How Kindle and Medium Calculate Reading Time

The Core Formula

Every reading time estimator uses the same fundamental calculation:

Reading Time = Word Count ÷ Reading Speed (WPM)

The differences between platforms come down to two choices: what they count as a "word," and what reading speed they assume. These two variables produce meaningfully different estimates for the same content.

Medium's Approach: 265 WPM

Medium uses an assumed reading speed of approximately 265 words per minute for its reading time estimates, rounded to the nearest minute. A post with 1,325 words would be estimated at 5 minutes (1325 ÷ 265 = 5.0).

Medium also applies a separate adjustment for images: each image in the body of a post adds approximately 12 seconds to the reading time for the first image, decreasing by 1 second per subsequent image (to a minimum of 3 seconds per image). This accounts for the time readers spend looking at images, which is not captured by word count alone.

The 265 WPM figure sits slightly above the research average for adult silent reading (238 WPM per Brysbaert, 2019). Medium's audience skews toward regular readers who self-select for long-form content, making a slightly higher baseline plausible.

Kindle's Personalised Approach

Amazon Kindle takes a different and more sophisticated approach. Rather than using a fixed assumed reading speed, Kindle estimates reading time based on the individual reader's own reading behaviour.

Kindle tracks the time you spend on each "page" across books you read. After gathering enough data, it calculates your personal reading speed and uses it to estimate how long remaining chapters and books will take you to read. The estimate typically appears as "X minutes remaining in chapter" or "X hours remaining in book."

This personalised approach is more accurate for individual readers than any fixed-speed estimate. A reader who typically takes 8 minutes per Kindle page (approximately 150 WPM) will get substantially different estimates than one who takes 3 minutes per page (approximately 400 WPM) — and both estimates will be calibrated to their actual behaviour.

Kindle's page-based approach also means it adjusts for font size and line spacing. A reader using large text with wider margins (fewer words per screen) will have more "pages" per chapter than a reader using small text. Kindle's algorithm accounts for the actual pages experienced rather than a raw word count.

Other Platforms

Substack

Substack calculates reading time using a fixed word-count approach similar to Medium, using approximately 200 WPM as its assumed reading speed — more conservative than Medium, closer to the research average for general audiences.

WordPress and blogging plugins

Most WordPress reading time plugins (such as Read Meter or Kadence) use 200–250 WPM as their assumed speed. Many are configurable — website owners can set the assumed reading speed to match their audience. A technical blog might set the speed lower (180 WPM) to account for denser content; a casual lifestyle blog might set it higher (270 WPM).

News sites and apps

Many news sites display reading time estimates for articles. The BBC, The Guardian, and New York Times have used reading time indicators at various points. Their implementations vary, but most use a fixed speed of 200–250 WPM calibrated to their audience's expected reading behaviour.

What Counts as a Word?

Most platforms count words as strings of characters separated by spaces — the same definition used by Microsoft Word's word count. This means:

• Hyphenated compound words like "self-aware" count as one word (one string between spaces)
• Contractions like "don't" and "can't" count as one word
• Numbers and codes count as words (12345 = 1 word, A1B2C3 = 1 word)
• Punctuation marks alone (an em dash — ) may or may not be counted depending on the implementation

Some platforms exclude words in code blocks, footnotes, or asides from the reading time calculation, on the grounds that these elements are processed differently from body text. Medium, for example, treats inline code differently from narrative prose.

Why Estimates Are Approximations

Reading time estimates are inherently approximate for several reasons:

• Individual reading speeds vary by roughly 100 WPM either side of the average
• Reading speed varies within the same person depending on content difficulty, familiarity, and purpose
• No estimator can account for pausing to think, taking notes, or re-reading sections
• Content density (how much information per word) varies enormously between texts

The value of a reading time estimate is not precision — it is the quick mental calculation a reader needs to decide whether to start an article now or save it for later. A 4-minute estimate tells you "this is a short read"; a 25-minute estimate tells you "save this for the weekend." The absolute accuracy within a minute or two rarely matters for this decision.

Estimate your content's reading time →