A good typing speed for most jobs is whatever the employer's own test requires, and where that requirement is written down, it comes with a test length and an error rule. U.S. federal clerk-typist and office automation positions ask for 40 words per minute (WPM), measured on a five-minute sample with three or fewer errors. A large published study of keyboard typing, with 168,960 volunteers, found an average of 51.56 WPM. So if you type around 50 WPM with few errors, you already clear the written office threshold that both the federal government and California use.
What counts as a good typing speed?
A well-documented reference point is a 2018 study from Aalto University and the University of Cambridge, presented at the CHI conference by Vivek Dhakal, Anna Maria Feit, Per Ola Kristensson and Antti Oulasvirta. Their 136 million keystrokes study asked volunteers to copy 15 English sentences as quickly and accurately as possible. The final dataset covered 168,960 people and 136.8 million keystrokes. The average was 51.56 WPM with a standard deviation of 20.2 WPM, which means typical speeds spread widely around the mean.
The edges of that distribution are more helpful than the average when you want a target. The researchers defined fast typists as the top 10 percent, above roughly 78 WPM, and slow typists as the bottom 10 percent, below roughly 26 WPM. Some participants reached 120 WPM or more. The average uncorrected error rate was 1.17 percent. Participants who reported taking a typing course were only about 5 WPM faster on average than those who had not.
Read in plain terms, around 50 WPM is average for this group, about 78 WPM or more puts you in its fastest tenth, and speeds below about 26 WPM sit in its slowest tenth. Those bands are a fair way to judge your own number when a job posting gives you nothing more specific. They are not, however, a grading scale that employers have agreed on, and the next section explains why the sample matters.
Who was measured
The participants were self-selected users of a commercial typing-test website who chose the study as an alternative to the site's regular test. Their mean age was 24.5, 68 percent were from the United States, and 72 percent said they had taken a typing course. They also reported typing about 3.2 hours a day on average. The authors note that the sample represents people interested in testing and developing their typing skills. Treat 51.56 WPM as the average of that group, typing English, rather than the average of all U.S. workers.
Phone typing is slower
The same research group ran a separate mobile typing study, published at MobileHCI 2019, with 37,370 volunteers. The average on phones was 36.2 WPM with 2.3 percent uncorrected errors, and the researchers describe mobile speeds as about 70 percent of desktop keyboard speeds. If a job will test you on a keyboard, a score from a phone test will likely understate what you can do.
Typing speed requirements by job
Searches for "WPM by job" return many ranges for data entry, customer support or programming, but most pages do not say where those numbers come from. The table below includes only requirements that an employer, government agency or certifying body has published. Each row shows the test conditions, because a WPM figure without them is hard to compare.
| Role | Written requirement | Test conditions | Source |
|---|---|---|---|
| Clerk-Typist, GS-2/4 (federal) | 40 WPM | 5-minute sample, three or fewer errors | OPM group qualification standard |
| Office Automation Clerk/Assistant and "(Typing)" titles, any grade (federal) | 40 WPM | 5-minute sample, three or fewer errors | OPM group qualification standard |
| Data Transcriber (federal) | 20 WPM for GS-2, 25 WPM for GS-3 and GS-4 transcription duties | 5-minute sample, three or fewer errors | OPM group qualification standard |
| Clerk-Stenographer, GS-3/4 (federal) | 40 WPM typing plus 80 WPM dictation | Typing on a 5-minute sample | OPM group qualification standard |
| Office Assistant (Typing), Office Technician (Typing) (California) | Certificate showing at least 40 WPM | 5-minute proctored test, certificate issued within 4 years | CalHR |
| Registered Professional Reporter (national certification) | 180, 200 and 225 wpm | Three 5-minute dictation tests on a steno machine, 95 percent accuracy each | NCRA |
The federal figures come from the clerical section of OPM's group qualification standards. Applicants can show proficiency by passing a performance test, presenting a certificate from an authorized organization, or self-certifying, and agencies may test self-certified applicants to verify. Test results and certificates stay valid for three years. California's typing certification rules from CalHR reject certificates from tests that were unproctored or shorter than five minutes.
The last row needs care. The NCRA's RPR requirements measure stenographic writing from dictation, not typing on a standard keyboard, so 225 wpm is not a keyboard goal. The same applies to OPM's reporting stenographer and shorthand reporter positions, which list dictation speeds of 120 to 175 words per minute with the equipment appropriate to the job.
How these thresholds compare with the average
Placing the 40 WPM federal and California threshold next to the Aalto distribution gives a rough sense of scale. It sits about 11.5 WPM below the study's 51.56 WPM average and well above the roughly 26 WPM line for the slowest 10 percent. Because the tests differ in length, text and error rules, this comparison only gives a rough position, and it would be wrong to turn it into a percentile. What it does show is that these written thresholds are minimums meant to screen for basic competence.
