QWERTY vs Dvorak vs Colemak: Is Switching Worth It?

QWERTY was finalised in the 1870s for a machine with metal typebars that jammed when adjacent levers fired in quick succession. That constraint has not existed for over a century, and the layout survived anyway. This is the standard opening for arguments in favour of switching, and it is true as far as it goes. The question is whether the historical accident costs you anything measurable today.

What the layouts actually do differently

Dvorak, patented in 1936, places all five vowels and the most common consonants on the home row. On typical English text, roughly 70% of keystrokes fall on the home row, against about 32% for QWERTY. It also alternates hands more aggressively, so consecutive letters tend to land on opposite hands rather than forcing one hand through several reaches in a row.

Colemak, released in 2006, is more conservative. It moves 17 keys from their QWERTY positions and leaves the rest alone. Home row usage lands around 74% — slightly better than Dvorak on most corpora — but the design priority was different. Colemak keeps Z, X, C and V in place, which means copy, cut, paste and undo stay exactly where your left hand already expects them.

The finger-travel numbers are genuine. Across a large sample of English text, QWERTY typists move their fingers meaningfully further than Dvorak or Colemak typists do for the same output. Depending on the corpus and the model used, the reduction is somewhere in the range of 30% to 50%. Nobody seriously disputes this part.

The evidence problem

What is disputed is whether reduced finger travel produces faster typing in practice.

The foundational studies claiming large Dvorak advantages were conducted by August Dvorak himself, or by researchers working directly with him, and they carry well-documented methodological weaknesses: small samples, no proper control groups, training conditions that differed between layouts, and outcome measures chosen after the fact. A 1956 US General Services Administration study found no meaningful advantage. Later analyses, most prominently by economists Stan Liebowitz and Stephen Margolis in 1990, argued that the case had been substantially oversold and that independent replications simply were not there.

That does not mean Dvorak is worthless. It means the honest summary is “plausible mechanism, weak evidence for a large speed effect.” Modern enthusiasts often report gains, but those reports come from people who chose to switch, invested weeks of deliberate practice, and had every reason to want the effort to have been worthwhile. That is close to the worst possible sampling conditions for drawing a conclusion.

The comfort claim is on firmer footing than the speed claim. Less finger travel and better hand alternation are reasonable things to want, even without a stopwatch benefit.

What switching actually costs

Expect two to four weeks during which your typing speed sits somewhere between 20% and 50% of your old rate. It is not a gentle dip. For the first several days you will be looking at the keyboard, hunting, and feeling roughly as competent as you did aged eleven. If you type for a living, this is a real productivity cost measured in working days.

Full recovery to your previous QWERTY speed typically takes one to three months of consistent daily use. Getting meaningfully past your old speed, if it happens at all, takes longer.

Then there is the problem nobody solves cleanly: other people’s computers. Shared workstations, a colleague’s laptop, a conference room machine, an interview terminal. Some switchers maintain both layouts and report that the two coexist fine after a few months; others find that keeping QWERTY competence permanently drags on the new layout. Which camp you land in is not predictable in advance.

Software support is mostly solved — Windows, macOS and Linux all ship Dvorak and Colemak natively — but games with hardcoded WASD bindings, remote desktop sessions, and recovery environments will hand you QWERTY without asking.

Who genuinely benefits

Three groups have a real case.

People with existing repetitive strain problems where finger travel and awkward reaches are contributing factors. If an occupational health assessment has flagged your hand movement patterns, reducing total travel is a sensible intervention alongside the ergonomic and workload changes that matter more.

People who type enormous daily volumes — transcriptionists, some writers, some developers — where a small per-keystroke comfort improvement compounds across millions of keystrokes a year.

People who genuinely enjoy the process. If relearning a layout is interesting to you rather than a chore, the switching cost is not really a cost.

For everyone else, the gain is small, uncertain, and expensive to obtain.

Colemak as the pragmatic middle

If you have decided to switch, Colemak is usually the better bet. It changes fewer keys, so the relearning burden is lower. It preserves the ZXCV shortcut cluster, which removes an entire category of daily friction. Its home row statistics are marginally better on English text. And Colemak-DH, a community variant, further adjusts two awkward index-finger reaches the original inherited. Dvorak’s main remaining advantage is age: it is more likely to be preinstalled on an unfamiliar machine.

The uncomfortable conclusion

For the vast majority of typists, layout is not the limiting factor. Technique is. Most people typing at 45 WPM are not held back by QWERTY’s key arrangement — they are looking at their hands, using inconsistent fingering, tensing their shoulders, and correcting errors constantly. Fixing any one of those produces a larger and faster improvement than a layout switch, at a fraction of the cost.

The test is simple. If you have solid touch-typing technique, accuracy above 97%, and you have plateaued north of 80 WPM while typing several hours a day, a layout switch is a defensible experiment. Below that, the ceiling you are hitting is not the letters’ positions.

Measure where you actually stand with a typing test before deciding, and if the answer is that technique needs work, structured typing lessons will move the number further than a new layout would. Developers who suspect their bottleneck is punctuation rather than letters — a common and misdiagnosed case — should try programming typing practice first, since no alphabetic layout addresses that at all.