CoreHRx Jobs

Apply in one click, save jobs and take free mock tests.

Register free Log in
Hiring? Post jobs with CoreHRx
My account
user

View & update profile
Interview guide

Data Entry Operator Interview Questions and Answers: 60 Questions from Easy to Hard

This guide is for freshers and working operators in India preparing for data entry operator interviews at offices, hospitals, bank back offices, e-commerce companies and government projects run through contractors. It has 60 questions with model answers in three levels: typing tests, Excel entry and accuracy habits for freshers; validation, cleaning, ERP uploads, error rates and targets for operators with one to three years; and quality checks, audits, team targets and privacy incidents for experienced operators and team leads. Every formula and number in the answers was run or worked out. For general office software such as Word, Outlook and basic Excel, see our MS Office and Computer Basics guide. Answer each question aloud first, then compare, and take the Data Entry Operator Test to practise under time.

  • 60 questions, 3 levels
  • Updated
  • CoreHRx Jobs

What data entry operator interviews cover

  • A typing or data entry test, often on the spot: copy typing scored in words per minute, or key depressions per hour in government and contract roles, with accuracy counted alongside speed.
  • Excel and form entry: keying records cleanly, dates and long numbers, validation, simple check formulas and cleaning messy columns.
  • The software you will use: an ERP, a hospital information system, a bank back-office screen or an upload template, and how master data differs from daily transactions.
  • Accuracy habits and confidentiality: how you check your own work, what you do with an unclear form, and how you handle personal data.
  • For experienced and lead roles: QC sampling, error rates, team targets, audits and privacy incidents.

How freshers should prepare

  • Practise touch typing for 20 to 30 minutes a day on names, addresses, PIN codes and amounts, not just paragraphs, and track speed and accuracy every week.
  • Build a small Excel entry sheet of 50 made-up records with dropdowns and a check column, and be ready to explain how it stops errors.
  • Learn the numeric keypad and the Excel keys you use every minute: Tab, Enter, Ctrl and semicolon for the date, and Alt and Down arrow for a dropdown.
  • If the job needs Hindi or another Indian language, practise on the layout the employer names, often Inscript with a Unicode font such as Mangal.

How experienced candidates should prepare

  • Know your own numbers: forms or records per hour, field accuracy from QC reports and the targets you met, and say how each was measured.
  • Prepare two STAR stories: an error you found and traced to its cause, and a change that cut errors or time for your team.
  • Be ready to describe the software you used, such as ERP screens, a hospital system, bank screens or upload templates, and the checks built into it.
  • For lead roles, revise QC sampling, weighted error scores, audit trails and how you would handle a privacy incident or a missed deadline.

Typing tests, notifications and job offers

  • Government and contract posts state the test in the notification: for example, SSC notices have asked many Data Entry Operator candidates for 8,000 key depressions per hour in a 15-minute test. Read which speed, language and keyboard layout your post needs.
  • Private employers often test on their own material, such as sample forms, so practise numbers and codes as well as text.
  • Work-from-home roles may ask about your laptop, internet connection and a private place to work. Be careful with typing job offers that ask for a registration or training fee first.
  • Contract roles in hospitals, banks and government projects often run in shifts with daily targets, so ask how output and accuracy are measured before you accept.

Common mistakes to avoid

  • Quoting a typing speed you have never measured, or a speed without an accuracy figure.
  • Saying you would guess an unclear entry to keep up speed. Say you would flag it and ask.
  • Treating Excel like a typewriter: merged cells, two values in one cell, long numbers typed as numbers and dates typed in mixed styles.
  • Describing a previous employer's customer records in detail, or saying you shared files through personal WhatsApp or email.
  • Talking only about speed. Always pair it with accuracy and how your work was checked.

Data Entry Operator interview questions: Easy questions (freshers)

20 questions

1. What does a data entry operator do, and which kinds of employers hire for the role in India?

A data entry operator turns information from paper forms, scanned documents, emails or other systems into correct records in a computer system. Hospitals hire operators for patient registration and bills in the hospital information system, bank back offices for account opening and KYC details, e-commerce firms for product listings and orders, and government projects, often through contractors, for scheme beneficiary or land records. A normal day is a batch of work: enter it, check your own entries, submit it to a checker, fix what comes back and report your count. Interviewers want to hear about accuracy and confidentiality, not only typing.

2. Which matters more in data entry, speed or accuracy? Explain with the cost of fixing an error.

Accuracy comes first, because an error costs far more time later than rushing saves now. Suppose rushing lets you finish a batch 30 minutes sooner but leaves 10 more errors, and each one takes a checker about 8 minutes to trace, compare with the source and correct. The team spends 80 minutes to save 30, a net loss of 50 minutes, and that counts only the errors the checker catches. An error nobody catches costs much more: a wrong account number sends a salary to a stranger, and a wrong date of birth can hold up a patient's insurance claim. So build accuracy first, then raise speed steadily.

3. What is touch typing, and how would you raise your speed from 25 to 40 WPM before joining?

Touch typing means typing with all ten fingers without looking at the keys. The fingers rest on the home row, A S D F and J K L and semicolon, and the small bumps on F and J let you find it without looking. To go from 25 to 40 WPM, practise 20 to 30 minutes every day on a typing tutor, first for accuracy of 98% or more and only then for speed. Practise on the kind of text the job uses, such as names, addresses, PIN codes and amounts, and measure once a week. The common mistake is practising fast with errors, which turns the errors into habits.

