How to type faster: a practical plan for real work

Where the time actually goes

A correction adds more than the wrong character. You can stop, press backspace, type the letters again, and then continue the sentence. A words-per-minute total combines all of that into one result, but it does not tell you which part repeated.

This guide is for ordinary work writing: emails, chat messages, notes, reports, customer replies, and project updates. It uses observable typing signals to choose a practice target. It cannot identify a medical, ergonomic, or personal cause.

A correction can have five observable stages.

1. 01 Wrong key One key press
2. 02 Stop You notice and interrupt the flow
3. 03 Backspace One or several presses
4. 04 Retype The same letters again
5. 05 Recover You find the sentence and your rhythm again

Backspace and retyping add key presses. Stopping and resuming add time. The full result tells you how fast the passage was completed, while the event sequence shows where the correction happened.

A study of 168,000 volunteers measured correction behaviour in an online transcription task.¹

2.29

Backspace or Delete presses per sentence, on average

6.3%

of key presses were Backspace or Delete

1.17

key presses per transcribed character

Those volunteers were copying set sentences, so the figures belong to that task rather than to your inbox. They show that correction behaviour was present and measurable in this study.

The headline number is useful for output rate, but not for identifying a practice target. Words per minute tells you what you produced, not where the time went. Accuracy tells you how many key presses were right, but not whether the wrong ones clustered somewhere. The list of mistakes names the keys and pairs that actually failed. The correction count reports the correction events recorded during the test. The pace curve shows whether your output rate stayed steady across the passage.

Even then, one test is one passage under one set of conditions. Take the typing test at a pace you can control, then take a few comparable tests on the same keyboard and layout and look for what repeats. If you want the detail of how a test decides that a key or pair is hard, that is in the methodology.

A lower score can contain different signals

The same low score can contain different observable signals. Each signal suggests a different next check, so choose from your own results instead of assuming one universal cause.

If your results have sat still for weeks rather than days, that is a different question again, and the plateau guide separates one variable score from a repeated pattern.

Set up the basics once

Keep the setup consistent so later test comparisons are easier to interpret.

In a standard QWERTY touch-typing example, rest your left fingers on A then S then D then F A +S +D +F and your right fingers on J then K then L J +K +L with the semicolon key. The raised marks on F and J can help you find that position without looking. From there, touch typing assigns each finger a group of keys and returns the fingers to the home row. This is one standard technique, not a rule that PeakTock can prove is best for you. Research from Aalto University also documents effective self-taught techniques.⁴

Choose a chair, desk, and keyboard position that feels comfortable and lets you reach the keys without strain. This guide does not prescribe a posture. Take breaks that fit your work. If typing causes continuing pain, numbness, or tingling, stop and speak to a qualified health professional.

Use a pace that you can repeat across the sentence. If the timing on your target becomes uneven or the correction count rises, slow down. Press and release each key with only the force needed for your keyboard to register it.

Keep one goal at a time: one pattern or one habit for each comparison. Changing one variable makes the before-and-after result easier to interpret.

Practise one pattern, not the alphabet

When the same key or pair repeats across comparable results and passes PeakTock's evidence gate, treat it as a candidate practice target. One occurrence is not enough.

A pair can require a movement that a single-key score does not show. Take E then D: standard QWERTY touch typing gives both keys to the left middle finger, so one finger has to leave the top row and come back to the home row before the next letter can be typed. A pair typed with alternating hands uses a different movement. This is the kind of movement to look for in your own result.

Whatever pair you end up with, practise it inside words. For ED, examples include needed, updated, confirmed, and attached. Put them into a sentence you might send at work, and use a pace at which you can repeat the pair accurately. Speed is not the goal of this step; a repeatable movement is.

If your goal is transfer to work, also use phrases you personally repeat: a customer name, a product name, your email sign-off, or a project code.

Find your own target

The test names a pair only when it has seen enough of your typing to justify one, and it offers the exact drill for whichever pair that is. If ED turns out to be yours, the full ED guide covers the movement in detail. To practise without disturbing important work, follow the work, practise, and review loop.

