A photograph can be powerful evidence, but it is not automatically an accurate answer. Without a defined standard, different representatives photograph different angles, shelves and moments, making comparisons unreliable.
Define what each image must prove
| Purpose | Recommended capture approach |
|---|---|
| Shelf availability | Whole relevant bay plus close-up of the SKU location if needed |
| Planogram placement | Consistent straight-on framing with the reference section visible |
| Display installation | Whole display and context showing it is on the selling floor |
| Price or label compliance | Legible ticket next to the correct product |
| Pack-out before/after | Same location and comparable framing at both times |
| Damaged packaging | Close-up showing the condition, with SKU context |
Do not demand photographs where retailer policies, privacy restrictions or safety requirements prohibit them.
Coach for evidence quality
A usable image is sufficiently sharp and bright, covers the relevant area, and is linked to the correct store, visit and task. Ask reps to explain gaps that cannot be photographed, such as retailer refusal or heavy customer traffic. More images do not always mean better evidence.
Audit a sample, not just totals
Have a supervisor periodically check a sample against three questions: Does the photo support the stated answer? Is the store and SKU context identifiable? Would a different reviewer reach the same conclusion? Track the proportion accepted and the reasons for rejection, then adjust field guidance.
When using image recognition
Computer vision may assist with counting visible products, facings or promotional elements, but accuracy depends on camera angle, packaging variations, occlusion, model coverage and validation. Start with a human-reviewed sample in each relevant category, measure false positives and false negatives, define confidence thresholds and retain a correction route.
Do not automatically treat a model score as ground truth. A correctly detected product can still be a local exception to the planogram, and a product outside the frame may be present elsewhere.
Privacy and worker trust
Avoid unnecessarily capturing customers or staff, particularly identifiable faces. Use clear photography policies, access controls and retention rules. Location and time metadata may be useful to establish visit context but should not become an unannounced employee-monitoring mechanism.
Where FieldKo fits
Photos taken in FieldKo are attached to the visit and the survey answer they belong to. Smart Camera checks each shelf photo for lighting, clarity and framing as it's taken, and Vision AI is available as a paid add-on. Have a person review any photo that leads to a significant decision.
See how FieldKo supports this →Further reading
Next: Turn observations into action
Reviewed October 26, 2026. Examples, scores and thresholds are illustrative, not benchmarks. See the research and sources library for independent references.