Those numbers aren't precise because the experiment was not allowed to run to conclusion.
From the post:
"It's of note that these conversion rates, are likely lower than their eventual steady state rates because I ended up accepting an offer which essentially stopped the process."
> * 1300 applications > * ~70 interviews (~5.4% conversion) > * About 6 offers
Which result in:
> [T]he bar here isn't all that high because you only need to make one sale to land a job. ... I had too many interviews to really optimize any more than I did.
We have to ask ourselves if the most effective use of the author's (and employer's) time is doing 64 failed interviews, or whether some portion of those 64 prospects could have been "funneled out" with a lower-cost pre-qualification step in the funnel.
(edit: Another criticism is whether the author is filling his funnel with the "best" jobs by using job postings Angel.co. Do a web search for "the hidden job market" for more on this.)