A percentage improvement is interpretable only when the starting value, comparison conditions, sample count, raw results, and calculation are clear.
The baseline determines the meaning
A percentage improvement is a relative change, normally calculated as the difference between a new value and a baseline value, divided by the baseline. Without the baseline, the percentage does not reveal the absolute size of the change. A large percentage can result from a very small starting value, while an important absolute difference can produce a modest percentage when the baseline is large.
Direction also matters. An increase may be desirable for one measure and undesirable for another. The claim must therefore name the measured characteristic, its units, the original value, the new value, and whether higher or lower is preferable.
The comparison must be like for like
A percentage is not a fair product comparison if the two values were obtained with different configurations, test methods, conditions, durations, or calculation rules. The comparison object must also be identified: it could be an earlier version, another product, a control, or a different operating condition. Each interpretation is materially different.
Raw values and the formula allow readers to reproduce the percentage and spot choices such as rounding, selective endpoints, or a denominator that makes the result appear larger.
Samples define how much the result represents
The sample count and selection method show whether the figure reflects one observation or a broader set. Results can also vary within each comparison group, so group averages alone may hide overlap, extremes, or instability. The NIST Engineering Statistics Handbook notes that information from a sample supports a bounded decision about a lot; it should not be treated as complete knowledge of all units.
A responsible improvement claim therefore states the baseline, comparison value, units, samples, conditions, raw results, and calculation. If those elements are unavailable or conditions are unequal, reporting the observed values without an improvement percentage is more informative. No SEWARD-specific percentage can be established without a matched comparison report.
Can a mathematically correct percentage still be misleading?
Yes. The arithmetic may be correct while the baseline is unusually small, the conditions are unequal, the samples are unrepresentative, or the absolute difference is practically minor.