Headings, subheadings and lists: how the ear follows an article's structure

A reader opening an article doesn't read it first, they scan it. In a second, their eyes catch the subheadings, count the sections, notice there's a bulleted list in the middle, and decide whether the piece is worth stopping for. All of that diagonal reading rests on visual cues. In audio, those cues no longer exist: the voice moves in a straight line, from the first word to the last, with no way for the listener to skim. So the question becomes concrete: what does a synthetic voice do with a heading, a subheading or a list, and how do you write so the structure can still be heard.

A heading read like a sentence, and that's intended

The first surprise is that a subheading isn't announced. The voice doesn't say "heading level two" before reading it: it simply speaks the heading text, then moves on. This is a choice, and a deliberate one. A voice that announced every heading level would sound like a technical screen reader, useful in an assistive context but exhausting for comfortable listening.

What our engine does instead is treat the heading as its own segment, which gives it a natural breath before and after. To the ear, a subheading is therefore marked by a slightly longer pause, not by a label. It's discreet, and it's enough if the headings are well written. Understanding this work on rhythm and reading is the whole point of our page on French text to speech.

The heading that can't stand on its own

From this follows a writing rule. A subheading has to make complete sense once detached from its layout. "Step 3" reads fine in a numbered page, but to the ear, lifted out of its column, it says nothing: step 3 of what? A heading like "Third step: linking the player to the right article" stands on its own, on screen and in the ear. Write your subheadings as short, self-contained sentences, and they'll serve you in both modes.

Lists, where items run together

Bulleted lists pose the opposite problem. On screen, the bullet visually separates each item: the eye knows where an item starts and ends. In audio, the bullet disappears, and if your items are fragments with no punctuation, they run into one another as a mush where the listener loses count. Three words stuck to three more words stuck to three more don't form a list to the ear, they form one confused sentence.

The fix is simple: write items that are real units, short but complete, ending in firm punctuation. A full stop between two items creates the break the bullet provided on screen. Our article on punctuation and how a voice breathes shows exactly how the voice leans on these marks to signal boundaries. A well-punctuated list stays a list when heard; a list of fragments becomes a blur.

When structure becomes real navigation

There's one case where structure doesn't just get heard, it gets steered. From your article's headings, the player can offer timestamped chapters: the listener sees the list of sections, taps the one they want, and the audio jumps to the right second. Listening then regains the ability to skip that reading always had. This is where clear subheadings pay double: first as audible landmarks, then as entry points into the audio file.

That navigation only has value if the headings are themselves informative. A chapter called "Introduction" or "Wrapping up" doesn't say what it holds. A chapter called "What the module writes into WordPress" can be picked out at a glance, or by ear. The quality of headings isn't a cosmetic detail, it's what makes the article browsable, on screen and in the headphones.

Writing once for both modes

The thread through all this is that a text well structured for the eye almost always is for the ear too, as long as you avoid two traps: headings that can't stand alone and list items that aren't punctuated. These aren't constraints specific to audio, they're good writing habits that audio merely makes visible. If you want to go further on what listening adds for a reader following the text with their eyes at the same time, our article on word-by-word guided reading carries the subject on. The simplest test, as always, is to listen: run one of your most structured articles through audio, and find the exact spot where you lose the thread. It's almost always a hollow heading or a list without full stops.

Read next

• Adding an audio version to a WooCommerce store
• Bold, italics, capitals: what text-to-speech doesn't hear
• Add an audio version to your HubSpot articles