Put the point where the reader is looking

Read this paragraph. Don't skip it, and don't read ahead. Give it one honest pass, at normal speed.

Deal with everything in the space before you start, because afterwards it is far harder. Take out whatever will come out, and protect whatever will not. Attend to any irregularities early: they will be much more obvious at the end than they are now, and by then very little can be done about them. Always work downward. Going the other way means repeating what you have already finished. Use less material than feels right, since two light passes give a better result than one heavy one, which gathers and runs. Wait properly between passes, even though the waiting is the hardest part of it. When everything has settled, put back what you moved and clean what you used straight away, because a day later that is a much bigger job.

You understood every word. You couldn't tell anyone what it was about.

Every word in that paragraph was ordinary English. Every sentence was short. It scores 69 on the Flesch reading-ease scale, which is comfortably plain English, and it was still unreadable. That's the narrow thing a formula can't see: it counts syllables and sentence lengths, and the problem here was in neither.

Context first, then the thing it contextualises

Gopen and Swan state this as the sixth of their seven structural principles in "The Science of Scientific Writing": "In general, provide context for your reader before asking that reader to consider anything new."

Keep the "In general". It's theirs, and they meant it: the same paper says "None of these reader-expectation principles should be considered 'rules.'" We come back to that at the end of this lesson, and lesson 8 is about it.

That sentence is doing more work than it looks. It isn't advice about being organised. It's a claim about what a reader is physically doing while they read you.

The mechanism is this. A reader who doesn't yet know where a passage is going can't file anything. Every fact arrives and has to be held, unsorted, in case it turns out to matter. "Attend to any irregularities early" is a burden until the moment you know it's a room being painted, at which point it costs nothing at all, because it drops into a slot that was already there.

So the cost of withholding context isn't that the reader is mildly annoyed. It's that you've made them carry your whole document at once. And they won't do it for long. Lesson 1's finding was that a reader takes in about a fifth of an average page; the sharper version of the same measurement is that readers get through half the information only on pages of about 111 words or fewer. Spend those first hundred words giving them nothing to hold on to and you've spent the part you could count on.

Front-loading at three levels

Front-loading is the practical form of principle 6, and it applies at three scales. It's the same move each time, for the same reason.

The document. The reader should be able to state what this is and what it wants from them inside the first hundred words or so. In an email that means the request or the decision in the first sentence or two. In a report it means an opening that says what was found, not what was done.

The section. Each section should say early what it's for, and that's mostly a job for headings. GOV.UK's guidance on structure asks for headings that are "descriptive" and "frontloaded", tells writers to "avoid generic headings like 'Introduction'", and gives the same instruction for the body: "Put the most important information first. The quicker you get to the point, the greater the chance your users will see the information you want them to." Lesson 7 takes headings properly. The principle is here: a section that opens with three sentences of preamble has recreated the whole problem in miniature.

The paragraph. A paragraph should usually declare its business in its first sentence. Not because of a rule about topic sentences, but because a reader who doesn't know what a paragraph is about is doing the painting-a-room paragraph again, at small scale, forty times a document.

The pattern to notice is that the reason never changes. At every level, the question is whether the reader can file what comes next.

The email with five clauses of throat-clearing

This is the kind of thing that arrives every day.

Following our discussion last week and having reviewed the figures from Q2 as well as the projections that finance circulated on Monday, and taking into account the feedback from the regional teams who raised a number of concerns about capacity during the summer period, I wanted to reach out regarding the launch date and share some thoughts about how we might want to proceed given everything we now know.

As you will remember, the regional teams have been flagging for some time that the summer months are difficult for them from a staffing point of view, and this year appears to be no different, with several of them noting that they expect to be running below their usual cover from late June through to the end of August.

The Q2 figures, read alongside that, do seem to support what the regional teams are saying, in the sense that the support volumes we saw in Q2 were already at the upper end of what we had planned for. The projections finance circulated on Monday, while they do not really change the picture in either direction, at least do not contradict it.

So on balance, and taking everything above together, I think the sensible course of action is probably for us to move the launch from May to July, unless there is something I have missed. Let me know what you think when you get a chance.

Two hundred and thirty-five words, and the decision arrives at word 211. A reader who gives this email the attention they give an average page has stopped somewhere in the second paragraph, and everything they read was preamble. They now know there was a discussion last week, that finance sent something, and that the regional teams have concerns. They don't know that anyone is proposing to move anything.

Before you read the rewrite, do the first step yourself.

Now the same email, front-loaded:

We should move the launch from May to July. Here's why.

The regional teams expect to run below their usual cover from late June to the end of August, and the Q2 figures back them up: support volumes were already at the top of what we planned for, so we'd be launching into our thinnest cover of the year. Finance's Monday projections don't change the case either way.

If you want to keep May, the thing to solve is summer cover. Let me know by Thursday and I'll put options together.

Ninety-one words, and the decision is the first thing in it: it begins at word three and is complete by word nine. The whole email now fits inside the hundred words the original spent clearing its throat.

