How to Read Faster: 7 Techniques That Actually Work (And 3 That Don't)

The honest version is short: five minutes a day for six weeks, with a pacer, a comprehension check, and a slow reduction of subvocalization. Done. You'll see a 30 to 60% gain in sustained reading speed at the same comprehension level.

The dishonest version sells you a weekend course for $499 and promises 1000 WPM. We've written about why those numbers are bullshit in a separate piece. This one is about what actually works, what doesn't, and why the difference matters.

A note first: you can't honestly train against a baseline you don't have. We've watched users start "speed reading" without ever timing themselves on a real passage, then declare success based on the feeling of having read faster. The feeling is unreliable. The clock is not. Measure first.

Take the 90-second baseline test →

The seven that work

Use a pacer. Any pacer.

This is the single biggest unlock. Your finger, the tip of a pen, a moving line on a screen — doesn't matter. Anything that moves across the page at a steady rate forces forward momentum and prevents the small backward eye-jumps that slow most adult readers.

The mechanism is so well-documented it's almost embarrassing. Rayner's eye-tracking research, going back to the 1970s, shows adult readers regress on 10 to 15% of words during normal reading. Most regressions aren't useful — they're an anxious habit dressed as caution. A pacer makes regression impossible because you can't re-read a line you've physically left behind.

If you try only one thing from this article, try this for a week. Run your index finger under the line you're reading, slightly faster than your normal pace. The first two pages feel like you're missing things. By page five your reading speed has risen to match the pacer, and your comprehension has stayed put. The gain typically holds at 20 to 30% — for free, forever.

Quiet the inner voice (don't try to kill it)

You can't eliminate subvocalization. Slowiaczek and Clifton's 1980 EMG study showed motor activity in the lips and tongue during silent reading even in trained subjects who reported no inner voice. The voice is built into how the brain accesses meaning from text. It's called lexical access via the phonological route, and it's a real thing your neurons do.

What you can do is reduce the voice's grip. The bottleneck isn't that the voice exists — it's that the voice runs at speaking speed (~180 WPM) and your eyes are capable of much more. The training goal is to read fast enough that the voice can't keep up, forcing your brain to recognise words visually before they're "spoken."

Three drills that work, in roughly increasing intensity:

Read while chewing gum or humming softly. Occupies the speech-motor circuit, breaks the verbalisation reflex. Try this for a week.

Read at 350 WPM via an RSVP drill for two minutes a day. Above the voice's ceiling. Your brain learns to short-circuit. (RSVP isn't the main meal — see our piece on why it fails as a main technique — but it's a fine warm-up.)

Count "1, 2, 3, 4" silently in the background while reading. Sounds ridiculous, works surprisingly well. The verbal channel is occupied, your visual recognition has nowhere to hide.

Expect 40 to 80 WPM of gain from this alone. Pair it with a pacer and you're already at most of the win.

Chunk in twos and threes

Slower readers fixate on every word. Faster readers fixate on groups. The eye-tracking literature calls this the perceptual span — how many letters of useful information you can extract per fixation. Trained adult readers have spans around 14 to 15 letters. Untrained ones run closer to 9.

The drill: take a printed page, draw two faint vertical lines about a third and two-thirds across, and force your eyes to land only on those two points per line. Three fixations becomes the goal instead of seven or eight. Painful for two days, then automatic.

The gains here are modest — about 15 to 25% — but they stack on top of the pacer gain and the subvocalization gain. The compounding is where the math gets fun.

Stop yourself from regressing

Specifically. Mechanically. Use a card or your hand to physically cover lines you've already read. You can't go back to what isn't visible.

This is the same principle as the pacer, applied retroactively. Most regressions weren't helping comprehension; they were a tic. The discomfort lasts about three sessions, then the habit dies, and your speed jumps another 5 to 10% with no comprehension cost.

A related fix that almost no one teaches: when you hit a word you don't fully understand, don't stop. Keep reading. Most of the time the meaning resolves from context in the next sentence. If it doesn't, then go back. This single rule converts maybe 80% of regressions into context-resolved understanding, which is faster and usually deeper.

