The badge sits at the top of the article like a price tag on a garment: 5 min read. It is small, gray, and almost always positioned next to the byline, where a newspaper once printed the dateline. It looks like metadata. It reads like a promise. It is neither.
This is a sentence about time, but it is not about reading. It is about the decision to read — a decision the badge is designed to make before you have any evidence about whether the piece is worth the minutes it claims to cost. The badge does not tell you what the article is about, who wrote it, or whether it will change your mind. It tells you the exchange rate. That is the transaction.
What the badge actually measures
The arithmetic behind most reading-time estimates is not complicated. Take the word count, divide by a words-per-minute constant, round up. The constant varies by platform and by era, but the logic is the same: reading is treated as a throughput problem, and the reader as a pipe.
The most commonly cited constant in web usability work is 200 words per minute, a figure Jakob Nielsen has used for years. In a 2008 analysis of 45,237 page views drawn from a naturalistic browser study by Weinreich, Obendorf, Herder, and Mayer, Nielsen found that users spent only about 4.4 seconds more for each additional 100 words on a page. At 250 words per minute — a faster pace he chose because the study participants were unusually literate — that works out to roughly 18 words read per 100 added. The rest of the time went to layout, navigation, images, and the general business of being on a web page rather than reading one.
Nielsen’s conclusion was blunt: on an average page, users have time to read at most 28% of the words during an average visit, and 20% is more likely. That figure comes from a specific dataset collected in 2005, and it should be treated as a historical finding rather than a permanent law of human attention. But the direction of the finding has been replicated often enough in usability research that it is not seriously contested: people do not read web pages the way the badge implies.
So the badge is not a measurement of reading. It is a measurement of text, divided by a constant that describes an idealized reader who does not exist. The badge is a unit conversion, not an observation.
The badge as interface rhetoric
Every interface string is an argument. A hold message argues that your time is less important than the company’s queue. A cookie banner argues that consent is a formality. A 5 min read badge argues that the value of an article can be expressed in minutes, and that you, the reader, are the kind of person who budgets minutes.
What does the badge promise? Three things, none of them stated:
- That the article is finite. It has a bottom. You will reach it. The badge converts an open-ended encounter with ideas into a bounded task.
- That the time is knowable in advance. You can decide now, before reading, whether to spend the minutes. This is the transaction.
- That the time is the relevant cost. Not difficulty, not density, not the possibility that the piece will require you to think. Just minutes.
What does the badge omit? Almost everything that determines whether reading is worth doing. It says nothing about comprehension. A 5-minute read of a dense legal analysis and a 5-minute read of a listicle are not the same experience, but the badge renders them equivalent. It says nothing about purpose. A reader looking for a definition and a reader looking for an argument will spend different amounts of time on the same words. It says nothing about the reader. A person reading in a second language, a person using a screen reader, a person with a cognitive disability, a person who stops to take notes — all of them are averaged into a single number that describes none of them.
The Web Content Accessibility Guidelines, now at version 2.1, are explicit that content should be perceivable, operable, understandable, and robust, and that conformance at the highest level still will not address every user need, particularly in the cognitive and learning areas. The reading-time badge is not an accessibility failure in the formal sense. But it is a small, pervasive example of the gap the guidelines acknowledge: a piece of metadata that assumes a uniform reader and a uniform reading act.
Who the badge protects
Read the badge as a clause in a contract between publisher and reader, and ask the forensic question: who does this sentence protect?
It protects the platform. A reading-time estimate is a click-through optimization. It reduces the perceived cost of a click, which increases the number of clicks, which improves the metrics that platforms report to advertisers and investors. The badge is not there to help you read. It is there to help you start.
It protects the publisher’s brand. A badge signals that the publication respects your time, which is a flattering thing to signal. It positions the article as a considered object rather than an endless scroll. The badge is a piece of tone, not a piece of information.
It protects the reader’s time anxiety, at least superficially. The badge offers a way to manage the feeling that there is too much to read and not enough time. But it manages that feeling by converting reading into a budgeting exercise, which is not the same as reducing the anxiety. It may deepen it: now every article has a price, and you are always spending.
What does it cost the reader? The badge costs you the possibility of reading without a clock. It frames the encounter as a transaction before it begins. It invites you to skip anything that looks expensive, which means it invites you to skip anything that looks difficult, which means it invites you to skip the things most likely to change your mind. The badge is a small machine for converting curiosity into triage.