The wrinkle: when the point should not come first

If this lesson stopped here you'd have a rule, and Gopen and Swan's own caveat is that none of this should be taken as one. Front-loading is a default. There are cases where it's wrong, and they aren't rare.

Bad news that needs its reason to be understood. Take a letter refusing a benefit claim. Buried, it looks like this:

Thank you for your application of 4 March and for the documents you supplied. We have now completed our review of your circumstances, taking into account the information you provided about your household and your recent change of employment. Having considered all of this carefully, and applying the rules currently in force, we have concluded that your claim does not meet the qualifying conditions and has therefore been refused.

Front-loaded without its reason, it looks like this:

Your claim has been refused. You can appeal within one month.

That is honest, and it's the sentence after which many readers stop reading and start reacting, which means they may never reach the part about appealing. The fix isn't to go back to burying it. It's to put the one reason that makes the decision intelligible next to the decision itself:

Your claim has been refused, because your savings are above the £16,000 limit. If that figure is wrong, you can appeal within one month and we will look again.

One sentence of reason, not four paragraphs of context. The reader understands rather than only resents, and the deadline is still in the first thirty words.

An argument whose conclusion would be rejected on sight. If you're writing to someone who currently believes the opposite of your conclusion, opening with the conclusion invites them to spend the rest of the document arguing with you instead of reading it. Here the premises genuinely do come first, because the reader has to agree to each step before they'll accept where the steps lead. Notice this is still principle 6: you're providing the context that makes the conclusion intelligible before asking the reader to consider it.

Narrative, where the withholding is the point. A joke front-loaded is not a joke. Neither is a mystery. This course is about writing that has to be understood, not writing that has to be enjoyed, and the two have different rules. Worth knowing which one you're doing.

What people get wrong

"The inverted pyramid is always right." Journalists put the most important information first, then detail, then background, and it works because a news reader may stop at any point. GOV.UK recommends the same shape for the same reason. It's one good instance of front-loading, not the definition of it. A document whose reader will finish it has different needs.

"Front-loading means being blunt." The two are unrelated. "We should move the launch from May to July, and here's why" is front-loaded and perfectly courteous. Bluntness is about what you say of people; front-loading is about what order the reader meets things in.

"A summary at the end does the same job." It doesn't, and the reason is in this lesson's first paragraph. A summary at the end helps a reader who finished. Context at the start helps a reader read. The passage at the top of this lesson, with its title printed underneath instead of above, is still unreadable.

"Front-loading means putting the conclusion first." Sometimes, and sometimes not. The principle underneath is context before the new thing. In an argument, the conclusion is the new thing and the premises are the context, which is why an unwelcome conclusion belongs after the reasons rather than before them.

"My document has an executive summary, so it's front-loaded." Only if the summary says what was found and what should happen. A great many executive summaries say what the document covers, which is a table of contents in prose and gives the reader nothing to file.

Practice

Where this goes next

You now know where the point should go. You don't yet know how to write the sentence that carries it, and that's the harder problem: an email can be perfectly front-loaded and still be exhausting to read, because the trouble is inside the sentences rather than in their order.

The next lesson goes down a level, to the single move that does more for a sentence than any other. Then lesson 5 comes back to arrangement, at the scale of the paragraph, and explains something this lesson has only asserted: why a passage can feel disconnected even when every sentence is in a sensible order.

Go deeper

• Gopen and Swan in full. It's about scientific prose, but six of its seven principles apply to any writing, and the closing section on why they are not rules is the best short thing written on the subject.
• GOV.UK on structure, which is front-loading written up as an operating standard for a whole government, with its heading rules and its own reading figures.
• Digital.gov's plain language guides, the American equivalent, which reach the same conclusions from a different drafting tradition.
• Bransford and Johnson's 1972 experiment, which did formally what the passage at the top of this lesson does to you. See the Sources note below on why this course links no copy of it.

Sources

• George D. Gopen and Judith A. Swan, "The Science of Scientific Writing", American Scientist 78 (1990). Principle 6, quoted in full with its "In general", is the foundation of this lesson, and the paper's own statement that "None of these reader-expectation principles should be considered 'rules.'" is quoted in the wrinkle section.
• GOV.UK, "Create a clear structure for your content". "Put the most important information first", the inverted pyramid, and the requirement that headings be descriptive and front-loaded rather than generic.
• The reading figures are carried forward from lesson 1, where they come from Jakob Nielsen's "How Little Do Users Read?" with his caveats: about a fifth of an average page, and half the information read only on pages of 111 words or fewer.
• The Flesch reading-ease score of 69 for the untitled passage was computed for this lesson.
• The three exceptions in the wrinkle section are reasoning from principle 6, not findings. The callout in that section says so.
• The untitled passage at the top of this lesson was written for this course, so that you could feel the effect on yourself rather than read about it. The demonstration is not original: a well-known 1972 experiment by Bransford and Johnson did the same thing formally with a passage about washing clothes, giving different groups its title before, after, or not at all, as secondary summaries describe it. The only freely available copies of that paper are scanned images, so this course has not read it, uses none of its text, and quotes no figure from it.