How do you make an A+ performance chart without exaggerating the data?

Start with the evidence table

Obtain the approved test record from the product owner. Identify the units, conditions, compared items and scope of the result. Do not infer data from a product photograph or ask the generator to supply reasonable numbers. If the claim cannot be traced to real evidence, do not present it as a performance chart.

In an illustrative battery comparison, duration measured under one workload is not a universal runtime promise. Keep the workload and compared configurations with the values. The example is a planning scenario, not a result from an CartVisual or customer test.

Document the visual scale

A chart should explain its scope without requiring access to the planning document. If the data is too detailed for a small image, provide a readable native table or accompanying description.

ONS guidance starts bar and area charts at zero because their extent represents the value. Line and scatter charts encode position differently and can use a cropped scale where justified and clear. Do not treat “label the axis” as permission to make a cropped bar comparison. Use an appropriate chart type if a zero-origin bar cannot show the needed distinction clearly.

Separate the chart from atmosphere

In A+ product pages, create a product illustration or restrained explanatory background from actual references. Build the numerical chart with an appropriate editor and place it into the approved module outside CartVisual. CartVisual does not export a data-bound chart whose bar heights are guaranteed to match your spreadsheet.

W3C’s complex-image guidance is useful when supplying a textual explanation of a chart. Amazon’s A+ design documentation provides destination context; it does not certify this proposed chart or remove the need for evidence.

Proof the finished chart against the record

Check each value, series name, axis boundary and test-condition note. Review the chart at normal phone size and confirm the accompanying page text conveys the same result. A truncated chart or missing unit can mislead even if the file is technically sharp.

Use comparison provenance for sourcing and diagram description for accessibility. Keep the exact source table and final chart together. Stop release if any visual comparison cannot be explained from the approved evidence.

Fictional battery data: A = 8 hours and B = 10 hours under the same invented workload. The difference is 2 hours; relative to A it is 25%, not threefold runtime. No batteries were tested.

Keep the source table, chart-type choice, scale, units and relevant conditions with the finished module. A textual value table helps access, but it does not excuse a misleading graphic. Neither an A+ template nor a generated chart certifies the test evidence.