4. A job notice asks for 8,000 key depressions per hour. What does that mean in a 15-minute test, and roughly what WPM is it?

A key depression is one press of a key, so letters, digits, spaces and punctuation all count. 8,000 an hour means 2,000 key depressions in 15 minutes. With the usual rule of 5 key depressions to a word, 8,000 an hour is 1,600 words, or about 27 words a minute. That sounds easy, but these tests count only key depressions that match the passage. SSC notices for the Combined Higher Secondary Level exam have used this standard for many Data Entry Operator posts, with a passage of about 2,000 to 2,200 key depressions, and a higher speed for some posts. Always read the current notice for your post.

5. How do you use the numeric keypad for entering numbers, and what can go wrong with Num Lock?

On the numeric keypad the right hand rests on 4, 5 and 6, with the middle finger on 5, which has a small bump. The index finger covers 1, 4 and 7, the middle finger 2, 5 and 8, the ring finger 3, 6, 9 and the decimal point, the thumb takes 0 and the little finger presses Enter. For long columns of amounts this is much faster than the top row. If Num Lock is off, the same keys act as arrows, Home, End and Page Up, so the cursor jumps and entries land in the wrong cells. Check the Num Lock light before a batch, and on a laptop without a keypad ask for an external one.

6. In Excel, how do Tab and Enter help you key one record after another?

Press Tab after each field to move right, and press Enter after the last field of the record. If you moved across the row with Tab, Enter takes you back to the column where you started, one row down, so your hands stay on the keyboard for the next record. If you fill one column at a time instead, Enter moves down by default, and its direction can be changed under File, Options, Advanced. Clicking into each cell with the mouse is slow and a common cause of entries in the wrong row. Keep your eyes on the source document and glance at the screen only to check.

7. Why should each column in a data entry sheet hold one kind of information and each row one record?

Because sorting, filtering, totals, lookups and uploads all depend on that layout. If one cell holds "Ravi Kumar, 9876543210, Pune", nobody can filter by city or check the mobile number, and an upload to another system fails. Keep one header row, then one record per row, with separate columns for name, mobile, city and so on. Do not merge cells, leave blank rows between records, put two phone numbers in one cell with a slash, or add a total row in the middle of the data. A clean layout also lets the checker compare your sheet with the source quickly.

8. Why should you use Paste Values instead of a normal paste when you copy data into an entry template?

A normal paste copies everything from the source cell: its value, its format and its data validation. So pasting from another file can quietly remove the dropdowns and rules set up in the template, change date or number formats, and bring in formulas that point to the other file. Paste Special, then Values, copies only the values, so the template keeps its rules; in Excel for Windows, Ctrl, Alt and V opens Paste Special. If the source columns are in a different order, compare the headers and paste one column at a time. Pasting over the template's formula cells is the worst case, because its totals stop working.

9. How should you type dates in Excel so that they are not misread, and how do you check that Excel understood them?

First find out which date order the computer uses. Office computers in India are usually set to day, month, year, but a computer set to the US style reads 05/06/2025 as 6 May, not 5 June. The safest entry is the month as a word, such as 5-Jun-2025, which cannot be misread, with the whole column shown in one format. To check, a real date is right-aligned by default and =ISNUMBER(A2) returns TRUE, while a date stored as text, such as 05.06.2025 typed with dots, returns FALSE. Never mix styles in one column, because sorting and filters then go wrong.

10. How do you avoid look-alike character errors, such as 0 and O, in codes like PAN and IFSC?

Know the pattern of each code, because the pattern tells you which character belongs in each place. A PAN has 10 characters: five letters, four digits and a final letter, so a round shape in the sixth to ninth places must be the digit 0, and one in the first five places must be the letter O. An IFSC has 11 characters: four letters for the bank, then a fifth character that the RBI says is always 0, the digit, then six characters for the branch. The same logic separates 1 and I, 5 and S, and 8 and B. Read codes in small groups against the source, and ask when the handwriting is really unclear.

11. One field on a handwritten form is hard to read, and another field is empty. What do you do?

I never guess. A guessed digit in a mobile or account number looks perfectly valid in the system, so nobody catches it later. I enter everything I can read clearly, mark the doubtful field in the way the team uses, such as a comment or a query flag, and note the form number, field and problem in a query sheet. For the empty field I follow the agreed rule for blanks, either leaving it empty or entering the agreed value, never my own choice such as NA in one place and a dash in another. At the end of the day I send the query list to the supervisor or the source team.

12. The same city is written as Bangalore, Bengaluru, B'lore and BLR in one sheet. Why is that a problem, and how do you prevent it?

A computer treats each spelling as a different city, so a count by city splits one city into four rows, a filter for Bengaluru misses most records and lookups fail. The fix starts before entry: agree one standard name for each value, keep the approved list on a separate sheet, and use a dropdown so operators pick instead of typing. For data already entered, replace each variant with the standard name using Find and Replace with Match entire cell contents switched on, so the same letters inside a longer value, such as a street name, are not changed. Then count the distinct values again.

13. A client's rule says: leave a field blank if the applicant skipped it, type 0 only if the amount is really zero, and type NA if it does not apply. Why does the client care?

