The State of Reading 2026: What 311,000 Minutes of Tracked Reading Reveal

TL;DR: We analyzed an anonymized snapshot of booksense reading data: 7,911 completed timed sessions, just over 311,000 minutes of reading, 8,605 notes, and 897 finished books from an early-stage community of readers. The headlines: the median reading session is 26 minutes, peak reading happens at 8pm to 9pm with a second shift before work, readers who quit a book do it at the 10 percent mark, only 1 in 9 readers has ever held a 7-day streak, and the top tenth of readers generate 90 percent of all reading time. National surveys say the median American finished 2 books in 2025. Our data shows what separates the people who beat that number: not speed, but showing up.

Every “state of reading” report you have read is built on surveys. Gallup asks people how many books they read, Pew asks whether they opened one at all, and everyone answers from memory, which is exactly where reading numbers go to get inflated. People misremember, round up, and count the book they mostly finished.

This report is built differently. booksense is a reading tracker for iPhone with a built-in timer, so its data is not what readers say they did. It is what they actually did: sessions timed to the second, pages logged at the moment of reading, notes anchored to real page numbers, books marked finished or abandoned as it happened.

That gives us a small but unusually honest dataset, and this report shares everything interesting we found in it, alongside the best national numbers for context.

About the data (and its limits)

All figures come from a frozen, fully anonymized export of the booksense database taken on July 28, 2026. We computed aggregates only. No individual reader’s data appears anywhere in this report, and none was inspected while producing it.

The snapshot covers:

Metric	Count
Completed timed reading sessions	7,911
Total tracked reading time	311,000+ minutes (about 5,185 hours)
Notes and highlights captured	8,605
Books finished	897
Books abandoned (DNF)	67
Books on want-to-read shelves	2,514
Distinct books in the catalog readers touched	1,701

Honesty first, because a data report lives or dies on it:

• This is an early-stage community. The snapshot spans 584 registered profiles, of which 157 readers completed at least one timed session. These are hundreds of readers, not millions. Treat every percentage accordingly.
• It is not a random sample of humanity. People who install a reading habit app are already trying to read more, and the community currently skews toward nonfiction and business titles (you will see this in the genre table). Where our numbers differ from national surveys, that selection effect is usually why.
• Every session was timed live. Reads imported from a Goodreads library export (a feature booksense shipped this summer) arrive backdated as history and involve no timer, so they cannot inflate the session data. Everything in the session analysis was tracked in real time, in the app.
• The numbers are frozen. This report reads from a single July 28 snapshot. Live counters elsewhere on this site keep moving and include activity beyond this export, so do not expect them to match to the digit.
• Timestamps are server-recorded in a single reference timezone, so hour-of-day figures may shift by a zone or two for any individual reader. The shape of the curve is what matters.

If you want the full survey-based picture of how the world reads, our companion page of reading statistics for 2026 has 58 sourced facts. This report is the other half: behavior, not self-report.

1. The reading session: 26 minutes, 14 pages

The median completed session in the dataset lasts 26 minutes. The mean is 39 minutes, pulled up by a long tail of marathon sessions, and the full distribution looks like this:

Session length	Share of sessions
Under 5 minutes	8.6%
5 to 15 minutes	16.7%
15 to 30 minutes	34.2%
30 to 60 minutes	28.7%
1 to 2 hours	9.3%
Over 2 hours	2.5%

Two things stand out. First, 63 percent of all reading happens in sessions of 15 to 60 minutes. The romantic image of the all-day reader is statistically a rounding error; real reading lives in the gap between dinner and sleep. Second, even the short stuff counts: a quarter of sessions run under 15 minutes, which matches what we argued in how to read more books. The habit is the unit, not the afternoon.

Within a session, the median reader moves 14 pages (the middle half of sessions cover 7 to 25 pages). Sustained pace works out to a median of about 36 pages per hour.

That pace number deserves a pause, because it is well below what lab studies imply. Brysbaert’s meta-analysis of 190 studies puts silent reading at roughly 238 words per minute, or close to a page a minute. Tracked real-world reading runs at barely 60 percent of that. The gap is not slowness; it is reality. Readers pause, reread, take notes (8,605 of them in this dataset), and read harder material than lab passages. If you have been benchmarking yourself against “a page a minute” and feeling behind, stop. Nobody reads like a lab subject.

