Every few years the industry rediscovers that people read on screens. The rediscovery arrives dressed as a trend: cards, then full-bleed heroes, then bento grids, then whatever is next. Each one is presented as a way of organising content, and each one quietly assumes the content is short.
It usually is not. Most of what people come to read is a sequence of paragraphs, and a paragraph has requirements that predate the web by about five hundred years. It wants a line length it can be scanned at. It wants enough leading that the eye returns to the right line. It wants a size that does not require leaning in.
Those three constraints are boring, well documented, and routinely ignored.1 They are also the entire difference between a page someone finishes and a page someone abandons at the second scroll.
Measure is not a style choice
The most common failure is line length. Text set across the full width of a desktop monitor produces lines of 140 characters or more. At that width the eye loses the start of the next line often enough that reading becomes work, and the reader stops without ever consciously deciding to.
The comfortable range is somewhere between 60 and 75 characters. This template sets prose at 68 and never lets it out, at any viewport, which is why the text column here stays narrow even when the window is wide. The whitespace on either side is not decoration. It is the feature.
A column of text is a machine for moving the eye from one line to the next. Widen it far enough and the machine stops working.
The same logic governs leading. Tight line-height looks dense and designed in a screenshot; at reading length it closes the visual gap between lines and makes the return sweep unreliable. Somewhere near 1.7 works for most serif faces at body size.
The reader is not admiring the layout.
They are trying to get to the end of a sentence.
What the trends are actually for
None of this means layout is useless. Cards are genuinely good at presenting a set of comparable things. Grids are good at browsing. Full-bleed images are good at establishing a mood in the two seconds before someone decides whether to stay.
The mistake is applying a browsing pattern to a reading task. An essay chopped into six cards is not more approachable; it is the same essay with the connective tissue removed and six new opportunities to leave.
A short test
Open any page you have designed and read it. Not skim it, not review it in a browser at 67% zoom with the inspector open. Read it, on a phone, at the size a reader would. Most pages fail this immediately, and the failure is nearly always one of the three constraints above.2
- Is the line length under about 75 characters?
- Is the body text at least 16px on a phone?
- Can you find where a paragraph ends without hunting?
If all three pass, the page is probably fine, whatever it looks like. If any fail, no amount of layout will rescue it, because the reader will be gone before the layout gets a chance.
Why this keeps happening
Design tools show text as a shape. In a canvas, a paragraph is a grey rectangle you resize to fit the composition, and the composition nearly always looks better when the rectangle is wider. The cost of widening it does not appear until the text is real, in a browser, being read by someone who is not you.
That is the whole mechanism. Not carelessness — a tooling gap between the thing being arranged and the thing being read. The fix is to check in a browser, with real words, early enough that the answer can still change the design.
Trends will keep arriving. The paragraph will keep being the unit. Set it well and most of the rest is detail.