Because each one means something different, and the client's figures depend on it. A blank says the value is unknown, 0 says it is known to be zero, and NA says the question does not apply, so it should not count at all. Excel follows the same logic: AVERAGE skips blanks and text such as NA but includes zeros. In the sheet shown, the average income as entered is 16,000. If Bharat's blank had been typed as 0, it would fall to 12,000, and if Chitra's real 0 had been left blank, it would rise to 24,000. One careless habit changes the report.

Example
     A          B           C                 D
1  Applicant  As entered  Blank typed as 0  Zero left blank
2  Anita      30000       30000             30000
3  Bharat     (blank)     0                 (blank)
4  Chitra     0           0                 (blank)
5  Deepak     18000       18000             18000
6  Esha       NA          NA                NA

=AVERAGE(B2:B6)  -> 16000
=AVERAGE(C2:C6)  -> 12000
=AVERAGE(D2:D6)  -> 24000

14. What changes when you enter data in a Google Sheet that several people edit at the same time?

Everyone works on the same live copy and every change saves at once, so there is no private version to fall back on. Three habits matter. Work only in the rows or tab assigned to you. Do not apply a normal filter or sort to the whole sheet, because a normal filter changes the view for everyone; use Data, then Create filter view, which only you see. And if something looks wrong, check before you fix it, because version history shows who changed what and when. Also leave the sharing settings alone: changing a sheet of customer data to anyone with the link is a privacy leak.

15. Why can default values and mandatory fields on an entry screen cause errors?

A default value fills a field before you type, such as today's date, the state Maharashtra or the gender Male. If the operator forgets to change it, the record looks complete and passes every check, yet it is wrong, so defaults hide errors instead of showing them. Mandatory fields cause a different problem: when the form has no value, people type something just to get past, such as 0000000000 as a mobile number or a full stop as the address. I handle both with care: I check every defaulted field against the form, and I follow the agreed code for missing values, so the data shows what is really unknown.

16. What is the difference between a scanned PDF and a searchable PDF, and why does it matter in data entry?

A plain scanned PDF is only a picture of each page, so you cannot search it or select its text. A searchable PDF has had OCR run on it, which puts a hidden layer of text behind the image, so Ctrl and F works and text can be copied. To tell them apart, try to select a word. For data entry this matters in two ways. A searchable PDF lets you find a name or number quickly in a long file. But text copied from the OCR layer can carry OCR mistakes, such as the letter O for the digit 0, so check every copied number against the image before you save it.

17. Is it all right to take a photo of a hard-to-read form on your phone to ask your supervisor about it?

No. A photo puts a customer's personal data on a personal phone, where it can sync to a cloud backup, sit in a chat app or be seen by anyone who uses the phone, all outside the company's control. Use the work route instead: note the form or record number and the field in the query sheet, show the supervisor the form or the scan on the work screen, or use the internal system or team mailbox if the supervisor sits elsewhere. The same applies to sending details on WhatsApp. Interviewers ask this because it is exactly the kind of shortcut that quietly becomes a habit, and it can cost the job.

18. The job needs Hindi typing. What are the Inscript and Remington layouts, and why does Unicode matter?

Inscript is the standard Indian keyboard layout for Indian scripts, and Windows includes it, for example as the Devanagari - INSCRIPT keyboard. Remington, sometimes called Remington Gail, follows the old Hindi typewriter layout that many experienced typists learnt. Phonetic typing, where you type namaste in English letters and get the Hindi word, is easy for beginners, but tests name the layouts they allow. Unicode fonts such as Mangal store real Hindi characters, so the text can be searched and opened on any computer. Older fonts such as Kruti Dev only look like Hindi and store English letter codes underneath. Check the notice before you practise.

19. In your first week you manage 200 forms a day, but the team target is 300. What do you do?

I would tell my supervisor early, with my daily numbers and error counts, and ask whether new joiners get a ramp-up target, for example 200 in week one, 250 in week two and 300 by week three. Then I would find where my time goes: reading unfamiliar forms, looking up codes, or moving between fields with the mouse. Learning the field order and the shortcuts, and asking an experienced colleague how they handle the tricky fields, usually gives the biggest gain. I would keep my accuracy where it is while the count rises, because a fast count full of errors is worth less to the team than a slower clean one.

20. Why do you want to work as a data entry operator, and how do you stay focused on repetitive work?

Answer honestly and link your answer to the job. For example: I am accurate and patient, I type at about 38 WPM with 98% accuracy, and I like work where the result can be checked. The role also shows me how a business runs, through its customers, bills and records, and I want to grow into a checker or MIS role. To stay focused I work in batches with a short check at the end of each, take my breaks on time, keep only the current batch on my desk, and track my count and errors so the day has small goals. Never say you want an easy job.

Ready to check yourself? Take the free Data Entry Operator mock test

Data Entry Operator interview questions: Medium questions (1-3 years)

20 questions

21. Your typing test score is 40 WPM, but at work your output is counted in forms per hour. How do you describe your speed in an interview?

I give both numbers and say how each was measured. Copy typing from a clean passage is faster than real entry, where I read handwriting, move between fields, pick dropdown values and check codes. So I might say: about 40 WPM at 98% accuracy on a typing test, and in my last job about 45 twelve-field KYC forms an hour at 99.6% field accuracy on the QC report. That shows the interviewer that I understand the difference and that my figures come from a real measure. Quoting a high WPM with no accuracy figure is the common mistake, and many employers will test you anyway.

