Common Errors in Bank Statement Conversion
By Sandra Vu
Bank statement conversion isn't perfect. Understanding common errors helps you catch and fix them before they cause problems.
Why Errors Happen
Conversion errors occur due to:
- OCR limitations - Character recognition mistakes
- Layout variations - Unexpected statement formats
- Data complexity - Multi-line entries, merged cells
- Quality issues - Poor scans, low resolution
Even the best converters have edge cases.
Common Error Types
1. OCR Character Misreads
Similar-looking characters get confused:
| Original | Misread As | Example |
|---|---|---|
| 0 (zero) | O (letter) | $1,O00.00 |
| 1 (one) | l (lowercase L) | $l,234.56 |
| 1 (one) | I (uppercase i) | $I,234.56 |
| 5 | S | $S00.00 |
| 8 | B | $B00.00 |
| $ | S or 5 | 5100.00 |
Impact: Amounts become text, formulas break, totals wrong.
Detection: Sort by amount—text values sort differently than numbers.
Fix: Find and replace, or manual correction.
2. Decimal Point Shifts
The decimal gets misplaced:
| Original | Error | Impact |
|---|---|---|
| $1,234.56 | $123,456 | 100x too large |
| $1,234.56 | $12.3456 | 100x too small |
| $1,234.56 | $1234.56 | Missing comma (might be OK) |
Impact: Totals wildly incorrect.
Detection: Compare sum to statement total. Large discrepancy indicates decimal issues.
Fix: Identify affected transactions, correct decimal position.
3. Missing Transactions
Entire transactions don't appear:
Causes:
- Page break in middle of transaction
- Multi-line description not recognized
- Unusual formatting skipped
- Image or watermark covering data
Detection: Count rows and compare to original transaction count.
Fix: Manually add missing transactions or reprocess with different settings.
4. Merged Rows
Multiple transactions combined into one:
Original:
01/15 AMAZON -$49.99
01/15 NETFLIX -$15.99
Converted:
01/15 AMAZON NETFLIX -$65.98
Impact: Loses transaction-level detail.
Detection: Row count lower than expected, amounts larger than typical.
Fix: Split manually or reprocess.
5. Split Rows
One transaction becomes multiple:
Original:
01/15 AMAZON MKTPL*2X9K7YT PURCHASE -$49.99
Converted:
01/15 AMAZON MKTPL*2X9K7YT
PURCHASE -$49.99
Impact: First row has no amount, second row has no date.
Detection: Rows with missing dates or amounts, row count higher than expected.
Fix: Merge rows or clean up incomplete entries.
6. Date Format Issues
Dates misinterpreted or malformed:
| Original | Error | Cause |
|---|---|---|
| 01/15/26 | 15/01/26 | DD/MM vs MM/DD confusion |
| 01/15/26 | 01/15/1926 | Wrong century |
| Jan 15 | Text, not date | Not recognized as date |
Impact: Sorting fails, date filters don't work, formulas error.
Detection: Sort by date—wrong formats sort incorrectly.
Fix: Convert text to dates using Excel functions or find/replace.
7. Sign Errors (Debit/Credit)
Amounts have wrong positive/negative:
| Transaction Type | Should Be | Error |
|---|---|---|
| Purchase | Negative | Positive |
| Deposit | Positive | Negative |
| Fee | Negative | Positive |
Impact: Running balances wrong, totals inverted.
Detection: Running balance doesn't match statement.
Fix: Identify pattern (all debits wrong? all credits?) and apply correction.
8. Missing or Extra Columns
Column structure doesn't match expected:
Missing columns:
- No balance column
- Debit/Credit combined when should be separate
Extra columns:
- Description split across multiple columns
- Phantom empty columns
Detection: Column headers don't match data, data in wrong columns.
Fix: Reorganize columns, merge or split as needed.
Verification Checklist
After every conversion, verify:
Count Check
☐ Transaction count matches original ☐ No duplicate rows ☐ No missing rows
Total Check
☐ Sum of debits matches statement ☐ Sum of credits matches statement ☐ Net change matches statement
Balance Check
☐ Opening balance correct ☐ Closing balance correct ☐ Running balances track correctly
Spot Check
☐ First transaction correct ☐ Last transaction correct ☐ 3-5 random transactions correct
Prevention Strategies
Use Quality Source Documents
- Download PDFs directly from bank (not scans of paper)
- Ensure full pages (no cropping)
- Avoid password-protected files if possible
Choose the Right Tool
- Use bank statement-specific converters
- Prefer tools with high accuracy claims
- Test on sample before bulk processing
Verify Immediately
- Check results before using data
- Don't assume accuracy
- Build verification into workflow
Fixing Errors Efficiently
Small Number of Errors
Fix manually in Excel:
- Identify error
- Correct value
- Verify running balance
Pattern-Based Errors
Use find/replace:
- Replace "O" with "0" in amount column
- Fix date formats globally
- Correct sign on all debits
Systematic Errors
Reprocess with different settings or tool:
- Try different converter
- Adjust OCR settings
- Process pages separately
Summary
Common bank statement conversion errors include OCR misreads, decimal shifts, missing or merged transactions, and date format issues. Always verify conversions by checking transaction counts, totals, and balances against the original statement. Build verification into your workflow to catch errors before they propagate into accounting systems.

About Sandra Vu
Sandra Vu is the founder of Data River and a financial software engineer with experience building document processing systems for accounting platforms. After spending years helping accountants and bookkeepers at enterprise fintech companies, she built Data River to solve the recurring problem of converting bank statement PDFs to usable data—a task she saw teams struggle with monthly.
Sandra's background in financial software engineering gives her deep insight into how bank statements are structured, why they're difficult to parse programmatically, and what accuracy really means for financial reconciliation. She's particularly focused on the unique challenges of processing statements from different banks, each with their own formatting quirks and layouts.
At Data River, Sandra leads the technical development of AI-powered document processing specifically optimized for financial documents. Her experience spans building parsers for thousands of bank formats, working directly with accounting teams to understand their workflows, and designing systems that prioritize accuracy and data security in financial automation.