Why CSS Multi-Column Layout Never Took Off
Multi-Column Layout has been around for roughly two decades, yet it remains one of the least used layout tools in CSS. Its syntax is among the simplest in the layout suite, so the neglect is not explained by complexity. The technical gaps and the usability concerns are real, but they are narrower than commonly assumed — and modern CSS has closed many of them.
Specification Gaps That Still Exist
The core limitation has not changed since Rachel Andrew described it five years ago:
"The column boxes created when you use one of the column properties can't be targeted. You can't address them with JavaScript, nor can you style an individual box to give it a background colour or adjust the padding and margins. All of the column boxes will be the same size. The only thing you can do is add a rule between columns."
There is no :nth-column() pseudo-class for alternating backgrounds, for instance. A column-rule accepts border-style values such as dashed, dotted, solid, groove and ridge, but it does not accept border-image, even though both features arrived at roughly the same time.
Browser Support and the Scrolling Objection
Early implementations were patchy, and when the module first shipped most designers had not yet accepted that sites needn't render identically in every browser. Implementations have since broadened considerably, though discrepancies remain, particularly around controlling content breaks. Many developers still describe CSS Columns as broken — an impression that outlives the actual state of support.
The more substantive objection is about reading. Andrew has noted that multicol can produce a reading experience that forces the reader to scroll in the block dimension:
"Neither of these things is ideal, and using multicol on the web is something we need to think about very carefully in terms of the amount of content we might be aiming to flow into our columns."
Careful thought, though, is exactly the job. Headlines, images and quotes can span columns and reset the column flow, which improves readability immediately. Combined with container queries and newer text-sizing units, the excuse for avoiding the module largely disappears.
The Core Properties
Two approaches put content into columns. column-count sets a fixed number of columns. column-width is often the better choice: you specify a width, and the browser fits as many columns along the inline axis as possible, distributing the leftover space between them. The gutter, column-gap, takes any length unit (1em is the default if omitted), and column-rule draws a divider in the gutter.
A Responsive Example Without Media Queries
A flexible article layout can adapt to screen size and to whether a <figure> is present, with no media queries at all. Columns narrow the measure, and text size follows the container rather than the viewport.

<section>
<main>
<h1>About Patty</h1>
<p>…</p>
</main>
<figure>
<img>
</figure>
</section>
The <main> gets a column-width of 40ch, with max-width and automatic inline margins narrowing and centering the content:
main {
margin-inline: auto;
max-width: 100ch;
column-width: 40ch;
column-gap: 3rem;
column-rule: .5px solid #98838F;
}
Flexbox is applied to the <section> only when it :has() a direct descendant <figure>:
section:has(> figure) {
display: flex;
flex-wrap: wrap;
gap: 0 3rem;
}
A shared min-width: min(100%, 30rem) on both <main> and <figure> relies on min(), which returns the smallest of several values — useful for size control under differing conditions. The column styles stay on throughout, so nothing has to be toggled:
section:has(> figure) main {
flex: 1;
margin-inline: 0;
min-width: min(100%, 30rem);
}
section:has(> figure) figure {
flex: 4;
min-width: min(100%, 30rem);
}
Text sizing uses clamp() together with container query units such as cqi, cqw, cqmin and cqmax; ClearLeft's Utopia can generate the values rather than requiring manual calculation.
h1 { font-size: clamp(5.6526rem, 5.4068rem + 1.2288cqi, 6.3592rem); }
h2 { font-size: clamp(1.9994rem, 1.9125rem + 0.4347cqi, 2.2493rem); }
p { font-size: clamp(1rem, 0.9565rem + 0.2174cqi, 1.125rem); }
<main> is declared as the container whose inline size drives those sizes:
main {
container-type: inline-size;
}
The result, when opened in a desktop browser, is an article layout that responds to screen size and to the presence of a <figure>, without a single media query.
Modern CSS Fixes the Older Objections
Geoff Graham, asking "When Do You Use CSS Columns?", raised the same practical worry:
"But an entire long-form article split into columns? I love it in newspapers but am hesitant to scroll down a webpage to read one column, only to scroll back up to do it again."
The column-span property now has solid browser support, allowing headlines, images and quotes to span columns and reset the flow. The solution is not purely technical: content creators and designers also need to consider how often spanning elements appear and to break a layout into shallower sections, so less scrolling is required.
h1, h2, blockquote {
column-span: all;
}
Detached headlines and figures whose images and captions were split across columns were another long-standing issue. The break-after property now has widespread support as well:
figure {
break-after: column;
}
Many problems that shaped early opinions of the module — rem and viewport units, calc(), clamp(), advanced selectors, container queries, and CSS Grid and Flexbox for precise control — simply did not exist when Multi-Column Layout first appeared. Those tools now make it more relevant, not less: adaptive text sizes inside columns, :has() to create columns conditionally on content, container queries to add columns only when a container is wide enough, and combination with Grid or Flexbox for more ambitious layouts.
None of this is new. Whether you last tried the module years ago or never have, it warrants a fresh look — its remaining flaws have not diminished what it can add to a design.