22. How would you add a check column that flags blank fields and wrong lengths in a sheet you have just entered?

I add a column at the end with a formula that tests each row and returns a short reason, then filter it for anything other than OK. In the sheet shown, row 4 has a 9-digit mobile number and row 5 a 5-digit PIN code. Row 3 is the trap: its DOB cell holds a single space, typed to get past a mandatory field. COUNTBLANK treats that cell as filled and returns 0, so a check built on it says OK. Testing key text fields with TRIM, as in TRIM(B3)="", catches it. Put the most serious check first, and correct every flagged row against the source, never from memory.

Example
     A            B           C            D        E
1  Name         DOB         Mobile       PIN      Check
2  Ravi Kumar   12-03-1998  9876543210   411001   OK
3  Sana Shaikh  (a space)   9123456789   400053   Blank field
4  Manoj Rao    05-11-1995  987654321    560034   Check mobile
5  Priya Nair   21-07-2000  9988776655   68203    Check PIN

E2: =IF(OR(TRIM(A2)="",TRIM(B2)=""),"Blank field",IF(LEN(C2)<>10,"Check mobile",IF(LEN(D2)<>6,"Check PIN","OK")))
=COUNTBLANK(A3:D3)  -> 0

23. A vendor form gives both a PAN and a GSTIN. What quick check shows that one of them was typed wrongly?

A GSTIN has 15 characters: a two-digit state code, then the 10-character PAN of the business, then three more characters. So characters 3 to 12 must equal the PAN, and the first two digits must be the state code of the registered address, such as 27 for Maharashtra or 29 for Karnataka. The formula shown tests the length and the PAN match: in row 3 two digits of the PAN part were swapped, and in row 4 a character is missing. It catches typing errors, not fake numbers, so a mismatch means going back to the documents or the GST portal's taxpayer search. The numbers are made up.

Example
     A              B            C                 D
1  Vendor         PAN          GSTIN             Check
2  Kiran Traders  ABCPK1234L   27ABCPK1234L1Z5   OK
3  Sai Agencies   AAGFS5678M   27AAGFS5687M1Z2   Check
4  Om Stores      AFHPM4321Q   29AFHPM4321Q1Z    Check

D2: =IF(AND(LEN(C2)=15,MID(C2,3,10)=B2),"OK","Check")

24. Mobile numbers arrive as +91 98765 43210, 098765-43210 and 9876543210. How do you bring them to one 10-digit format?

Remove the spaces, hyphens and plus signs with SUBSTITUTE, then keep the last 10 digits with RIGHT, as shown. That handles the +91 and leading 0 versions, and the check column flags the 9-digit number in row 6. One trap: RIGHT also cuts an 11-digit typing mistake down to 10 digits that look valid. So count the digits before trimming: 10 is fine, 11 only if it starts with 0, 12 only if it starts with 91, and anything else goes back to the source. Keep the result as text, because a phone number is an identifier, and paste values before any upload.

Example
     A                  B             C
1  Mobile as written  Cleaned       Check
2  +91 98765 43210    9876543210    OK
3  098765-43210       9876543210    OK
4  9876543210         9876543210    OK
5  91-98765 43210     9876543210    OK
6  98765 4321         987654321     Check

B2: =RIGHT(SUBSTITUTE(SUBSTITUTE(SUBSTITUTE(A2," ",""),"-",""),"+",""),10)
C2: =IF(LEN(B2)=10,"OK","Check")

25. Addresses exported from an online form have line breaks inside the cells, and the upload file breaks. How do you clean them?

When someone presses Enter inside an address box, the cell holds a line break, character 10, and some systems read it as the end of a record, so one address splits into two rows or the upload fails. CLEAN removes the line break but joins the words, giving Shanti NiwasMG Road. Replacing character 10 with a space and then using TRIM keeps the words apart. For a quick fix on a whole column, open Find and Replace, press Ctrl and J in the Find box to enter a line break, and replace it with a space. Some exports also carry character 13, so replace that too.

Example
A2: Flat 12, Shanti Niwas(line break)MG Road, Pune 411001

=CLEAN(A2)
  -> Flat 12, Shanti NiwasMG Road, Pune 411001
=TRIM(SUBSTITUTE(A2,CHAR(10)," "))
  -> Flat 12, Shanti Niwas MG Road, Pune 411001

26. The same customer seems to be entered twice under slightly different spellings. How do you find such duplicates?

Exact duplicates are easy to find, but here the names differ, so I match on a field that should be unique and is copied from a document, such as the mobile number, PAN or account number. COUNTIF on the mobile column flags every number that appears more than once: Suneeta and Sunita Sharma share one, and so do Ravi Patil and R. Patil. Then I compare each flagged pair with the source forms, because family members can share a phone, and merge or delete only with the supervisor's approval. Prevention is better: search the system by mobile number before creating any new record.

Example
     A               B            C
1  Name            Mobile       Flag
2  Suneeta Sharma  9876501234   Possible duplicate
3  Ravi Patil      9822012345   Possible duplicate
4  Sunita Sharma   9876501234   Possible duplicate
5  Imran Khan      9930011122
6  R. Patil        9822012345   Possible duplicate

C2: =IF(COUNTIF($B$2:$B$6,B2)>1,"Possible duplicate","")

27. Before you upload 3,000 rows into an ERP through its Excel upload template, what do you check?

First the template itself: the column order and headers exactly as the system expects, with no columns added or renamed. Then the data: mandatory fields filled, codes such as customer, item or branch codes already present in the system, dates in the format the template asks for, numbers without commas or text, and only values, no formulas. If the file must be saved as CSV and has Hindi or other non-English names, choose CSV UTF-8 so they are not garbled. Upload 10 rows first and look at them in the system. After the full upload, match the row count and the amount total with your file.