A minimum is not the same as a competitive target. A posting may set 40 WPM as the floor while other candidates arrive with much higher scores, and a hiring manager may weigh accuracy, formatting or domain vocabulary more than raw speed. If you are already above the written requirement, accuracy and familiarity with the kind of text the job involves are usually the more productive things to work on.
What about programmers, writers and customer support?
We did not find an official or published employer standard that sets a typing speed for programmers, content writers, journalists or customer support agents in general. Many guides list ranges such as 50 to 70 WPM for support roles or 60 to 80 WPM for data entry, but the pages we reviewed did not link those ranges to job postings, surveys or measurements. They may be reasonable guesses, but you cannot rely on them as hiring criteria.
For these roles the reliable source is the posting you are applying to. If it names a WPM figure, ask three questions before you practice toward it. How long is the test? Is the score gross or net? How are errors counted? If the posting names no figure, the Aalto bands are a sensible way to see where you stand: about 52 WPM was average in that sample and roughly 78 WPM marked its fastest 10 percent.
Why two typing scores rarely match
Two tests can give the same person different scores because they define and count things differently. In the Aalto study, one word was any five characters, including spaces, and speed was measured from the first to the last keypress of each sentence. Other tests count the actual words in a passage. Some report gross speed, which divides everything typed by the time; others report net speed, which subtracts errors or ignores incorrect words.
The table below compares the measurements used in this article on the same criteria, so you can see which figures are directly comparable.
| Measurement | Word definition | Length | Error rule |
|---|---|---|---|
| Aalto 136M keystrokes study (2018) | Five characters | 15 sentences | Uncorrected error rate reported separately |
| OPM clerical standards | Not specified | 5-minute sample | Three or fewer errors |
| CalHR typing certificate | Not specified | 5-minute proctored test | Not specified on the requirements page |
| NCRA RPR skills test | Dictated words, steno | Three 5-minute legs | 95 percent accuracy per leg |
| Turkish court clerk exam (2025) | Actual words in the passage | 3 minutes | Each wrong word is an error; over 25 percent wrong invalidates the text |
The takeaway is that a 60 WPM score from a one-minute online test with no error penalty is not the same as 60 WPM on a five-minute test that allows three errors. Test length is part of the definition, which is why CalHR does not accept certificates from tests shorter than five minutes. The text also changes results. The Aalto sentences came from email and newswire collections, contained at most 70 characters, fewer than five numerical symbols and only simple punctuation. Work documents full of numbers, capital letters and symbols are a different task, so a score on simple sentences may not carry over.
A real-word example from outside the U.S.
Turkey's Ministry of Justice shows how much the word definition matters. Its 2025 court clerk hiring notice requires candidates to type at least 90 words in three minutes on a computer, without errors or repetition, counting the actual words of a passage rather than five-character units. That works out to 30 real words per minute. If more than 25 percent of the typed words are wrong, or 22 or more words are skipped, the text fails. Because a real word can be longer or shorter than five characters, 30 real words per minute is not directly comparable with 30 WPM on an English five-character test. Candidates who pass are then ranked by their number of correct words.
How to check whether your speed is enough for a job
The goal is to measure yourself the way the employer will. These steps keep the comparison fair.
- Find the test length in the posting or the agency's rules. OPM and CalHR both use five minutes, so a one-minute test will not tell you much.
- Type on a physical keyboard, since phone speeds in the Aalto mobile study averaged about 70 percent of desktop speeds.
- Use text similar to the job's work, including numbers and punctuation if the job involves them.
- Work out your gross speed first. You can enter the words you typed and the minutes it took into a typing speed calculator to get gross WPM.
- Count your errors separately and apply the employer's rule. For federal clerical positions, that means three or fewer errors over five minutes.
The NumerSpace calculator divides word count by time and does not measure errors or accuracy, so step five is yours to do by hand. Once you have a speed calculated under the employer's rules, compare it with the written requirement. If you are below it, work on accuracy first, because under rules like OPM's a fast sample with more than three errors does not qualify. If you are comfortably above it, clean output is usually a better use of practice time than extra speed.
Your next step
Before you start any practice plan, write down the target test's length, its error rule and whether it reports gross or net WPM. Then take one baseline measurement under those exact conditions and repeat the same measurement once a week. Comparing like with like is the most reliable way to know whether practice is moving you toward the requirement.
If short, regular sessions suit you, a Pomodoro timer can structure them, and the reading time calculator estimates how long a reader will need for the text you produce. The full set of study and productivity tools is on the education and productivity page. When a specific job is in view, the posting's own test is the number that counts.