Image Optimization Report Workflow
Short answer: BatchWebP creates a CSV report from completed items in the current browser queue. Each row records the source filename, output filename, original and output bytes, rounded saved percentage, original and output dimensions, output format, and applied quality. The report documents what BatchWebP produced; it does not measure page speed, visual quality, Core Web Vitals, CDN behavior, or bytes delivered by a production website.
Report format verified: August 2026 against the current CSV export implementation.
Which fields are included?
| CSV field | Meaning | Important limitation |
|---|---|---|
| Source file / Output file | Original local name and generated output name. | Names do not prove where a file was published. |
| Original bytes / Output bytes | Blob sizes before and after the current conversion. | Transfer compression and CDN variants are not included. |
| Saved percent | Rounded percentage calculated from original and output bytes. | A negative value is possible when output is larger. |
| Original / Output dimensions | Decoded width and height before and after scaling. | Does not record CSS display size. |
| Output format | Actual generated extension recorded by the queue. | Production CDNs can later serve another format. |
| Quality | Applied browser encoder quality, including target-size search results. | Quality numbers are not visually equivalent across formats or browsers. |
How is saved percentage calculated?
For each completed file, BatchWebP calculates (1 - output bytes / original bytes) × 100 and rounds the result to a whole percent. The batch summary uses the corresponding aggregate byte totals. This is a file-size comparison, not a performance score. If an output is larger than its source, the saved percentage can be negative and should not be hidden.
How to create a defensible handoff report
- Group files by purpose so hero images, product images, and thumbnails are not evaluated against one arbitrary rule.
- Choose output format, dimensions, quality, or Target KB settings.
- Convert the batch and inspect representative images in the comparison view.
- Resolve failures and obvious outliers before exporting.
- Download the ZIP and CSV from the same completed queue so filenames correspond.
- Record the conversion date, BatchWebP version or commit, browser version, and intended publishing context outside the CSV.
- After deployment, measure the real page separately with browser or field-performance tools.
What does the CSV not prove?
- It does not prove that an image looks acceptable; visual review remains necessary.
- It does not show which responsive candidate a real visitor downloaded.
- It does not measure LCP, CLS, INP, cache hits, network latency, or total page weight.
- It does not preserve a permanent audit trail inside BatchWebP; the queue is cleared when the tab closes.
- It is not a compliance certificate or independent optimization audit.
Spreadsheet safety
CSV cells are quoted, embedded quotes are escaped, and filenames beginning with =, +, -, or @ receive a leading apostrophe to reduce spreadsheet-formula injection risk. Review untrusted filenames before importing reports into automated systems.
Run a batch and export its CSV report, or read how Target KB results are selected.