Guides

How long is a bestseller? Two studies, two answers

Ask whether bestsellers are getting longer or shorter, and the answer depends entirely on which decade you happen to be counting from.

In 2021, the analytics firm WordsRated looked at 3,444 titles that had appeared on the New York Times bestseller list between 2011 and 2021. Average length, they found, had dropped from 437.5 pages to 386, a fall of 11.8 percent. The share of bestsellers running over 400 pages fell too, from 54 percent to 38 percent of the list. And since 2016, short books have simply outlasted long ones on the chart: an average of 7.8 weeks for the shorter titles against 5.9 weeks for the longer ones.

Six years earlier, a different company reached the opposite conclusion about the same kind of list. Flipsnack, in a report picked up by Smithsonian Magazine in June 2015, examined 2,515 bestsellers drawn from the New York Times and Google spanning 1999 to 2014, and found average length rising, from 320 pages to 407, a jump of 25 percent.

Same category of book. Overlapping methods, in the loose sense that both are counting pages on bestseller lists. Opposite headlines. Neither study publishes enough detail about its exact sampling to say definitively where the disagreement comes from, but the most likely answer is also the simplest one: they're looking through two different windows onto the same twenty-five-year stretch, and a trend that looks like growth from one angle can look like contraction from another, depending entirely on where the window starts and ends.

A third data point complicates things further, in an interesting way. In May 2025, a writer named Ray Tabler examined a smaller, differently built sample: 118 of the most-read novels published between 1722 and 2012. Rather than a single steady climb or decline, he found a dip. Bestselling length falls sharply after 1875, bottoms out around 1925 to 1930 (almost nothing in that stretch clears 80,000 words), and then climbs back up in the decades that follow. That shape is compatible with both of the earlier studies without confirming either of them: if you happen to sample mostly from the rising side of that curve, you'll see growth; if you sample mostly from a plateau or a later dip, you'll see decline.

None of the three studies is wrong, exactly. They're measuring different populations, over different years, with different definitions of what counts as a bestseller in the first place, and the resulting numbers simply don't reconcile into one tidy verdict. What's genuinely interesting is how confidently each report states its finding as though length were a single, stable line moving in one direction. It isn't. It's more like weather than climate: real, measurable, and true for the period it covers, but a poor guide to what the next decade's numbers will say.

WordsRated's own data offers a small hint about why the confusion persists. The firm marks 2016 as a turning point, the year after which short books started outlasting long ones on the list. If that pivot is real, then a study sampling mostly the years before it and a study sampling mostly the years after it would report opposite trends without either one making an error, simply because the underlying pattern itself seems to have bent partway through the period both studies claim to cover.

A reader curious about "the" length of a bestseller today is really asking three separate questions at once: which years, which list, and which two points on that list you're comparing. Answer those first, and the two contradictory-sounding studies stop contradicting each other quite so loudly. They were just never describing the same thing.