Data Entry Typing Test
Real structured records — field data, codes and identifiers, with instant WPM and accuracy.
Type what a data entry operator actually keys — customer records, invoices, claims, trade confirmations and batch logs, complete with the codes, dates and identifiers. Built for keyers, claims processors, indexers and anyone facing a KPH assessment.
What you will actually be typing
Taken straight from the test. Read a few lines and notice what is missing: there are almost no ordinary words holding it together. That absence is the entire subject of this page.
Customer Record 1 of 4. Customer ID: CUS-10482. First Name: Daniel. Middle Initial: R. Last Name: Whitfield. Address Line 1: 4821 Oakwood Drive. Address Line 2: Apartment 3B. City: Springfield. State: IL. Postal Code: 62704. Country: United States. Phone: 217-555-0148. Alternate Phone: 217-555-0192. Email: d.whitfield@examplemail.com. Preferred Contact: Email. Account Opened: 03/14/2022. Status: Active. Customer Record 2 of 4. Customer ID: CUS-10483. First Name: Aleksandra. Middle Initial: M. Last Name: Nowakowski. Address Line 1: 77 Beaumont Terrace. Address Line 2: None. City: Portland. State: OR. Postal Code: 97209. Country: United States. Phone: 503-555-0217. Alternate Phone: None. …
Order Processing Record. Order ID: ORD-2026-BC-09421. Order Date: 04/12/2026. Order Time: 14:37:22 UTC. Channel: Web. Customer ID: CUS-20841. Customer Name: Bright Future Electronics. Purchase Order Reference: BFE-PO-77214. Currency: USD. Line 1. SKU: MER-MK880-BLK. Quantity: 40. Unit Price: 89.99. Line Total: 3,599.60. Line 2. SKU: MER-VM220-BLK. Quantity: 40. Unit Price: 39.99. Line Total: 1,599.60. Line 3. SKU: MER-DS900-BLK. Quantity: 15. Unit Price: 159.99. Line Total: 2,399.85. Subtotal: 7,599.05. Discount Code: VOL10. Discount Amount: 759.91. Net Subtotal: 6,839.14. Tax Rate: 8.25. Tax Amount: 564.23. Shipping Method: Two Day. Shipping Cost: 142.00. Order Total: 7,545.37. Payment Method: Account. …
Trade Confirmation Entry. Trade ID: TRD-20260412-EQ-0047821. Trade Date: 04/12/2026. Settlement Date: 04/16/2026. Executing Broker: Meridian Capital Markets LLC. Broker Code: MCM. Clearing Broker: Northgate Clearing Services. Clearing Code: NGC-0447. Account Number: ACC-8821-4470-EQ. Account Name: Kestrel Pension Trust. Side: Buy. Security Description: Halloway Industries Ordinary Shares. Ticker: HWI. ISIN: US4471882093. CUSIP: 447188209. SEDOL: B0X4471. Exchange: NYSE. Quantity: 24,500. Price: 38.4750. Gross Consideration: 942,637.50. Commission Rate: 0.0450. Commission: 424.19. Exchange Fee: 18.85. Regulatory Fee: 20.74. Net Consideration: 943,101.28. Currency: USD. Settlement Currency: USD. FX Rate: 1.000000. Trader ID: TRD-4471. Order Reference: ORD-20260412-0088. Execution Venue: XNYS. Capacity: Agency. …
Who this test is for
Data entry keyers, claims processors, order processing clerks, records and registry staff, document indexers and research coders — and anyone facing a pre-employment KPH or alphanumeric assessment. It is also the right page if a general typing test has told you your speed is fine and you suspect record capture is a different problem. It is.
| Durations | 1, 3, 5 or 10 minutes |
|---|---|
| Target speed | 45 WPM net — but accuracy is the constraint here, not speed |
| Scoring formula | 5-character standard word; one standard word deducted per incorrect word |
| Backspace | Allowed (configurable) |
| Difficulty levels | Easy (routine batches) · Medium (cross-referenced records with exceptions) · Hard (specialist regulated capture) |
| Passage corpus | 30 passages of structured records |
| Field density | 12.82 digits per 100 characters and 290 proper nouns per 1,000 words — both the highest on this site |
| Cost | Free, no login, results never leave your browser |
This is not prose, and that changes everything
Every other typing test on this site puts sentences in front of you. Sentences have grammar, and grammar is a prediction engine — it tells your fingers what is probably coming next. Structured records have no grammar at all, and the difference is measurable.
