Word count is the number authors track, but it is not the number readers experience. A reader thinks in evenings, commutes and sittings, and a listener thinks in hours. This tool converts one into the other by dividing your text by a set of fixed rates and reporting the result four different ways.
Those rates are constants rather than settings, and it is worth knowing which ones. Reading time assumes 225 words a minute. Spoken time assumes 140, which is the slower pace of reading aloud. Printed length assumes 250 words to a page. The normalised page estimate divides by 187. None of the four is adjustable, so the outputs are one consistent model rather than a range.
Because they are constants, the tool is most reliable as a comparator. Whether a particular reader finishes your book in four hours or six depends on them, not on the arithmetic; but the difference between your 60,000-word draft and your 74,000-word revision is real, and it is the thing this tool measures cleanly.
No. The four rates are fixed constants rather than adjustable settings, which is a deliberate limitation: it means every text you measure is measured identically, so comparisons between drafts and between chapters are consistent. If you need a different pace, divide the word count yourself by the rate you prefer.
Because speech is slower than silent reading, and the tool uses a materially lower rate for it. Silent reading skips ahead, takes in whole phrases and rarely sounds words out, whereas anything read aloud is limited by how fast words can actually be articulated clearly.
Loosely indicative at best. It divides the word count by a flat figure per page, while real page count depends on trim size, margins, type size, leading, how chapters open and how much white space the design carries. Two books of identical word count can differ by dozens of pages.