28. Excel keeps completing a value as you type, for example Mumbai Central when you meant Mumbai. Is AutoComplete helping or hurting?

Both. AutoComplete offers a value already typed in the same column as soon as the first letters match, which saves time on repeated text. The risk is that Enter or Tab accepts the offer, so a fast typist whose eyes are on the paper enters Mumbai Central where it should be Mumbai and never notices. When you want the shorter value, type it fully and press Delete before Enter to remove the suggested part. For columns with fixed values, a dropdown list is safer. If it causes more errors than it saves, switch off Enable AutoComplete for cell values under File, Options, Advanced.

29. You must put the same branch name in 200 cells, today's date in a column and the value from the row above in several rows. Which keys do you use?

To fill many cells at once, select them, type the branch name once and press Ctrl and Enter, and every selected cell gets it. For the date, press Ctrl and semicolon, which types today's date as a fixed value. Do not use the TODAY function for an entry date: it recalculates, so next week every record would show next week's date. To copy the value from the cell above, press Ctrl, Shift and the double-quote key, and to copy the top cell of a selection into the cells below it, press Ctrl and D. Then spot-check a few cells to make sure each took the right value.

30. You are setting up a Google Form for field staff to submit customer details. How do you stop bad entries at the source?

Make the key questions required and choose the question type that fits: a dropdown or multiple choice for fixed values such as state or product, and a date question for dates. Then add response validation to short-answer questions: a regular expression for the mobile number, such as ^[6-9][0-9]{9}$, which accepts 9876543210 and rejects 5876543210, 987654321 and +919876543210, and a number rule with a range for amounts. Write the error text in plain words so staff know what to fix. Responses land in a linked sheet; do any cleaning on a copy or a separate tab, never in the response columns.

31. Some rows you entered in a shared Google Sheet yesterday have changed or vanished. How do you find out what happened and fix it?

Do not type anything again yet. Right-click a changed cell and choose Show edit history to see who changed it and when. Cell edit history does not show deleted rows, format changes or changes made by formulas, so for vanished rows open version history and compare yesterday's version with today's. Restoring an old version undoes everyone's work since then, so it is usually better to copy the lost rows from the old version back into the current sheet. Then tell the supervisor and the person concerned. To prevent a repeat, ask for protected ranges and a separate tab for each operator.

32. The ERP accepted a supplier bill that had already been entered last month. How could that happen, and how do you stop it?

Duplicate warnings usually need an exact match on the vendor and the bill number, and some systems, such as SAP, can also compare the invoice date and amount. So the same bill gets through if one operator typed INV/245 and another INV-245 or just 245, if the date or amount was keyed differently, or if the supplier exists twice in the vendor master. To stop it, type bill numbers exactly as printed, including slashes and letters, and before saving search the vendor's bills by amount and date as well. Merge duplicate vendors through the master data owner, and if a repeat bill was already paid, tell accounts at once.

33. You are given OCR output from 2,000 scanned forms to correct. What kinds of errors do you expect, and how do you work?

OCR reads clean print well but makes predictable mistakes: 0 and O, 1, l and I, 5 and S, 8 and B, rn read as m, lost decimal points and broken words where a stamp or fold crosses the text. Handwriting comes out much worse. So I never trust a number field without looking at the image: I keep the scan and the extracted fields side by side, check every number, date and code against the image, and read names and addresses more quickly. If the tool marks low-confidence fields, I start with those. Faint or skewed scans go back for rescanning at a higher resolution.

34. What is document indexing in a digitisation project, and why is it checked so carefully?

Indexing means entering a few search fields for each scanned file, such as file number, name, department, document type and date, so the file can be found in the document system later. In projects such as digitising old office or land records, the index is often the only way to find a scan, because nobody reads the images to search. A wrong file number or date means the document is effectively lost, even though the scan is perfect. That is why index fields usually get a full check, not a sample: a second person compares each index entry with the file, and the page count is matched too.

35. A friend asks you to look up a patient's test result, or a caller says he is a customer and asks you to read out his account details. What do you do?

I politely refuse both. My access to the system is for my assigned work only, and opening a record without a work reason is misuse even if I never share what I see. Many systems also log every record a user opens, so it would show up. For the caller, I cannot confirm on the phone who he is, so I do not read out any details; I direct him to the official channel, such as the branch or the customer care number, where his identity is checked. If someone keeps pushing or offers money, I report it to my supervisor.

36. You find that 120 records you entered last week have the wrong month in the date. What do you do?

I tell my supervisor the same day, with the facts: which batch, how many records and how I found it. Then I find the full extent, for example by filtering the records entered under my user ID that week, because there may be more than the ones I noticed. I correct them the way the process allows, usually through an edit that keeps an audit trail or a correction file approved by the supervisor, never by quietly overwriting. I check whether any report or payment has already used the wrong dates. Finally I find the cause, such as a default month on the entry screen, and fix it.