We counted function words — the, and, of, to, is, for — the connective tissue that holds writing together. In real prose they make up roughly a quarter to a third of everything. Here they barely appear.
What is left when you remove the grammar is a sequence of discrete facts. Consider what your hands do with the patient was admitted to the — they run ahead, because only a few words can follow and you have typed all of them a thousand times. Now consider Policy Number: POL-2019-HO-338821. Nothing predicts anything. There is no redundancy, no context to catch an error, and no rhythm to settle into.
This is why people who type 70 WPM on prose are often startled by their first data entry assessment. Their speed has not gone anywhere. The material simply stopped helping them.
| Level | What it is | Digits / 100 chars | Capitals / 100 letters | Proper nouns / 1,000 |
|---|---|---|---|---|
| Easy | Routine record batches | 12.98 | 18.5 | 265 |
| Medium | Cross-referenced with exceptions | 12.65 | 18.6 | 304 |
| Hard | Specialist regulated capture | 12.82 | 17.6 | 300 |
KPH and WPM: the conversion nobody explains
Because the material is not made of words, the industry does not count words. Data entry is measured in keystrokes per hour, and job postings quote figures like 8,000 or 10,000 KPH that sound enormous next to a WPM number. They are the same thing in different units, and the conversion is one division.
KPH ÷ 300 = WPM
An hour contains 60 minutes and a standard word is 5 characters, so 60 × 5 = 300. Divide any KPH figure by 300 and you have its equivalent in words per minute. That is the whole calculation.
| Keystrokes per hour | Equivalent WPM | Where you tend to see it |
|---|---|---|
| 6,000 | 20 WPM | Entry level, often with training provided |
| 8,000 | 27 WPM | Common minimum for general data entry roles |
| 10,000 | 33 WPM | The most frequently advertised benchmark |
| 12,000 | 40 WPM | Experienced keyer, production environments |
| 15,000 | 50 WPM | Specialist high-volume capture |
Two things worth taking from this. First, the numbers are far less intimidating than they look — the widely quoted 10,000 KPH is about 33 WPM, which most competent typists already exceed on prose. Second, and less comfortably, hitting it on this material is a different matter from hitting it on sentences, for the reasons set out above.
One caution on the conversion. It assumes every keystroke counts, which is how most systems measure it, but some employers count only the characters that end up in the field and exclude tab presses and corrections. Ask which method applies if the difference matters to you.
Why serious capture is keyed twice
In clinical trials, registries, census work and regulated financial capture, the same document is often keyed independently by two operators and the results compared automatically. Where they differ, someone made an error and the record goes for a third look.
It is a deliberately expensive control — you are paying twice to capture once — and it exists for exactly the reason described above. Coded data cannot be proofread. A reviewer reading back POL-2019-HO-338821 against the source has no linguistic cue to alert them to a wrong digit; they are performing a character-by-character comparison, which is slow and which humans do poorly. Two independent keyings converge on the truth far more reliably than one careful proofread does.
The practical lesson for practice is what it implies about your own error rate. You almost certainly do not know it, because backspace hides it — you fix mistakes as you notice them and the corrected output looks clean. Run one session a week with backspace disabled and you will see your genuine first-pass rate, which is the figure that predicts how many errors would reach a live record.
Character accuracy flatters you; field accuracy does not
General typing tests report character or word accuracy. Data entry employers usually assess accuracy at the level of the field or the record, and the two measures can tell very different stories about the same performance.
The same error, scored two ways
A policy number is 20 characters. You get one digit wrong. Character accuracy: 95%. Field accuracy: 0% — because a policy number that is nearly right is simply wrong, and it will match no policy at all.
Now scale it. Key 50 records of 30 fields each and make 15 single-character errors spread across them. Character accuracy stays around 99%. Record accuracy is 70%, because 15 records contain a defect. Both numbers are honest. Only one of them describes the job.
This is the gap that surprises candidates who arrive confident from a general typing test. It is also why the practice plan below puts accuracy targets ahead of speed targets in every single week rather than building to them.