2. When people actually read: the 9pm peak and the second shift

Aggregating all 7,911 sessions by hour produces a curve with two humps:

• The evening peak. The two busiest single hours in the dataset are 8pm and 9pm. The 7pm to 11pm block accounts for about 24 percent of all sessions.
• The morning shift. The 6am to 11am block is even bigger in total, at 29 percent of sessions, spread across the before-work and commute hours rather than spiking in any single one.

The middle of the day is a dead zone. Reading in 2026 is something people do at the edges of the day, which is worth knowing if you are designing your own habit: the readers in this dataset did not find time, they claimed the same two windows over and over.

The weekly rhythm is just as clear:

Day	Share of sessions
Sunday	17.7%
Saturday	17.4%
Monday	15.5%
Tuesday	13.3%
Thursday	12.5%
Friday	12.0%
Wednesday	11.6%

Sunday is the biggest reading day of the week, with 53 percent more sessions than Wednesday. The weekend as a whole takes 35 percent of all sessions despite being 29 percent of the week. Midweek is where reading habits go to die quietly.

And yes, the New Year effect is real, and it is brutal. January 2026 saw 42 percent more completed sessions than December 2025. By February, session volume had fallen by half from the January peak, to the lowest month in the dataset. If your resolution survived February this year, you beat the curve.

3. Finishing and quitting: the 10 percent cliff

Of the reads in the snapshot that reached a conclusion, 93 percent ended in a finish and 7 percent were abandoned. That is a far higher finish rate than casual reading folklore suggests, and the timing data explains why.

Among abandoned books with recorded progress, the median quit point is 10 percent of the way through the book. A quarter of DNFs happen by the 5 percent mark, and three quarters happen before 30 percent. Almost nobody abandons a book in the second half.

Read together, those two numbers describe a filter, not a failure. Readers audition books hard in the first few chapters, cut fast, and then overwhelmingly finish whatever survives the audition. That is close to the strategy we recommend in when to quit a book, and it turns out disciplined readers already behave this way without being told.

The books that do get finished are not short, either. The median finished book in the dataset runs 320 pages. At the community’s median pace of 36 pages per hour, that is about 9 hours of reading, which at the median session length of 26 minutes means a typical finished book represents roughly three weeks of daily sessions. That is the honest price of a book, and it is lower than most people fear.

Meanwhile, the want-to-read shelves tell their own story: for every book finished in the snapshot, 2.8 books sit on a want-to-read shelf. The TBR grows faster than any reader can burn it down, which is why we wrote a whole guide on organizing your TBR instead of pretending it will shrink.

4. Streaks: rare, fragile, and worth it

booksense is built around daily streaks, so the streak data is the part we watched most nervously. Here is the unvarnished version, computed as each reader’s longest-ever run of consecutive reading days:

• The median reader’s longest streak is 1 day. Most people who try tracked reading have never yet chained two days together.
• About 11 percent of readers have reached a 7-day streak at least once.
• About 3 percent have reached 30 days.
• The longest streak in the dataset is 365 days. One reader read every single day for a year.

That distribution mirrors the habit research almost exactly: Lally and colleagues found new daily habits take a median of 66 days to become automatic, and most attempts collapse long before that. A 7-day streak already puts a reader in the top ninth of this community; the mechanics of getting there are in our guide to reading streaks.

The concentration data makes the same point from another angle. Among readers with at least one completed session, the top 10 percent account for 90 percent of all tracked reading time, and the top quarter account for 99 percent. National surveys show the same shape at country scale: YouGov’s December 2025 survey found the 4 percent of Americans who read 50 or more books accounted for 46 percent of all books read. Reading is not a normally distributed activity anywhere. It is a small core of consistent people doing almost all of it, surrounded by a large group deciding whether to join them.

5. What people read (and how they rate it)

The genre table confirms the selection effect we flagged up front. Counting genres across active and finished reads (books can carry more than one genre):

Genre	Reads
Business	226
Fiction	217
Self-Help	182
Entrepreneurship	109
Psychology	109
Success	75
Fantasy	72
Science	69

A reading-habit app attracts self-improvers, and the shelf shows it. Fiction holds its own in second place, with fantasy leading the fiction subgenres, but the center of gravity is unmistakably nonfiction. (Nationally the picture is different: romance is print’s leading growth category, per the market data in our reading statistics roundup.)

The most-read book in the community, by number of distinct readers, is Atomic Habits, with twice as many readers as the runner-up, The 7 Habits of Highly Effective People. A habit-tracking app whose most-read book is the habit-building manual is either a coincidence or the least surprising finding in this report. Thinking, Fast and Slow, Zero to One, The Alchemist, and The Hobbit round out the top of the list.

