Autograding

In pricing terms, autograding removes one of the largest sources of grade variance between sellers: human subjectivity. When two sellers grade the same device differently, one will systematically underprice or overprice relative to market reality. Platforms like Back Market increasingly reward sellers with low return rates, and those rates correlate strongly with grading consistency. For pricing intelligence, autograded inventory is more comparable across competitors because it follows a repeatable methodology rather than individual technician judgement.

Autograding systems typically assess cosmetic condition through image recognition using standardised lighting and camera angles. The device is photographed from multiple sides, and the AI model identifies and scores visible defects, producing a condition classification without a human making a subjective call. The accuracy of current systems varies by defect type: screen scratches and housing dents are detected reliably, while fine micro-scratches on black housings under certain lighting conditions remain more challenging. Operators deploying autograding systems should benchmark accuracy against their existing human grading outcomes before fully replacing human review.

The commercial benefit of autograding compounds over time. Consistent grading improves seller score, better seller score improves search placement, and better placement enables sustaining prices slightly above the market floor without losing volume. The cost of grading inconsistency therefore goes beyond individual returns: it affects the platform relationship and pricing power in ways that accumulate over months.