Ask what the data entry test measures

Confirm whether the assessment uses WPM or KPH, its duration, allowed corrections, numeric keypad requirements, field separators, and accuracy threshold. The same label can describe different scoring rules.

Practice the same field pattern

Names, addresses, codes, amounts, and tabs create different demands from prose. Use structured fictional records rather than random words, and practice a 10-key test separately when a numpad is required.

Optimize for net output and accuracy

A slightly lower raw rate with fewer errors often produces stronger net KPH. Build a cadence that remains accurate when fields change; do not call a result job-ready without knowing how the employer handles corrections.

Repeat under comparable conditions

Use the same duration, keyboard, layout, and correction policy for practice and comparison. A short burst can be useful for technique, but it is not evidence of sustained data-entry performance by itself.

Check the environment

Use a reliable keyboard, verify Num Lock, close distracting software, and understand reasonable accessibility accommodations before the test begins.

Seven days of synthetic amounts and records

Choose Day 1 through Day 7 in the data-entry tool. Every stage proposes two active minutes: baseline; quantity/amounts; names/addresses; full mixed records; transitions; repeat the baseline stream; new transfer values. The selector does not claim that seven calendar days have elapsed or that selecting a stage completes it. The final result says whether you finished the planned duration.

Days 1, 6 and 7 use eight fields in the same order. Day 6 intentionally repeats Day 1. At the same record index, Day 7 changes every field value while matching field lengths. This controls length and structure, not all reading difficulty. Keep your declared layout, equipment, duration and correction rules the same. Focused Days 2 and 3 have different field counts, so do not compare their KPH to full records.

10-key KPH and field-entry KPH: a worked comparison

In a hypothetical two-minute 10-key run, 240 counted attempts (including eight target Tab separators) give 240 / (2/60) = 7,200 raw KPH. If 220 target characters remain correct, net KPH is 6,600. Correct final characters divided by five and then by two minutes would be 22 WPM, but it is a conversion of that count, not a separate prose assessment or an employer threshold.

In the data-entry form, Tab, Enter and Backspace are excluded from character attempts. An exact first baseline record contains 64 characters across eight fields, so one exact record in two minutes gives 1,920 net KPH. If 69 characters were attempted while correcting, raw KPH is 2,070 and exact-field accuracy can still be 100%. These different rules explain why comparing the two headline numbers alone is misleading.

Names and addresses: an error you can reproduce

The original synthetic first baseline record is TW-100000; Ada Moreno; 42 Cedar Road; Demo Town; Cable kit; 10; 100.40; Review. The new transfer record is TW-600000; Eva Montes; 73 Maple Lane; Test City; Label pad; 60; 600.20; Queued. These are fictional, not customer records.

Deliberately submit Ada instead of Ada Moreno while copying the other seven fields exactly. The name field earns zero exact-field characters, the other fields remain independent, and the record is not exact: seven of eight fields = 87.5% field accuracy, zero of one records = 0% record accuracy. With two active minutes, 54 exact-field characters give 1,620 net KPH. Check focus by moving through all eight inputs with Tab before a timed run.

Numbers or field changes: a controlled worksheet

Use the same two numeric strings, 10 and 100.40, in two short manual trials: A places them in one continuous line with a separator; B puts them into two separate labeled fields with Tab. Record the text, time, corrections and navigation count. Reverse A/B order and repeat at least twice under the same layout and equipment.

This worksheet can expose a repeatable transition cost; it does not diagnose its cause from one attempt. Keep the two scoring systems separate: a target Tab may count in 10-key mode, while navigation Tab does not count in the field form. The Day 2 quantity/amount exercise practices the second format. Do not treat a phone number row as a physical-numpad technique test.

Fresh fictional emails instead of one memorized passage

Three new original messages in the email practice pool use a delivery confirmation, a meeting correction and an invoice exception. Example: Hi Lena, the sample kit will arrive on March 12 at 09:30. Please confirm the loading-bay number and keep the delivery receipt with order TW-7310. If the time changes, send the revised window before 16:00.

The complete delivery message has 204 characters including spaces, 35 whitespace-separated words, 14 digits and 9 punctuation characters. The meeting correction has 228 characters, 40 words, 10 digits and 9 punctuation characters; the invoice exception has 227 characters, 37 words, 14 digits and 7 punctuation characters. These counts describe the individual source messages, not a whole timed passage: the workout can combine messages. Whitespace word counts are descriptive; scoring still uses five characters per conventional word.

The delivery message mixes a hyphenated term, an order reference and clock times. The meeting correction requires distinguishing two room numbers and a time; the invoice includes a reference and two decimal amounts. These are content differences to record, not a validated difficulty scale. The same active-time WPM formula applies to all three: correct final characters / 5 / active minutes. Keep duration, layout, equipment, visible window and correction rules constant; their different lengths and familiarity can still affect a comparison.

Use the email workout for prose WPM with punctuation. Change passages between trials, retain the same duration, and note lengths and unfamiliar names or numbers. A shorter or easier message is a changed condition. Grammar quality, reply judgment and actual job performance are not scored. Never paste confidential correspondence to reproduce this case.

Understand a gain and export its conditions

Compare Day 1 baseline, Day 6 repeated stream and Day 7 new matched-length stream. A larger gain on Day 6 than Day 7 may include familiarity; it is not proof of a training effect or a causal estimate. Use several attempts and retain field accuracy, exact records, active time and any shortened run beside KPH.

The data-entry result exports a local JSON with summary counts, selected stage/cohort, field IDs, layout, equipment, planned duration and actual active time. It excludes field values and typed text. Completed session metrics also remain in local history, with the practice stage tagged. Export before clearing browser data. A selected stage without submitted records does not create a completed session.

The downloadable executed dossier records three actual browser UI runs on October 22, 2026: baseline (Day 1 selector), focused practice (Day 2) and repeat (Day 6). All input was automated CUA typeText/Tab events using fictional records, with Other layout and Other input declared. No physical keyboard, human typing performance or six elapsed training days is represented. Before looking at scores, check the conditions: all three were configured for 120 seconds but ended early; the practice run used two fields instead of eight. The viewport was 1280 × 720 at the captured observations, while continuous visibility between captures was not measured. The error criterion remained exact-field-characters-v1, and the product excludes hidden-tab time.

The original UI exports report about 35.45 active seconds and 64 exact characters for baseline, 4.63 seconds and nine characters for focused practice, and 7.39 seconds and the same 64-character source record for repeat. Their headline net KPH values are 6,499, 6,995 and 31,191, respectively; every submitted field matched. These are shortened automation runs, not comparable full-duration assessments. The same KPH formula scales different active-time denominators, and the focused task also changes the field scope. No score difference is attributed to learning or training. The dossier keeps the unmodified exports, observation timestamps, source/result captures, screenshots and condition checks; it adds only deliberately fictional source text. Ordinary private result exports still exclude text and field values.