The badge in the wild
Consider three artifacts.
The news analysis. A 2,000-word piece on a complicated policy change carries a 9 min read badge. The badge is accurate in the sense that 2,000 divided by 220 is about 9. But the piece is dense with names, dates, and competing interpretations. A reader who actually wants to understand it will spend far longer than 9 minutes, and a reader who spends exactly 9 minutes will have skimmed. The badge has told the truth about the text and a lie about the reading.
The personal essay. A 1,200-word essay carries a 5 min read badge. The essay is about grief. The badge is technically correct and tonally absurd. It is the equivalent of printing a calorie count on a condolence card.
The terms of service. A 12,000-word contract carries no badge at all. This is the most honest thing about it. The publisher knows that no one will read it, and the absence of a badge is an admission that the reading-time estimate is a courtesy extended only to content that is meant to be consumed, not enforced.
What the badge could be instead
The badge is not inevitable. It is a design choice, and design choices can be unmade or replaced. Three alternatives are worth considering, not because they are obviously better, but because they clarify what the badge is doing by contrast.
Difficulty level. Instead of 5 min read, a badge could say introductory, technical, or assumes familiarity with X. This would tell the reader something about the experience of reading rather than the length of the text. It would be harder to compute and harder to fake, which is probably why it is rare.
Reading purpose. A badge could say skim for the argument or read in full for the evidence. This would acknowledge that different readers want different things from the same words. It would require the writer to have a theory of who the reader is, which is a higher bar than counting words.
No badge at all. The absence of a badge would return the decision to the reader, who would have to judge the article by its title, its opening, and its reputation. This is how reading worked for most of its history. It is not obviously worse.
What the badge reveals
The reading-time badge is a small artifact, but it is a legible one. It reveals a set of assumptions that have become so common they are rarely stated: that attention is a resource to be allocated, that reading is a service to be consumed, that the value of a piece of writing can be expressed in minutes, and that the reader is a rational agent who wants to know the price before buying.
Some of these assumptions are defensible. It is not unreasonable to want to know how long something will take. It is not unreasonable for a publisher to respect a reader’s time. The problem is not the impulse. The problem is the reduction. The badge takes a rich, variable, context-dependent act — reading — and flattens it into a single number that describes none of the things that matter.
The badge is not a lie. It is a category error. It measures the text and calls it the reading. It prices the article and calls it the value. It converts an encounter with ideas into a transaction, and then it invites you to complete the transaction before you know what you are buying.
You can still read the article. The badge does not stop you. But it has already told you what kind of thing you are doing, and that framing is hard to unsee. The next time you see 5 min read, try to notice what you are actually doing in those five minutes. It is probably not what the badge said.
FAQ
Where did the reading-time badge come from?
The badge became widely visible in the 2010s as part of a broader shift toward quantified, scannable web content. Medium is often credited with popularizing the format, though the underlying arithmetic — word count divided by a words-per-minute constant — predates any single platform. The specific origin story is difficult to pin down with the sources available here, and I would rather say that than invent one.
How is reading time calculated?
Almost always by dividing the word count by a constant. The constant varies. Nielsen has used 200 words per minute as a general assumption and 250 words per minute for highly literate readers. Many platforms use a figure in the 200–265 range. The calculation is simple; the interpretation is not.
Is the badge accurate?
It is accurate as a measure of text length divided by a constant. It is not accurate as a prediction of how long you will spend reading, because reading speed varies enormously by reader, by purpose, and by difficulty. Nielsen’s 2008 analysis found that users read about 20% of the text on an average page, which suggests that the badge describes a kind of reading most people do not do.
Should publishers remove reading-time badges?
That is a judgment call, not a finding. The badge has real benefits: it signals that an article is finite, and it can help readers decide whether to start. The argument here is not that the badge is worthless. It is that the badge is doing more than it says — it is framing reading as a transaction — and that publishers who use it should be honest about what it is and what it leaves out.
What would a better badge look like?
It would say something about difficulty, purpose, or the kind of attention the piece requires. It would acknowledge that readers are not uniform. It would be harder to compute and harder to fake. Whether any of that is practical at scale is a separate question, and one I do not have a clean answer to.