6 Common Data Quality Issues (and How to Fix Them)

Data quality issues are the silent killers of analytics projects. According to Gartner, poor data quality costs organizations an average of $12.9 million annually. Here are the six most common problems and how to eliminate them.

1. Trailing Whitespace

Invisible spaces at the end of values cause failed lookups, duplicate entries, and incorrect aggregations. A report showing “duplicate” customers may really just have whitespace differences that the human eye can’t see but every database engine treats as distinct.

Fix: Trim all string fields as the first step in any cleaning pipeline. This alone often resolves 30–40% of data quality issues before you even reach more complex rules.

2. Inconsistent Date Formats

Different source systems produce dates in different formats — MM/DD/YYYY, DD-MM-YYYY, YYYYMMDD, Unix timestamps, and even natural language. Merging data from multiple sources multiplies the chaos and makes cross-source analysis nearly impossible without normalization.

Fix: Normalize all dates to a single standard format at ingestion time. ISO 8601 (YYYY-MM-DD) is recommended for its sortability and unambiguous parsing across all modern systems.

3. Amount and Currency Inconsistencies

Financial data arrives with dollar signs, commas, parentheses for negatives, and varying decimal precision. "$1,234.50", "1234.5", and "(1,234.50)" represent the same magnitude but won’t join, sum, or compare correctly without normalization.

Fix: Strip currency symbols, remove thousand separators, and standardize decimal places. Apply rounding rules consistently across all monetary fields to avoid floating-point drift.

4. Case Sensitivity Mismatches

"[email protected]", "[email protected]", and "[email protected]" may refer to the same user, but case-sensitive systems treat them as three distinct entities. This is especially common with email addresses, product codes, and categorical labels.

Fix: Normalize case based on the column’s purpose. Emails should be lowercased; proper names can be title-cased; product codes should match their canonical source format precisely.

5. Null and Empty Value Confusion

Null (no value), empty string, "N/A", "null", "None", and "-" are all used interchangeably to indicate missing data. Analytics tools handle each differently, often producing different row counts for the same query depending on the placeholder convention.

Fix: Choose a single representation for missing values and apply it consistently across all columns. Document what null means in your context so downstream consumers can reason about it correctly.

6. Outlier and Range Errors

Extra zeros in quantity fields (10000 instead of 100), impossible dates (February 30th), and negative ages are common data entry errors that skew averages, medians, and any analysis that relies on statistical distributions.

Fix: Define valid ranges for each field. Flag outliers for human review rather than silently accepting them (which distorts analysis) or discarding them (which loses data).