Operational evidence
Observed results, with the boundaries in view.
We publish figures that can be tied to documented checks and runs. We do not present them as market averages, general error rates, or promises for every store.
Estimator dataset
Real checks, with one eligible sample per distinct store.
The estimator uses the latest eligible result for each store in the dataset, so repeated checks of the same storefront do not receive extra weight.
What the dataset means
The figures describe exactly what was observed in the eligible runs used by the estimator.
- The dataset contains one completed run and eleven partial runs.
- The product count is the number actually analyzed, not the volume initially requested for a scan.
- Aggregate findings are not a market average and do not predict the result of an individual store.
Published coverage
67 public catalog checks
PROVEN uses a public catalog of 67 standard checks to identify observable problems without reducing the whole catalog to a synthetic score.
The number and descriptions of checks come from the Capabilities page. You can inspect each check in Explorer.
Case study · image optimization
A real catalog, from evaluation to final verification.
This evidence describes a real run that formed the basis for the image-optimization feature. The figures are reused directly from the same public source shown on the feature page.
- 01
Observed situation
Images from a real catalog were evaluated. Not every analyzed image met the conditions for optimization.
- 02
Controlled decision
Eligible files are handled separately; an optimized file may replace the original only if it is smaller and passes the required checks.
- 03
Verification outcome
The final check reported no eligible images left unprocessed in that run.
These figures describe that real run. Savings depend on the images in each store.
See image optimizationInterpretation
What these figures support — and what they do not.
You can conclude
- PROVEN has been exercised on real catalog data, not only demonstration examples.
- Documented runs exist where the result and final verification stay in the same workflow.
- You can check feature availability and limits on the public Capabilities page.
You cannot conclude
- that these values are a general issue rate for e-commerce;
- that every store will have the same finding distribution or cost;
- that one image run guarantees the same saving for other catalogs.
Check your own catalog
Public evidence is the starting point. Your store is the relevant test.
Try PROVEN checks publicly available Shopify store information without Admin access and without modifying the catalog.
Request access to PROVEN The app is not yet in the Shopify App Store. A request is not an installation or subscription.