Roles and how each one is measured
As elsewhere on this site, the caveat first: no professional body publishes a required speed for any of these roles. Thresholds are employer conventions. What is more useful is knowing which unit you will be judged in.
| Role | What you key | How it is measured | Where errors bite |
|---|---|---|---|
| Data entry keyer | Bulk record capture from forms, scans or dictated lists | KPH, usually with a separate alphanumeric and numeric score | The only role where raw keying rate is the primary metric |
| Claims processor | Insurance and medical claims, coded fields, policy references | Claims per hour with an accuracy floor | A wrong policy number routes the claim to the wrong contract entirely |
| Order processing clerk | Purchase orders, SKUs, quantities, delivery addresses | Orders per hour, error rate sampled | Address errors surface days later as failed deliveries |
| Records / registry clerk | Registrations, licences, membership, patient demographics | Records per shift, verification pass rate | Frequently double keyed because the record is the legal source |
| Document indexer | Metadata for scanned documents, dates, reference numbers, types | Documents per hour | An indexing error makes the document unfindable rather than wrong |
| Survey / research coder | Response coding, open-text classification, batch validation | Cases per hour with inter-coder agreement checks | Accuracy assessed against another coder rather than a source document |
On the market, an honest note. Automated capture has removed a great deal of routine keying, and one signal of that is administrative rather than statistical: the Bureau of Labor Statistics no longer publishes a full Occupational Outlook profile for data entry keyers, listing the occupation in its summary table of occupations not covered in detail instead [2]. Wage and employment figures remain available through the occupational employment statistics series [1].
What that means practically is that straight bulk keying is a shrinking market while exception handling, verification and audit of machine-captured data are not. Those roles need the same skill and generally pay better, but they ask for demonstrable accuracy rather than speed alone.
A four-week plan
Around 20 minutes a day. Note that every week leads with an accuracy target rather than a speed one — on this material that is not a motivational choice, it reflects how you will be assessed.
| Week | Focus | What to run | Move on when |
|---|---|---|---|
| Week 1 | Field rhythm | Easy passages, 5-minute runs | 97% accuracy; colons and labels typed without hesitation |
| Week 2 | Codes and identifiers | Medium passages, plus the alphanumeric test | Zero transpositions in any reference across a full run |
| Week 3 | Numeric pad | Medium and Hard, plus the numbers test for keypad drilling | Digits typed by touch, no glances at the keypad |
| Week 4 | Verification discipline | Hard passages, 10-minute runs, backspace off once per session | 45+ WPM net at 98% accuracy, and a known first-pass error rate |
Pair this with the numbers test for keypad work and the alphanumeric test for reference codes. If your general speed is below about 35 WPM, work through the free touch-typing course first — records are a poor place to build basic technique, because there is no rhythm to learn from. For finance-specific numeric work, the accounting typing test covers the same digits inside prose rather than inside fields.
Before a data entry assessment: a self-check
- You know whether the employer measures KPH or WPM, and which conversion they apply.
- You know whether they score character accuracy or field accuracy — the difference is large.
- Accuracy holds above 98% on Medium passages, not just Easy ones.
- No transposition appears in any reference code across a full 10-minute run.
- You type digits by touch, on the keypad or the number row, without looking.
- You have run a session with backspace off and know your true first-pass error rate.
- You know whether numeric and alphanumeric are tested separately.
Frequently asked questions
What is a good data entry typing speed?
It depends which unit the employer uses, and that catches people out constantly. General typing is quoted in words per minute; data entry is usually quoted in keystrokes per hour. The conversion is simple once you see it — divide KPH by 300 to get WPM, because an hour holds 60 minutes and a standard word is 5 characters. So a common 10,000 KPH benchmark is about 33 WPM, and 12,000 KPH is about 40. Those are far less intimidating than the raw numbers look.
Why is data entry measured in KPH instead of WPM?
Because there are barely any words. We measured this corpus against the others we publish: only 3.5% of it is function words — the, and, of, to — against 25.8% in general typing prose and 36.7% in accounting text. Structured records are field labels, codes, dates and identifiers. A "word" is not a meaningful unit when a third of your keystrokes are a policy number, so the industry counts keystrokes instead. The unit follows the material.
Why does data entry feel harder than typing at the same speed?
