Why there is no number

To answer this question honestly, somebody would have to run a study that does not exist. It would need a defined starting population — people who currently type how, at what speed? — a defined criterion for what counts as having learned, and long enough follow-up to see people cross it. Nobody has published that study, so every specific figure in circulation was produced some other way.

You can watch this happen if you look at the numbers side by side. One site says twenty hours, another says two weeks, another says ten minutes a day for a month, another says three months. None of them names a dataset, none of them defines what "learned" means, and none of them says who was measured. They are not competing estimates from different studies. They are the same non-answer written in different fonts.

That does not make the question useless. It makes it the wrong shape. Broken into pieces it becomes answerable, and the pieces are worth more than a headline number would have been.

Decide what "learned" means first

There are three thresholds hiding inside the question, and they arrive in this order:

  1. You can find every letter by feel. You can place your hands by locating the F and J bumps, and you know where each letter is without checking. This one is genuinely quick — it is a memorisation task with 26 items and a strong spatial structure.
  2. You can type continuous prose without looking down. Any speed at all, at an accuracy you would accept in real writing. This is the threshold most people mean, and it is much further away than the first one because it requires the movements to happen without deliberate thought.
  3. You are at least as fast without looking as you used to be while looking. For someone who was already a competent hunt-and-peck typist, this is the expensive threshold, and it is the one that determines whether the switch felt worth it.

Most of the total effort sits in the gaps between these, not inside any one of them. Someone who reports learning touch typing in a weekend has almost certainly crossed the first threshold. Someone who says it took months has usually been talking about the third.

What the answer actually depends on

Four things dominate, and only one of them is under your direct control.

  • How much you type anyway. This is the largest factor by a wide margin — but only if the typing you already do is done with the new technique. Two hours of email a day is enormous practice volume. Two hours of email typed the old way, plus ten minutes of drills, is ten minutes of practice a day and two hours of rehearsing the habit you are trying to replace.
  • Whether you switch in real work or only in practice. Switching in real work is the accelerator and the cost at the same time: your actual output drops for a while, in front of colleagues and deadlines. Practising only in drills protects your output and stretches the timeline, possibly indefinitely.
  • How fast you already are. Counter-intuitively, a fast existing typist has the harder road. The first two thresholds are no worse for them, but the third is much further away, because they have more old speed to re-earn.
  • How much accuracy you are willing to lose on the way. Pushing pace before the movements are automatic keeps you in the uncomfortable stage longer, because the errors you make while guessing are themselves practice.

Expect to get slower first

The dip is not a sign that something has gone wrong, and there is one well-documented data point about how deep a comparable dip can be. In 1956, Earle Strong ran a controlled study for the US General Services Administration in which ten government typists were retrained onto the Dvorak keyboard layout. They needed well over 25 days of four-hour training merely to get back to the speed they had already had on QWERTY, and Strong concluded the retraining cost could not be recouped. The study is reported in detail by Liebowitz and Margolis in Reason.

Read that as a bound rather than a forecast. Strong measured a full layout change — every letter in a new physical location — which is a strictly larger relearning problem than keeping the same layout and changing which finger presses each key. Your dip should be shallower than theirs. What the study establishes is the shape: skilled typists who change how they type get slower first, by a lot, for longer than they expect, and they recover.

The plateau, and why it happens

Almost everyone stalls somewhere. It helps to see what a plateau looks like in units of actual movement, because the intuitive picture — that you need to press keys much faster — is wrong.

Words per minute converts cleanly into milliseconds per keystroke. At 60 WPM you are producing 300 characters a minute, which is one character every 200 milliseconds. At 70 WPM you are producing 350 characters a minute, or one every 171 milliseconds. Ten words per minute is therefore worth about 29 milliseconds per keystroke. At the top end the same arithmetic gives 120 WPM an average interval of 100 milliseconds.

Nobody makes a 29-millisecond improvement by pressing keys harder. Intervals that short are not really "how fast your finger moves" at all; they are dominated by whether there was a pause before the movement started. A plateau is usually a pause problem: the letters you know cold already arrive near your physical floor, and the remaining time is sitting in the handful of sequences that still make you think.

How to break a plateau

What follows is reasoning from how the metrics are computed, not measured effect — no published study establishes that any particular drill breaks a typing plateau, and this page is not going to invent one. With that stated plainly, four moves follow logically from the arithmetic above.

  • Slow down until the errors stop, then rebuild. If your raw WPM sits far above your net WPM, the plateau is made of mistakes rather than speed, and pushing pace makes it worse.
  • Watch consistency instead of speed for a while. Keyboardo's consistency score is built from your per-second rhythm, so it measures precisely the pauses that the millisecond arithmetic says your remaining time is hiding in. Speed can sit flat for weeks while consistency climbs, and when that happens the speed usually follows.
  • Change the material, not the effort. Turning on punctuation and numbers, moving to a longer test, or switching to another language all force the sequences you have not automated to show up more often. Practising what you are already good at is comfortable and does nothing.
  • Read the per-word breakdown, not the headline. After every test the results screen lists exactly which words you got wrong and marks your errors on the speed chart. A plateau usually has three or four specific culprits in it, and they are named on that screen.

How to measure your own answer

You cannot get a general answer to this question, but you can get your own, and it is more useful than the general one would have been. The method is unglamorous: fix a measurement condition and never change it.

Pick one — a 60-second English test with punctuation and numbers off is a good default, because it is also the condition the leaderboards use — and run exactly that, once, at the end of each practice session. Everything else you do that day is practice; that one run is the measurement. Changing the mode, the duration or the toggles between measurements makes the series meaningless.

A free account then does the bookkeeping for you. Your profile shows the average WPM of your last ten tests, a chart of your daily average over the past 90 days, a heatmap of which days you actually practised, and your personal best on each of the eight time and word boards. The heatmap and the chart answer different questions — did you turn up, and did it work — and you need both, because a flat curve during a month of blank heatmap squares is not a plateau. Without an account, your most recent 200 results are still kept in your browser, and can be merged in later if you decide to sign up.

To set your baseline, run a 60-second test now and write the number down. For what to do between measurements, see building a typing practice routine; for the technique itself, see the touch typing guide.