Ratings run hot: across the 73 finished reads that received a star rating, the average is 4.55 out of 5, and about two thirds are five-star ratings. Small sample, clear signal: when readers choose their own books and feel free to quit early (see the 10 percent cliff above), the books they finish are books they love. Filtering early does not make reading worse. It makes it better.

6. Notes: the 82 percent page anchor

The snapshot contains 8,605 notes and highlights, and 82 percent of them are anchored to a specific page number. Readers do not just react to books; they cite them.

Note-taking is as concentrated as everything else in reading: among books that have notes at all, the median is 3 notes, while the most annotated tenth carry more than 50 notes each. Some books are read; a few are mined. If you want the mining to pay off later, the page anchor is what makes a note findable again, and spaced resurfacing is what makes it stick. That system is the core of how to remember what you read, and it is built into booksense as note review.

7. How this compares to the national picture

Zoom out from our small, motivated community to the country-scale surveys and the contrast is instructive:

• Gallup’s most recent measure put the American average at 12.6 books per year, its lowest since 1990.
• YouGov (December 2025) found the median American finished 2 books in 2025 and 40 percent finished none. The full breakdown of why the mean and median disagree so violently is in how many books the average person reads.
• Pew Research Center (October 2025) found 75 percent of US adults read at least part of a book in the past year, but only 7 percent belong to a book club. Reading remains mostly solitary.
• Social reading platforms keep growing anyway: The StoryGraph passed 5 million registered users in January 2026, per its Wikipedia timeline, after years of readers migrating away from Goodreads.

Now put our numbers next to those. The median reader in our dataset who actually built the habit reads in 26-minute sessions at 36 pages per hour. Run that arithmetic: a single median session per day finishes a median-length book every three weeks, or about 17 books a year. That does not just beat YouGov’s median of 2; it beats Gallup’s self-reported 12.6, on less than half an hour a day.

The gap between the national median (2 books) and the daily-session reader (17 books) is the entire story of reading in 2026. It is not talent, speed, or free time. It is a 26-minute window, defended daily, usually at 9pm or before work, ideally on Sundays.

Five takeaways for 2026

1. The 26-minute session is the atom of reading. Plan for it. Protect it. Two hours of Sunday ambition is worth less than 26 minutes every evening.
2. Quit at 10 percent or commit. The best readers in the data audition books ruthlessly in the first chapters, then finish 93 percent of what survives.
3. February is the boss fight. The New Year bump is real (+42 percent in January) and it halves by February. A streak that survives February usually survives the year.
4. A 7-day streak makes you a top-ninth reader. Consistency is so rare that a single week of daily reading puts you ahead of 89 percent of people who set out to build the habit.
5. Benchmark against behavior, not lab speeds. Real tracked reading runs at 36 pages per hour, not the mythical page a minute. Set goals for the reader you are, and they will actually compound.

Citing this report

Everything above is yours to use. Cite figures as “booksense, The State of Reading 2026” with a link to this page, and note that booksense figures describe an early-stage tracked-reading community, not a national sample. External figures should be credited to their named primary sources (Gallup, YouGov, Pew, Brysbaert, Lally et al.), which are linked where they appear. We plan to rerun this analysis on a larger dataset next year; if you want your reading in the 2027 edition, booksense is free on the App Store.

FAQ

How long is a typical reading session in 2026?

In booksense data, the median timed reading session lasts 26 minutes, and 63 percent of all sessions fall between 15 and 60 minutes. Marathon reading is rare: only 2.5 percent of sessions run longer than two hours. The typical session moves the reader about 14 pages.

What time of day do people read the most?

Evenings, with a strong second shift in the morning. The two busiest single hours in booksense session data are 8pm and 9pm, but the 6am to 11am block accounts for 29 percent of all sessions. Sunday is the biggest reading day of the week, with 53 percent more sessions than Wednesday, the quietest day.

At what point do readers give up on a book?

Early. Among abandoned books with recorded progress in booksense data, the median reader quit at just 10 percent of the way through, and three quarters of all abandonments happened before the 30 percent mark. Books that survive the first third overwhelmingly get finished: 93 percent of concluded reads ended in a finish rather than a DNF.

Do reading streaks actually last?

Mostly no, which is exactly why they are worth building. The median reader in booksense data has never strung together more than a single day of reading, only about 11 percent have ever reached a 7-day streak, and about 3 percent have reached 30 days. The outlier proves the ceiling: one reader logged a 365-day streak, a full year without missing a day.