Because prose carries you and records do not. When you type "the patient was admitted to the", your fingers run ahead — grammar, spelling and context all predict what comes next, and a small error usually still reads correctly. In "Policy Number: POL-2019-HO-338821" nothing predicts anything. Every character is an independent fact with no redundancy behind it. You are transcribing rather than writing, and transcription runs at the speed of verification rather than the speed of your hands.
What is double-key verification?
Two operators key the same source document independently, and the system compares the results. Where they disagree, one of them made an error and the record is flagged for a third look. It is deliberately expensive — you are paying twice to capture once — and it exists because numeric and coded data cannot be proofread by reading. It is still standard for anything where the record is the legal source: clinical trials, registries, census work, and regulated financial capture.
What is the difference between character accuracy and field accuracy?
Character accuracy counts wrong characters against total characters. Field accuracy asks whether each field is entirely correct. They diverge sharply. A 20-character policy number with one wrong digit scores 95% on characters and 0% on that field, because a policy number that is nearly right is simply wrong. Most data entry employers assess field or record accuracy, which is why a 98% character-accuracy figure from a general typing test can badly overstate how you would actually perform.
Should I learn the numeric keypad?
Yes, if numeric fields are a real part of your target role. It is a separate skill from touch typing on the main rows. The home position is 4-5-6 under the index, middle and ring fingers, with the raised dot on 5 as the anchor. Most competent typists still look down at the keypad, which quietly caps their numeric rate. This corpus runs at 12.82 digits per 100 characters, so if you are looking down here you would be looking down constantly on the job.
What kind of text is in this test?
Three graded levels of real structured records. Easy passages are straightforward batches — customer records, invoices, employee files, shipping manifests. Medium passages add cross-referenced records with validation and exception handling. Hard passages are specialist capture: trade confirmations, customs entries, payment instructions, pharmaceutical batch records, aircraft maintenance logs and specimen chain of custody. The corpus holds 30 passages.
Do employers test numeric and alphanumeric separately?
Often, yes, and the two scores can differ considerably. Alphanumeric tests present mixed field data of the kind on this page. Numeric tests present figures only and are usually keyed on the numeric pad. Some employers also run a 10-key test as a third component. If a posting quotes two numbers, that is what they are, and it is worth knowing which of your two rates is the weaker one before you sit the assessment.
Is data entry work disappearing?
It is contracting faster than most office roles, and worth being straight about. Optical character recognition, automated form capture and system-to-system integration have removed much of the routine keying that once existed. One telling signal: the Bureau of Labor Statistics no longer publishes a full Occupational Outlook profile for data entry keyers, listing the occupation instead in its summary table of occupations not covered in detail [2]. The work that remains is exception handling, verification, and capture that machines still cannot read reliably.
Does that mean the skill is worthless now?
No, but its value has moved. Bulk keying is a shrinking market. Accurate, verifiable capture of the records machines get wrong is not, and neither is the adjacent work — validating automated output, resolving exceptions, and auditing captured data against source. Those roles pay better than straight keying and they still require exactly the skill this page trains. What has changed is that speed alone is no longer the product; speed plus demonstrable accuracy is.
Can I use backspace during this test?
Yes, and it is on by default. Turning it off for one session a week matters more on this page than most, because backspace lets you fix a mistyped digit the moment you notice it — which flatters your accuracy in a way that real capture does not, since nothing in a record prompts you to look again.
Is this data entry typing test free?
Yes. No login, no signup, no limit on attempts. Scoring runs in your browser and your results are never uploaded to our servers.
Sources
Digit, capital and proper-noun figures are computed from our own corpora and can be regenerated with scripts/corpus-stats.mjs. The function-word percentages use a fixed list of 45 common English function words applied identically to every corpus; the list is a judgement rather than a mechanical property of the text, so the absolute figures are indicative while the comparison between corpora is sound. The KPH to WPM conversion is arithmetic from the 5-character standard word and is shown in full rather than cited. Typing and keystroke ranges are drawn from commonly advertised roles and are labelled as such — no professional body publishes a required figure for this occupation.
- Occupational Employment and Wage Statistics: Data Entry Keyers (43-9021)U.S. Bureau of Labor Statistics · link checked 21 Aug 2026
- Occupational Outlook Handbook: Data for Occupations Not Covered in DetailU.S. Bureau of Labor Statistics · link checked 21 Aug 2026
- Employment Projections — 2024–2034 news releaseU.S. Bureau of Labor Statistics · link checked 21 Aug 2026