Define rework before counting it
A correction brings an output back to the approved brief. A change introduces a new instruction. If the brief requested a natural shadow and the file has none, that is a correction. If the buyer later chooses no shadow, that is a change. Both take time, but combining them hides the cause.
Also separate technical handling from image editing. Renaming a delivery, downloading it again and matching views to SKUs can consume staff time even when no pixels need changing.
A worked example
The following figures are hypothetical planning assumptions, not Rupato pricing or measured customer results.
| Activity | Calculation | Time |
|---|---|---|
| Review corrections | 120 images × 3 minutes | 6 hours |
| Clarify instructions | 12 exchanges × 10 minutes | 2 hours |
| Reconcile and rename files | One monthly total | 2 hours |
| Total | 10 hours × £25 internal hourly cost | £250 |
At the same volume and workload for twelve months, that is £3,000 of internal handling. It is not a forecast if your work is seasonal. It is not a saving you can claim simply by changing suppliers. You would need to measure how much of that work actually disappears.
Do not count the same minutes twice. If correction review is already included in a wider staff-time total, use one total or subtract the overlap.
Compare complete alternatives
For each option, add editing charges, platform fees or minimums, extras, transfer costs where relevant and expected internal handling. Use the same product mix and delivery requirements. A simple cutout and reflective jewellery retouching are not interchangeable units.
Keep rush fees and changes of brief visible. State whether quoted amounts include tax and use a consistent currency. An annual estimate should show its volume assumptions and the date of the quotation.
Keep a short log for the next batch
Record the product or batch, reason for intervention, minutes spent and whether it was a correction, change or handling task. Include the reason, such as missing reference, wrong crop or mismatched filename. After a representative period, group by cause rather than by who complained.
Then test one improvement. An annotated reference may reduce repeated questions; a delivery manifest may reduce reconciliation. Measure the next comparable batch and keep the extra cost of the improvement in the calculation. That gives you a decision grounded in your operation rather than a promised percentage.