Pre-read the structure (the highest-ROI thing on this list)

Spend 30 to 60 seconds looking at headings, the first sentence of each section, any bolded terms, and the final paragraph. Then read.

This is mundane and absurdly effective. Pre-loading the scaffold of a text before reading the body makes the body 20 to 30% faster and improves comprehension. The mechanism is well-understood in cognitive psychology: your brain knows where the argument is going, so it can compress predictable parts and pay attention to the surprising ones.

If you only have time for one technique on this list and you're already a competent reader, do this one. It's free, takes a minute, and works on the very first try.

Read more in your field

Reading speed is gated by word-recognition speed. Familiar words trigger lexical access in roughly 200 milliseconds. Unfamiliar ones can take ten times longer, because your brain has to assemble meaning rather than retrieve it.

The practical implication is that the most reliable way to read faster in any specific field is just to read more in that field. The first 50 hours are slow. The next 50 are noticeably faster, with no technique change at all. By 200 hours you're reading domain papers at the same speed you read general non-fiction, and the gain is entirely from vocabulary fluency.

This isn't a "drill." It's a feature of how speed grows. Just useful to know — because it means your job-related reading will get faster every year you stay at your job, even if you do nothing else.

Practice daily — and stop after six weeks

Five to ten minutes a day, six days a week, for six weeks. That's the dose with the most evidence behind it. Less than that doesn't stick. More than that hits diminishing returns fast.

Two hours of practice on a weekend, then nothing all week, doesn't compound. Asymmetric is the killer. Daily-small beats weekly-large the same way running 2 miles every day produces fitness that running 10 miles on Sundays doesn't.

After six weeks, you don't have to stop — but the heavy lifting is done. Maintenance is one session a week. Most of the lifetime gain is locked in by week eight.

The three that don't

Photo-reading. The claim that you can absorb a page by glancing at it for one or two seconds while your subconscious processes the content. Every controlled study of "photo-reading" has produced comprehension scores at chance level — i.e., the participants did no better than people who never saw the page. It's not a marginally-effective technique; it's not a technique at all. If you spent money on a photo-reading course, you bought certainty, not capability.

Speed-only training with no comprehension check. Most speed-reading apps optimise the wrong number. They show your WPM rising; they never ask what the passage said. This is how people end up "reading at 800 WPM" on the app and gaining zero on real books. Without comprehension as a constraint, your brain learns to disengage at speed — a habit that doesn't transfer to anything you care about. The fix: every session ends with three questions about what you just read. If you can't answer two of them, the speed was fake.

One-time weekend courses with no daily practice. The intensives produce a real boost on day two — and that boost is gone by week four. Reading speed is a habit, not a fact. The brain de-trains the way it de-trains anything you stop doing. The weekend-course business model trades on the temporary boost without disclosing that the gains don't survive a month of normal life.

What six weeks actually looks like

Numbers we see in our own user data, which roughly match the intervention studies:

• Weeks 1–2. Pacer training plus regression elimination. Median user goes from ~240 to ~290 WPM. Comprehension holds.
• Weeks 3–4. Subvocalization reduction starts working. Chunking begins to feel natural. Median climbs to ~340.
• Weeks 5–6. Pre-reading habits compound. Domain vocabulary gets pruned. Median lands around 380 to 420 with comprehension held.

That's a 60–75% gain. Sounds modest next to "1000 WPM!" until you do the math: a 200-page non-fiction book that took 6 hours now takes 3:30. Across a decade of reading, that's hundreds of additional books you'll actually finish. The compounding is the entire point.

What to do this week

Get a baseline. Take the test — 90 seconds, including a comprehension check. Write down the number.

Pick two techniques. Make it the pacer plus subvocalization reduction. Skip the rest for the first two weeks. Trying to do all seven at once is how people quit.

Train five minutes a day for six weeks. Re-test on day 14, day 28, day 42. The number either moves or it doesn't, and if it doesn't, change one variable and try again.

That's it. The whole thing fits on an index card. The reason it doesn't get marketed that way is that index cards don't sell for $499.