Filling a layout with no content: placeholder text that does not look like a template

When I was laying out a ticket list for an admin back office, I filled the detail panel on the right with Latin placeholder text. Five people were in the review meeting; three of them stared at the block and asked whether this was the final copy, and another asked whether the block should be bold. The real copy was still held by marketing, promised for the following week.

Another time I was lazy myself: into the heading slot I typed the words heading placeholder, and the front-end developer set the column width from the width of those four characters. When it shipped, the real heading was 13 characters long and the first row pushed the table out of shape. Together those two taught me one thing: placeholder text is not about putting in some characters, it fills visual volume, and volume has to be calculated.

Placeholder text answers one question only: does it fit

My definition of it is now narrow: what it replaces is not the copy but the length of the copy. It can settle three things in advance: how many lines the body needs, how tall a card must be, how wide a table column should be. It cannot settle three things: tone, the real length of domain terms, and where a line will break. The last one is the sneakiest, because a break point only shows up once real text arrives.

As for the three colleagues who questioned the Latin, I read that as a signal that the placeholder had failed: the reader was reading it as content. My fix is not a different kind of placeholder but two separate versions. The one for discussing structure uses Latin, deliberately opaque; the one for discussing placement swaps in Chinese placeholder text sized to match the real copy.

Turning paragraphs and sentences into volume

The generator I use has two settings: paragraphs from 1 to 20, sentences per paragraph from 1 to 10, plus an option to start with the classic opening line and a switch between Latin and Chinese placeholder. Before doing layout I want to know how many words this version will produce, so I measured the classic passage myself: four consecutive sentences total 69 words and 445 characters, an average of 17.2 words and 6.39 characters per word; the first sentence alone is 19 words and 123 characters. Multiplying by those averages gives the table below, which is a scale reference, not a promise.

The last row is what I use to stress-test scrolling on a list page; to fill a normal card body I use 1 paragraph of 2 to 3 sentences. I reversed the order of choosing the volume: measure the character count of the real copy first, divide by the number of sentences I want per paragraph, and work the settings backwards, instead of deciding that three paragraphs feels about right. For this step I usually go back and forth twice between the Lorem Ipsum Generator and a plain text editor, because the sentence count per paragraph has to line up with the real content.

Chinese and Latin placeholders do not test the same layout

This is where I think layouts fail most often. How much fits on a line comes from content width divided by character width; a Chinese character takes about one font size, a Latin letter about half. Take a 360px container, 14px type and a line height of 1.6, about 22.4px: a Chinese line holds roughly 25 characters, a Latin line roughly 51 characters, or about 7 words at 6.4 characters per word.

Two consequences, and I have hit both several times. First, Chinese has no spaces, so the browser can break between any two characters and horizontal overflow is almost impossible; Latin only breaks at spaces, so a URL, a long identifier or a model code with no spaces leaves the last line short, and the real line count is one or two more than a character estimate suggests. Second, the same content set in Latin and in Chinese has a different height, so testing one version is the same as not testing. If a card is built for three reserved lines at 67px, 200 Chinese characters need 179px, and the extra 112px either overflows or gets clipped. That is exactly the moment a review should decide how many lines to allow and whether to expand, not a style fix after launch.

Three rules I set for myself

One: no placeholder text in high-risk slots. Headings, button labels, table cells and navigation items get the longest real candidate string I have, even if it is not final. Placeholder text only goes where length can be approximated, such as body paragraphs.

Two: no x, no TODO, no labels like "heading goes here". They have almost zero volume, so whoever builds the interface sizes the column to zero and every problem is deferred to launch day. If you fill it, fill it with enough characters.

Three: process placeholders in bulk before pasting them into a template. I use Line Operations to add a prefix and suffix to every line, a <p> prefix and a </p> suffix, so dozens of placeholder lines become a fragment that can go straight into the template. Then Word Count checks the gap between the placeholder character count and the real copy count; within 30 percent I consider the test valid, and a bigger gap means starting again.

<p>Lorem ipsum dolor sit amet, consectetur adipiscing elit.</p>
<p>Placeholder copy of the same visual volume goes here; the wording can wait.</p>

That last placeholder line is one I wrote myself, 41 characters, kept specifically to run against the Latin version: the same CSS, the same width, the two side by side, and whichever one overflows is obvious at a glance.

Q: Can placeholder text ship as is?

No. It only guarantees volume; tone, terminology and the key information all have to be replaced. Leave it in and search engines will also treat those paragraphs as meaningless content.

Q: Why not just wait for the real copy?

Because the layout has to be decided now. Waiting means line counts, card heights and column widths all hang in the air, and the front end either stalls or reworks. What placeholder text buys you is hitting the line-count problem today.

Conversely, if the real copy will arrive within two days and the layout only needs a heading and a button, placeholder text is not worth it: leave it empty or put the longest candidate in, at a lower cost.

Q: Do I need to run both Chinese and Latin placeholders?

If the product ships in two languages, yes, both. Even for a Chinese-only build, run the Latin version once at extreme widths, because long strings without spaces expose overflow best.

Q: How do I decide the paragraph count and sentences per paragraph?

Measure the character count of the real copy, divide by the line capacity of the container to get a line count, then work backwards from how many lines I plan to give each paragraph. I always avoid filling by eye in blocks of about three paragraphs, because that approach needs rework nearly every time.

Q: What if people at the review still question the placeholder text?

I swap that version for a Chinese placeholder on the spot and pull the discussion back to structure. If the argument about wording continues after the swap, the deadline for the real copy should move earlier; the problem is not the placeholder.

Back to that review. What I delivered was two versions: one entirely in Latin, one with Chinese placeholder sized to the real character counts, and the card-height table was one I worked out row by row in a 360px container and pinned next to the design. The marketing copy finally arrived at 11 characters per item, narrower than the Latin version but far wider than the four characters of heading placeholder; the column width was set from the 11-character version and nothing went crooked on launch day. That is what placeholder text is for: it does not stand in for the copy, it makes the layout show you the error before the error happens.