37. A QC check of 150 of your forms, with 20 fields each, found 18 wrong fields spread across 15 forms. What are your field and form error rates?

Field error rate is wrong fields divided by fields checked: 150 x 20 = 3,000 fields, and 18 / 3,000 = 0.6%, so field accuracy is 99.4%. Form error rate is forms with at least one error divided by forms checked: 15 / 150 = 10%. Both matter. A client who sends back any form with an error feels the 10%, while a team measuring keying quality looks at the 99.4%. So first ask which measure your target uses. Then sort the 18 errors by type: if most are in one field, such as dates, the fix is specific, not just being more careful.

38. A client sends an Excel file for you to enter into your system. What do you check before you start?

First that it is the right file: the period, branch or batch the email mentions, and not a copy of a file already entered, which I check against the intake log. Then I look for anything that can hide data: hidden rows, columns or sheets, filters left on, and merged headers, so that the count I see is the real count. I compare the row count with the count in the client's email and query any difference. If Excel shows a security warning about macros, I do not enable them and I ask IT. Only then do I log the file with its record count and start entering.

39. Forms come with names in Hindi, but the system accepts only English letters. How do you enter them consistently?

First check what the process says. Often the rule is to copy the English spelling from the ID document, such as the English side of the Aadhaar or PAN card, when one is attached. If you must transliterate, the whole team should follow one set of rules, so the same person is not Shrivastava in one record and Srivastava in another, which creates duplicates and failed searches. Never correct or shorten a name yourself, and keep initials as written. Where a spelling is unclear, flag it instead of guessing. A short list of common surnames with the approved spelling helps new operators.

40. Each form needs about 220 keystrokes, you type at 9,000 key depressions per hour, and reading and moving between fields adds 15 seconds per form. How many forms can you do in a day?

Typing time per form is 220 / 9,000 of an hour, which is 88 seconds. Adding 15 seconds for reading and navigation makes each form 103 seconds, so about 35 forms an hour. In seven hours of real entry time that is about 244 forms. I would use this kind of working before accepting a target: if the target is 300, I would show the calculation and ask whether shortcuts, a better field order or default values can cut the 103 seconds, instead of simply typing faster and risking errors. Measure real speed over a week, not one good hour.

Ready to check yourself? Take the free Data Entry Operator mock test

Data Entry Operator interview questions: Hard questions (experienced)

20 questions

41. You are hiring 10 operators for a hospital billing project. How would you design the skill test?

I would test the real job, not just copy typing. Part one is a short typing test for a minimum speed and accuracy. Part two, the main part, is entry from 20 sample bills and registration forms into a practice screen or sheet, scored on fields per hour and field accuracy, with critical fields such as patient ID, amounts and dates weighted more. I would set the pass mark from what our current good operators score on the same test, not from a figure found online. All test documents use made-up patients, never real records, and every candidate gets the same material and time.

42. When is double-key entry worth its cost compared with checking a sample of records?

In double-key entry two operators key the same fields independently, the system compares the two versions, and a third person settles every mismatch against the source. Because two people rarely make the same mistake in the same field, it catches almost all random keying errors, but it nearly doubles the keying effort. So it is worth it where one error is costly or hard to find later: account numbers, amounts, ID numbers, and survey or exam data that feeds big decisions. For low-risk fields such as remarks or a second address line, checking a random sample is enough. Many teams mix the two, field by field.

43. QC checked a 5% sample of 4,000 records with 12 fields each and found 6 wrong fields. What does that tell you about the whole batch?

The sample is 200 records, or 2,400 fields, so the error rate is 6 / 2,400 = 0.25%, an accuracy of 99.75%. Fixing those 6 does not clean the batch: at the same rate, the other 45,600 fields still hold about 114 errors. With only 6 errors found the estimate is rough, as the true rate could be anywhere from about 0.1% to 0.5%, and it holds only if the sample was random across operators and days. Against a 99.9% target the batch fails, so check more or rework. If any error is in a critical field, such as an account number, check that field in all 4,000 records.

44. How do you score quality fairly when a wrong account number and a misspelt middle name each count as one error?

Use weighted errors. Agree the categories with the client or manager first: for example critical errors, such as a wrong account number, amount, ID number or date of birth, worth 5 points; major errors, such as a wrong address or phone number, worth 3; and minor errors, such as a misspelt middle name, worth 1. Compare operators on points per 1,000 fields checked. On the same number of fields, if Asha has two critical errors and one minor one, she scores 11 points; Vikram has nine minor errors and scores 9. On a plain count Vikram looks worse, but Asha's errors can send money to the wrong account. Publish the categories so everyone knows what matters most.

45. An auditor will pick 50 records your team entered last quarter. How do you prepare, and what will they check?

An auditor checks that each record matches its source and that the process was followed. For each sampled record we must show the source document or its scan, who entered and who checked it, with dates, and the approval and reason for any later correction. The auditor may also look at who had access to the system, whether leavers lost access on time, and whether documents were kept only as long as the policy allows. I prepare by keeping the audit trail complete every day, not in the week before, and by running our own test: pick five random records and see whether each can be traced in a few minutes.

46. Your team of 8 must enter 24,000 records in 10 working days, and two of them joined this week. How do you plan it?

The daily need is 24,000 / 10 = 2,400 records. If the six experienced operators do 350 a day and the two new joiners 150, that is 2,100 + 300 = 2,400, which leaves no buffer at all. QC rework of even 3% adds about 72 records a day, or 720 over the period. So I would raise this on day one with the numbers and agree a buffer: overtime on two or three days, a temporary operator, or a later date for part of the work. I would give the new joiners the simpler forms, check all of their work at first, and track the count against plan every day.