Shortcuts handle editing and navigation outside typing-test speed

Shortcuts do not change how fast you enter letters in a typing test. They act on the editing, selection, and navigation around those letters. The exact keys can vary by application, so use this table as a starting point and check the menus in your own software.

Take one a week, starting with the action you already repeat most. Some applications reassign these keys. The table follows current Apple and Microsoft shortcut references.⁵ Many Linux applications use similar Ctrl shortcuts, but check the application menu.

What four weeks of this looks like

A structure, if you want one. Sessions are five to ten minutes, away from anything with a deadline, and a week that does not fit around your work is better repeated than skipped. This is an untested planning template, not a program derived from the studies below.

Spreading the sessions across days borrows a general learning principle, not a tested typing dose.³

Week one is measurement. Take a controlled test on days one, two, and three, writing down the WPM, accuracy, and correction count each time, and keeping the keyboard, layout, and test length the same. Three results give you a small starting range that one result cannot. On day four, look for a key or pair that repeats and pick one candidate target. Practise that target slowly on days four and five, first in real words and then inside a sentence you would actually send. Day six you type normally at work and do not think about it at all. Day seven, retest and look only at the target.

Week two holds the pace. Use the same target. Find a pace at which corrections stay low and hold it rather than pushing past it. Add one shortcut from the table above.

Week three checks technique. If looking down was one of your observed signals, practise a familiar short passage while keeping your eyes on the screen. Otherwise, continue the same target. Add a second shortcut if it helps your normal work.

Week four checks the transfer. Take a controlled test and compare it with day one. Then either pick the next single target or run week one again with a new one.

One awkward session is not enough to judge the plan. Learning research separates how well you perform while practising from what you retain afterwards, so do not judge the plan from how a single practice session felt.² The return to normal work is the transfer check.

Common questions

Sources and their limits

These support the structure of the guide. None of them proves a typing method, prescribes a schedule, or predicts a result for an individual.

1. Dhakal, V., Feit, A. M., Kristensson, P. O., and Oulasvirta, A. (2018). Observations on Typing from 136 Million Keystrokes. CHI 2018. DOI. An online study of 168,000 self-selected volunteers copying set sentences. It reports an average of 51.56 words per minute, 2.29 Backspace or Delete presses per sentence, and 1.173 key presses per transcribed character. Backspace or Delete key presses averaged 6.3 per cent of all key presses. It also reports that words per minute correlates negatively with error corrections (r = −0.36) and with key presses per character (r = −0.4), and positively with the self-reported number of fingers used (r = 0.38). These are correlations across a group of volunteers. They are not causes, not norms for office work, and not predictions for you. The task allowed other editing actions, so these key counts do not capture every possible edit.
2. Soderstrom, N. C., and Bjork, R. A. (2015). Learning Versus Performance: An Integrative Review. Perspectives on Psychological Science, 10(2), 176–199. DOI. A broad learning review that separates temporary practice performance from later learning. It is not touch-typing research.
3. Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., and Willingham, D. T. (2013). Improving Students’ Learning With Effective Learning Techniques. Psychological Science in the Public Interest, 14(1), 4–58. DOI. This review finds distributed practice useful across many learning conditions. It is not a typing trial.
4. Feit, A. M., Weir, D., and Oulasvirta, A. (2016). How We Type: Movement Strategies and Performance in Everyday Typing. Aalto University research page. The page compares trained touch typists with diverse self-taught techniques. It does not show that touch typing is necessary or best for every person.
5. Apple Support. Mac keyboard shortcuts. Official reference, and Microsoft Support. Keyboard shortcuts in Windows. Official reference, plus Keyboard shortcuts in Word for word selection, reviewed August 18, 2026. The sources note that some shortcuts vary by app.

PeakTock’s own results come from the key presses it measures in your test or, for the Mac app, in your everyday typing, and they are subject to its evidence gate. It measures typing, not people: it can show you what repeated, but not why, and it is not a medical or ergonomic assessment.