Skip to main content Accessibility Feedback

How I use utility classes to write more efficient CSS

Today, I wanted to talk about my approach to CSS. It’s a strategy I’ve fine tuned over a decade of working with the web, and helps me create websites and apps that are easier to build and maintain.

Let’s dig in!

Styling elements

I tend to broadly style elements directly whenever I want all elements of that type to look the same.

For example, in my designs, I often want all input, select, and textarea elements to be full-width elements with a slight margin at the bottom. I style those elements directly.

textarea {
	display: block;
	width: 100%;
	margin: 0 0 0.5em;

Same thing with elements like tables. I want every table to have a similar look.

table {
	border-collapse: collapse;
	border-spacing: 0;
	margin-bottom: 1em;
	max-width: 100%;
	width: 100%;

I could use a class for these things, but why would I? It’s extra code in my HTML to achieve the same exact result.

Classes for components

Where I do use classes is when I have a component that does not have a native element, and instead uses a div or span styled a specific way.

Grid layouts are a good example of that. Here’s what the HTML for a grid with three equally sized columns typically looks like in my sites.

<div class="row">
	<div class="grid-third">1</div>
	<div class="grid-third">2</div>
	<div class="grid-third">3</div>

I favor clear, obvious class names that overtly describe what they do. As my students have heard me say a million times, readability is more important than brevity.

The CSS for that might look something like this.

.row {
	display: flex;
	margin-left: -1.4%;
	margin-right: -1.4%;
	justify-content: space-between;

.grid-third {
	margin-left: 1.4%;
	margin-right: 1.4%;
	width: 33.33333%;

I’m using flexbox here, but CSS Grid would also be a great choice, depending on what you’re trying to do.

Modifier classes

Sometimes, I have a typical presentation for a component, but occasionally need to modify it slightly. I use modifier classes for that.

For example, perhaps I want the columns in my .row to have a specific markup order, but be displayed visually in reverse in the UI. I would create a .row-reverse class for that, and use it with my .row class.

<div class="row row-reverse">
	<div class="grid-third">1</div>
	<div class="grid-third">2</div>
	<div class="grid-third">3</div>

Here, the columns are displayed as 3 2 1 in the UI, but are written as 1 2 3 in the source code.

The CSS behind the .row-reverse class might look like this.

.row-reverse {
	flex-direction: row-reverse;

I use modifier classes for elements, too, not just components.

For example, I style my table elements directly, but have a .table-striped class I can use to add zebra stripes to them.

<table class="table-striped">
	<!-- ... -->

The CSS for it looks like this.

 * Adds zebra striping
.table-striped tbody tr:nth-child(odd) {
	background-color: #e5e5e5;

Modifier classes are designed to work with specific elements or components, and typically wouldn’t work if applied to other parts of the design.

Utility classes

Utility classes are the workhorses of my CSS. I use them to nudge-and-tweak my designs in very general ways for a variety of elements or components.

They typically do things like…

  • Modify margin or padding
  • Adjust font-size
  • Center and align things

For example, the .margin-bottom class adds or adjusts the bottom margin on an element.

<div class="row margin-bottom">
	<!-- ... -->

The CSS for it looks like this.

.margin-bottom {
	margin-bottom: 1em;

This could be used on a wide variety of elements and components, not just my grid layouts.

If I wanted to align my text to the right and display it in all-caps, I might do this.

<p class="text-right text-capitalize">Hello, world!</p>

And the utility classes used here look like this.

.text-capitalize {
	text-transform: capitalize;

.text-right {
	text-align: right;

I could create element or component styles to do the same thing, but these are little adjustments that I’d likely make across a wide range of stuff in my design.

Utility classes let me keep my CSS smaller, and my designs for consistent.

One last little trick!

Using the right heading element is important for accessibility.

If your last heading was an h2, and you’re using a heading to show some content that’s a subsection of it, you should use an h3.


<h3>Top 10 best rappers of all time</h3>

<h3>Bad rappers who are for some reason popular anyways</h3>

But sometimes, you want a heading that looks like a different heading level than it should be semantically. For example, I might want headings that look like h5 level headings, even though semantically they need to be h3.

Heading levels should never be used for style. They always convey semantic meaning.

I define heading classes alongside my element styles. Every heading level has a matching class.

.h1 {
	font-size: 1.5em;
	padding-top: 0.5em;

.h2 {
	font-size: 1.3125em;
	padding-top: 1em;

.h3 {
	font-size: 1.1875em;

h4, h5, h6,
.h4, .h5, .h6 {
	font-size: 1em;

.h4 {
	font-size: 0.8125em;
	text-transform: uppercase;

Because classes have more specificity than elements, I can use them to modify the appearance of a heading element while maintaining it’s semantic meaning.


<h3 class="h5">Top 10 best rappers of all time</h3>

<h3 class="h5">Bad rappers who are for some reason popular anyways</h3>

What is this approach called?

I don’t have a cute name for how I write CSS, but I should probably come up with one!