47. Management wants to pay data entry incentives on forms per hour alone. What problems do you see, and what would you propose?

Paying for speed alone rewards the wrong behaviour. Operators rush, pick the easy forms, skip their own checks and push errors onto the checker, so errors rise while the headline output looks better. I would propose that the incentive counts only forms that pass QC, with an accuracy gate, for example no incentive for a month below the agreed accuracy target. Forms should be weighted by complexity, so a 40-field account opening form counts for more than a 6-field change request, and work should be allocated in order, not picked. I would run it as a one-month trial and check that accuracy and rework did not get worse.

48. You find that a shared folder holding 20,000 scanned KYC forms has been open to everyone in the company for three months. What do you do?

Close it first, then report. I would restrict the folder at once to the people who need it, without deleting anything, and tell my manager and the company's security or data protection contact the same day under the incident process. IT can use the access logs to show who opened which files, which decides how serious it is. Some reporting may already apply: CERT-In's 2022 Directions require listed incidents, including unauthorised access to data, to be reported within 6 hours. From May 2027 the DPDP Rules 2025 add reporting without delay to the Data Protection Board and the people affected. The legal or privacy team decides. Then find out why the permissions were wrong and check other shared folders.

49. A client says the DPDP Act makes your entry accuracy a legal matter. Is that right, and what changes for your team?

Largely right, but if you key data for a client, the duty sits with the client. Under the DPDP Act 2023, the company that decides why and how data is processed is the data fiduciary. A BPO or contractor that keys data for it is usually a data processor working under a contract, so the fiduciary's duties reach the team through that contract and the client's procedures. Section 8(3) touches entry directly: where data will be used for a decision about the person, or shared with another fiduciary, it must be complete, accurate and consistent. Expect tighter accuracy targets and audit trails for corrections. Most duties under the Act and the DPDP Rules 2025 apply from May 2027.

50. Your team handles Aadhaar, PAN and bank details every day. Which controls would you insist on as team lead?

First, least access: each operator sees only the screens and fields the task needs, and full ID numbers are masked wherever the work allows, for example showing only the last four digits of an Aadhaar number. Second, no easy exits for data: USB ports and personal email blocked, printing limited, phones kept off the work floor where the client requires it, and files shared only through approved folders. Third, accountability: individual logins, no shared passwords, access logs reviewed, and access removed the day someone leaves. Fourth, people: training at joining and every year, a signed confidentiality agreement and a clear way to report mistakes without fear.

51. Two operators keyed the same 500 forms for double entry. How do you list every mismatch in Excel?

First make sure both sets are in the same order: sort each by form number and check that both have the same 500 form numbers, row for row. Then compare field by field. In the layout shown, the check column tests whether any field in the row differs between operator 1 and operator 2, and one SUMPRODUCT counts the differing fields in the whole batch. Form F102 has two digits of the mobile number swapped, and F103 has 7250 keyed as 7520. Filter the check column for Check, settle each mismatch against the paper, and record which operator was wrong. The plain comparison ignores letter case, so use EXACT where case matters.

Example
     A     B, C, D (operator 1)            F, G, H (operator 2)            I
1  Form  Name        Mobile      Amount  Name        Mobile      Amount  Check
2  F101  Sunil Rao   9876512345  4500    Sunil Rao   9876512345  4500    Match
3  F102  Kavita Jain 9823011122  12800   Kavita Jain 9823011212  12800   Check
4  F103  Arif Khan   9930044455  7250    Arif Khan   9930044455  7520    Check

I2: =IF(OR(B2<>F2,C2<>G2,D2<>H2),"Check","Match")
=SUMPRODUCT((B2:D4<>F2:H4)*1)    -> 2

52. A vendor says its OCR tool reads your forms with 98% accuracy. How do you test that claim before the company buys it?

First ask what the 98% measures. Character accuracy sounds high but hides a lot: if each character of a 12-character account number is right 98% of the time, the whole number is right only about 78% of the time, because 0.98 to the power 12 is about 0.785. What matters is field accuracy on your own forms. So run a pilot on 200 real forms, handwritten and printed, check every extracted field against the image, and report accuracy for each field, above all the critical ones. Also time the human review, because a tool that needs every field checked saves little. Decide on those numbers, not on the brochure.

53. The same error keeps coming back: dates of birth are entered in the application date field. How do you fix it for good?

Repeated errors usually come from the process, not only from people, so I look for the cause before retraining anyone. I would sit with two operators and watch them work. Common causes are a screen whose field order differs from the paper form, two date fields side by side with similar labels, or a cursor that jumps to the wrong field after a dropdown. The fixes are then specific: ask IT to reorder or relabel the fields, and add checks that catch the swap, for example that a date of birth for an adult account is at least 18 years ago and an application date falls in the last few weeks. Then track the error count for two weeks.

54. Payroll tells you that 40 employees had their salary bank details entered wrongly by your team. How do you handle it?

First limit the damage: confirm the 40 names, check each account number and IFSC code against the source documents and correct them with payroll the same day, before the next salary run. If any salary has already gone to a wrong account, payroll and finance must start recovery with the bank at once, because the money now sits in someone else's account. Then find the cause: unclear forms, one operator or one batch? For the future, I would add a second-person check on all bank fields, since they are critical, and suggest an account validation check before payment if the payroll system offers one.

