How we measure read-aloud time, word count and re-read tolerance

A parent looking for a picture book at eight in the evening is not browsing. They have a specific constraint - four minutes before the child is too tired, a first day of school on Monday, a toddler who will not sit down - and the recommendation sites answer with a theme and a cover image. The one figure they all repeat, the publisher’s reading age, is a marketing range rather than a measurement.

So Owl and Page publishes the numbers that decide whether a book works tonight: minutes, words, sentences per page, and whether the book can take being read fourteen nights running. Those figures either mean something or the site is worth nothing, which is why the method is written out here in full, including the parts that are judgement.

Every word of the main text, counted from the book in hand. That includes refrains and repeated lines, because a child hears them and the adult reads them. It excludes the title page, the copyright page, the dedication, author biographies, and any activity or discussion pages the publisher has bolted on to the back.

Words printed as part of the artwork - a shop sign, a label on a truck, a speech bubble - are counted only when they are part of the read: if you would say it out loud on a normal reading, it counts. Sound effects count as one word each (“CRASH!”), and hyphenated compounds count as one word.

Where a figure is an estimate rather than a count - derived from page count and text density on a book we have not been able to sit down with - the record carries an internal verification flag and the number is treated as provisional. We would rather publish four fields than five fields where one is invented.

The signature figure, and the one no competing site publishes. It is derived, then checked:

1. Divide the counted words by 110 words a minute, the pace of an adult reading aloud to a small child with expression. Conversational speech runs at roughly 150 wpm; a picture book read at that speed is being rushed, and any parent reading one knows the difference.
2. Round to the nearest half minute. Precision beyond that would be false.
3. Read the book aloud and compare. Books with heavy rhythm, long pauses built into the page turns, or a refrain that invites a response come out slower than the arithmetic predicts, and the published figure is adjusted to the real reading.

What the figure does notinclude: questions, pointing at pictures, going back three pages to look at the dog again, or reading it twice. Treat it as the floor - the time the text takes - and add your own child’s overheads on top. In practice a book listed at two minutes is a four-to-six minute bedtime with a chatty three-year-old, and that is still a far more useful number than “ages 0-7”.

The typical number of sentences on a spread, rounded to a whole number. It is the fastest way to separate a book for a two-year-old from a book for a five-year-old, and it does the job that a reading age cannot: one sentence a page is a book a toddler can follow while moving about the room, four or five is a book that needs a child who will sit down.

We count a spread rather than a leaf, because that is the unit a child sees. Where a book varies wildly - a quiet opening and a dense middle - we publish the typical figure and say so in the caveat rather than averaging the difference away.

The only judgement on the site, so here is the whole rubric. It measures how the book holds up for the adultacross repeat readings, because the child’s enthusiasm is rarely the constraint.

5 out of 5

Still fine on the fortieth reading. Rhythm carries it, the jokes work for the reader as well as the listener, and there is something in the pictures you had not noticed in a month of nights.

4 out of 5

Survives a fortnight of nightly repeats, which is the real test. It may wear eventually, but not before the child has moved on.

3 out of 5

Holds up. Bearable through the repeat-request phase without ever becoming the one you hope they pick.

2 out of 5

Good for a handful of readings. The gag or the reveal is the whole book, and it does not survive knowing the ending.

1 out of 5

Wears thin fast. It grates on the adult long before the child is finished with it - worth borrowing rather than buying.

Difficulty of the language relative to the age band the book is aimed at, not in absolute terms. A 1 uses words a child already owns; a 5 is deliberately stretching and expects you to stop and explain. Neither is better. A stretching book is the right choice for a bedtime with time in it and the wrong choice for a child who is already half asleep, and the number lets you tell which is which before you buy.

We use six bands - 0-2, 2-3, 3-4, 4-5, 5-6 and 6-8 - and most books genuinely span two of them. The band is judged from sentence density, vocabulary, whether there is anything for the child to do while listening, and whether the emotional content lands at that age. Publisher ranges are ignored: they exist to widen the market, and “reading age 0-7” describes no child.

Each book carries one specific reason it works - what the book actually does, page by page, for a particular kind of child. Generic praise is treated as a defect here: if the sentence would fit any other book, it gets rewritten.

The caveat is only written when a real one exists. We do not invent a flaw for the sake of balance, and we do not soften a genuine one. Where a book only works read a certain way, or assumes a US school, or has a page that upsets a sensitive four-year-old, the caveat says so plainly.

Where a book is one of ours, it is labelled as ours everywhere it appears, and it is measured on exactly the same terms as everything else.

• A word count, a read-aloud time or a product code we have not verified. The field is omitted instead.
• A star rating aggregated from other people’s reviews. It measures popularity, not fit.
• A price. Prices change weekly and a stale one is a lie, so we link out instead.
• A list assembled to hold a keyword. If we cannot name eight books that genuinely belong on it, the page does not get built.

If a figure here is wrong, it is worth more to us to fix it than to defend it. A disputed word count is settled by counting the book again and republishing the record, and a corrected book is corrected everywhere it appears, since every page is a view of the same record rather than a copy of it.

Owl and Page earns a commission from qualifying Amazon purchases. It costs you nothing and it never decides which books are listed or how they are ranked - the specs and the caveats are written before any link is added.

Links carry a separate tracking code per section of the site, which tells us which shelves are useful and nothing about you personally. Affiliate links are never used in our emails or inside our books.

Method last reviewed 2026-08-13.