalise the page, so i can see how it causes confusion.