Info-Dumping Is a Timing Problem
Nobody has ever put a book down and said there was too much information in it.
They say they got bored. They say they lost the thread. They say they started skimming and then realized they'd skimmed four pages and hadn't missed anything. Then a writer hears that, goes looking, finds a long paragraph about the history of the mining guild, and concludes the paragraph was too long.
Maybe. (Sometimes it genuinely is too long.) But that diagnosis sends you off cutting... and cutting is not reliably the fix here. Now and then it makes things worse.
What the standard advice gets right
Search this and you'll find the same four instructions everywhere, from people who know what they're talking about. Spread it out. Cut what the reader doesn't need. Deliver it through scene and action instead of explanation. Don't have two characters tell each other things they both already know.
All four are correct. I'd follow all four. But notice they're all answers to HOW you deliver information, and not one of them answers WHEN. They're method. Underneath the method there's a mechanism, and the mechanism is where the actual failure lives.
Information needs a question already waiting for it
Here's the whole thing in one line: a reader takes in information in proportion to how much they already wanted it.
Not how clearly it's written. Not how gracefully it's folded into action. Whether, at the moment it arrives, there was a question open in their head that this information answers. Give somebody an answer to a question they've been carrying and they'll absorb it without noticing they're being told anything. Give the same person the same paragraph before the question exists and it lands as material to get through.
That's why the volume framing misleads. A paragraph can be too long. A single sentence can also be an info dump, if nothing in the story has made anybody care. And two pages of history can read as fast as anything in the book, provided the reader has spent three chapters wanting to know.
Robert McKee's version, in Story, is that exposition should be released when the audience needs and wants it, and that the best of it should be saved for last. He's writing about screenplays, so take the mechanics with a grain of salt (a screenwriter gets to assume a two-hour runtime, which changes the arithmetic more than you'd expect), but the principle travels intact. Gardner comes at the same thing from the other side in The Art of Fiction: fiction has to keep a vivid and continuous dream running in the reader's head, and needless explanation is one of the things that breaks it. Neither of them is offering you a revision procedure. They're both describing the same underlying fact from different angles.
Why you do it anyway
Because you're frightened the reader will be lost, and that fear is not stupid. It's just badly calibrated, and it's badly calibrated in a way you can't fix by trying harder.
You know everything. You know what the guild is, why it matters in chapter nineteen, and what the reader will need to have understood by then. That knowledge makes you the worst available judge of what somebody needs on page 40, because you can't unknow it to check.
So the failure has a predictable trigger. A new character walks on, you feel the reader's ignorance as a kind of danger, and you hand over the history right there, before anything in the scene has made anybody want it. It feels responsible. It reads as an interruption.
"As you know, Bob" is the symptom, not the disease
The trope is real and it's worth being able to spot: two characters explaining to each other something they both know, so the reader can listen in.
But look at what's actually happening. The information had nowhere to land, and rather than build somewhere for it to land, the draft borrowed a character as a delivery van. Fix only the dialogue and you've changed the vehicle. The information still shows up before anybody wanted it, now dressed as description or a flashback instead.
Worth noting that Browne and King had this in a chapter title back in 1993. Chapter two of Self-Editing for Fiction Writers is "Characterization and Exposition," and the subtitle is "How much and when to fill your readers in." How much AND when. The timing half has been sitting there for thirty years while the craft internet mostly argued about the volume half.
The question-before-answer test
Run it on a finished draft.
- Go through and mark every passage that hands the reader information: world, history, rules, prior events, how the technology works, who this person used to be.
- For each one, find the exact sentence where the information starts.
- Read backward from there. Find the most recent moment where the story gave the reader a reason to want this. An event that raised it. A gap somebody's behavior opened. A tension that only makes sense once you know it.
- If that moment exists and it's close enough that a reader would still be holding it, the passage is placed. If you look backward and there's nothing, you've got an info dump. Length is IRRELEVANT to this step.
- Then the part nobody expects. DON'T cut it.
The fix is almost never deletion
That's the move everybody reaches for, and it's why info dumps get shorter across drafts without ever getting better. You trim the guild history from six paragraphs to two, and the two paragraphs are still arriving in a vacuum.
What actually works is going upstream and building the thing that was missing. Find or make the moment that opens the question, put it before the passage, and then read the passage again. Usually you'll find it needs less trimming than you thought, because half of what felt like bloat was just the reader having no reason to be there. (Bloat and boredom feel identical from the inside of a draft.)
Sometimes the question-opening moment already exists and it's in the wrong place, twenty pages downstream of the answer. Moving it is a ten-minute fix that reads like a rewrite.
Where FirstReader lands on this
It can't run this test, and I want to be straight about why. Knowing what question your reader is holding on page 40 means knowing your book, and that's you.
What it does have is the checks next door. There's a named "As You Know, Bob" check in the dialogue analysis, looking for characters explaining things to people who already know them, and it's specifically told not to flag the legitimate cases, like recapping for somebody who genuinely was absent. The pacing analysis flags significant beats and revelations that got summarized in narration instead of played out.
There's also a guard I'm oddly fond of. Any exposition or info-dump finding gets measured against the passage it cites before it's allowed to survive, using average sentence length and how much of the passage is dialogue. If the prose is brisk or it's mostly people talking, the finding gets dropped as not exposition-like. It exists because a tool that flags every long paragraph as an info dump is worse than no tool, and I'd rather FirstReader miss one than cry wolf at your good pages.
So it gives you candidates. The backward look in step three is yours.
The fear is pointed the wrong way
The thing writers are scared of is the reader being confused, and I think that fear is aimed at the wrong target.
A reader will carry an open question a remarkably long way, as long as something is moving. Not knowing why she won't go back to the house is a reason to keep reading. It's the engine... and writers keep treating it as a leak and plugging it.
What you should actually be frightened of is indifference, and indifference doesn't arrive because a question stayed open too long. It arrives because no question ever got opened. Every time you answer something early, you're spending the fuel before you've built the engine.
So the revision question stops being "is this too much?" which you can argue about forever, and becomes "had I given them any reason to want this?" That one has an answer, and it's in the ten pages before the passage.
If you enjoyed this, please leave a comment below to let me know. If you DIDN'T enjoy it, well, I'd like to hear from you too!