55. Your ERP has built up about 1,200 duplicate customer masters over the years. How would you clean them up without breaking past records?

Plan it as a project, not a delete. First find the duplicates on fields that should be unique, such as GSTIN, PAN or mobile number, then review each group and pick one surviving code. Never delete a master that has invoices or receipts against it: block it or mark it inactive and record which surviving code replaces it, so old documents still open and reports can map old codes to new. Move open orders and balances only with the accounts team's approval. Then stop new duplicates: make a unique field mandatory, have the system warn on similar names, and allow only a few trained people to create masters.

56. A government department wants 5 lakh pages of old files scanned and indexed in six months. How would you plan the indexing and QC work?

Work out the daily flow first. At six working days a week, six months is about 150 working days, so the project needs about 3,333 pages scanned a day. If an average file has 50 pages, that is 10,000 files, or about 67 files indexed a day, which sets the size of the indexing team. Every index entry gets a full second check, because a wrong file number loses the document, while scan quality can be sampled. Keep a register of every file leaving and returning to the record room, since these are original government records. Fragile pages need careful handling, and the personal data in them must stay on project systems.

57. Your team must enter records in Hindi and English for a state scheme. What problems would you plan for?

First, fonts and encoding. Everyone must type in Unicode, for example with the Inscript layout and a font such as Mangal, because text typed in older fonts like Kruti Dev turns into meaningless English letters on another computer and cannot be searched. Second, exports: save CSV files as CSV UTF-8, or the Hindi turns into question marks. Third, names: agree how each name is written in both scripts, so one person does not appear twice. Fourth, skills: test each operator's Hindi speed separately, since many are fast in English but slow in Hindi, and split the work around it. A small pilot batch shows most of these problems early.

58. Fifteen operators and a client team will work in one Google Sheet of customer data. How do you set up access and protection?

Share the sheet only with named people or the company's domain, never with anyone who has the link, and give the client team view or comment access unless they must edit. In the share settings, turn off Allow editors to change permissions and share, so only the lead can add people. Give each operator a tab or block of rows, and use Protect sheets and ranges on everything else, including headers, formula columns and the dropdown lists. Ask everyone to use filter views. If the client requires it, stop viewers and commenters from downloading, printing and copying. Review the sharing list every month.

59. An operator's accuracy has slipped from 99.5% to 97% over the last month. How do you find out why and fix it?

I look at the evidence before talking to the person. The error log shows whether the new errors sit in one field, one form type or one time of day, and each points to a different cause: a new client form layout, a changed screen, unclear handwriting from one branch, eyesight or screen trouble, or tiredness from extra shifts. Then I meet the operator privately, show the pattern and ask what has changed for them. We agree a plan, such as a short refresher on that field, a full check of their work for two weeks and a daily review. I recognise the recovery, and escalate formally only if it does not come.

60. Clients send data to your team by email, WhatsApp and pen drives. What safer intake process would you propose?

One controlled entry point. I would propose a single team mailbox or a secure shared folder, or SFTP for large files, so files are not scattered across personal inboxes and phones. WhatsApp and pen drives should stop for personal data, because they leave copies outside company control. Files with personal data should be password-protected, with the password sent through a different channel. Every file is logged on arrival with sender, record count and date. After entry and the agreed retention period, files are deleted from the intake area and the deletion is logged. I would write this as a one-page procedure and agree it with each client.

Practice questions for November 2026

New every month

3 questions from the Data Entry Operator mock test, the way they appear in it. A new set comes on the 1st of every month. Pick your answer, then open it.

  1. FresherProofreading

    When proofreading a keyed letter, what should Gopal compare it with?

    • The original source document, line by line
    • Only the spell checker's suggestions
    • His memory of what the letter said
    • The previous letter sent to the same customer
    Show the answer

    Answer: The original source document, line by line. Proofreading checks the copy against the source, because the eye reads what it expects when working from memory. A spell checker misses wrong numbers, names and real words in the wrong place, so it cannot replace the source.

  2. Intermediate (1-5 years)Number and Date Formats

    How is Rs 3,45,60,000 written in crore?

    • Rs 34.56 crore
    • Rs 345.6 crore
    • Rs 3.456 crore
    • Rs 0.3456 crore
    Show the answer

    Answer: Rs 3.456 crore. 1 crore = 1,00,00,000 (ten million), so 3,45,60,000 / 1,00,00,000 = 3.456 crore. 345.6 is the amount in lakh, a hundred times the crore figure, which is the usual mix-up.

  3. Experienced (5+ years)Basic Maths for Data Work

    Three operators had the errors shown on different volumes. What is the overall error rate for the team?

    Op 1: 6 errors in 100 records
    Op 2: 8 errors in 400 records
    Op 3: 15 errors in 1500 records
    • 6%
    • 1.45%
    • 3%
    • 2%
    Show the answer

    Answer: 1.45%. Overall rate = total errors / total records = 29 / 2,000 = 1.45%. Averaging the three rates (3%) gives the small operator too much weight compared with the one who keyed 1500 records.

Take the full Data Entry Operator mock test

Ready to apply?

Browse the open jobs on CoreHRx. Create a free profile once and apply with it.