CSV date handling
Why CSV Percentages Change—and How to Validate Rates Safely
Percentages usually look wrong after CSV import because CSV stores text values, not a universal percentage type or display format. One source may represent twelve percent as 12%, another as 0.12, and another as 12. The importing tool may remove the percent sign, infer a number, or apply percentage formatting again. Preserve the original, confirm the source convention and metric definition, compare known rows, and normalize a working copy to one documented representation. Recalculate rates from their numerators and denominators when possible; do not multiply or divide every value by 100 until the column's meaning is verified.
Short answer
Percentages usually look wrong after CSV import because CSV stores text values, not a universal percentage type or display format. One source may represent twelve percent as 12%, another as 0.12, and another as 12. The importing tool may remove the percent sign, infer a number, or apply percentage formatting again. Preserve the original, confirm the source convention and metric definition, compare known rows, and normalize a working copy to one documented representation. Recalculate rates from their numerators and denominators when possible; do not multiply or divide every value by 100 until the column's meaning is verified.
Upload
Bring the CSV you already use.
Understand
Start with the brief and evidence.
Act
Ask, edit, filter, and export.
A practical starting point
Percentages usually look wrong after CSV import because CSV stores text values, not a universal percentage type or display format. One source may represent twelve percent as 12%, another as 0.12, and another as 12. The importing tool may remove the percent sign, infer a number, or apply percentage formatting again. Preserve the original, confirm the source convention and metric definition, compare known rows, and normalize a working copy to one documented representation. Recalculate rates from their numerators and denominators when possible; do not multiply or divide every value by 100 until the column's meaning is verified.
Concrete examples
Example 1: Percentage formatting is applied twice
A discount column stores 0.15, meaning fifteen percent. A user changes the value to 15 and then applies percentage formatting, producing 1500%. Comparing the raw file with a known invoice shows that the original decimal fraction was correct; only the display format needed adjustment.
Example 2: Two exports use different scales
One campaign file stores conversion rates as 0.08, while another stores them as 8. Appending the files without a source-specific conversion makes one campaign appear one hundred times stronger. The analyst retains a source column, confirms each export contract, normalizes a working copy, and recalculates rates from conversions and visits.
Example 3: A simple average gives the wrong overall rate
Outlet A records 1 conversion from 10 visits, or 10%. Outlet B records 90 conversions from 1,000 visits, or 9%. Averaging 10% and 9% gives 9.5%, but the combined rate is 91 divided by 1,010, about 9.01%. The dashboard uses the combined counts rather than an unweighted mean.
Example 4: A value above 100% is valid but poorly labelled
A monthly growth field shows 125%. It means the new value is 125% higher than the prior value, not that a share of a fixed whole exceeds its maximum. The team verifies the formula, changes the label to monthovermonthgrowthpercent, and keeps the distinction from percentoftotal clear.
Example 5: Missing denominators make a rate unverifiable
A supplier scorecard contains only ontimerate = 96 and no shipment count. The value may mean 96%, but its weighting and sample size cannot be checked. The report labels the rate as source-supplied and avoids combining it with other suppliers until shipment counts or a documented aggregation rule are available.
Examples are illustrative and are not customer results.
Common questions
What is the safest way to work with this CSV problem?
Percentages usually look wrong after CSV import because CSV stores text values, not a universal percentage type or display format. One source may represent twelve percent as 12%, another as 0.12, and another as 12. The importing tool may remove the percent sign, infer a number, or apply percentage formatting again. Preserve the original, confirm the source convention and metric definition, compare known rows, and normalize a working copy to one documented representation. Recalculate rates from their numerators and denominators when possible; do not multiply or divide every value by 100 until the column's meaning is verified.
Keep exploring
See what your own CSV is saying.
Try the browser workspace with a real file and keep the public guide open when you need help.
Start free