Customer photos at PDP, UGC moderation rules for body-type + size?

Customer photos drive conversion in lingerie more than in any other category, shoppers buying intimates desperately want to see real bodies wearing the SKU, not just professional models. But unmoderated UGC for intimates is a brand-safety landmine.

Moderation rules I default to:

• Photo must show the product fitting, not full-body shots, not bikini-only shots without the SKU visible.
• Face is optional, many customers want to share fit-check photos without their face. Allow head-cropped or face-blurred uploads.
• Size + body-type tag required, customer self-reports their size at upload. AI moderation (Hive Moderation, AWS Rekognition, or a custom CV model) auto-assigns body-type bucket (petite, mid, plus, post-partum, post-mastectomy) for the gallery filter.
• Manual moderation queue, every UGC photo passes through a human reviewer before going live. Rejection reasons: nudity beyond the SKU coverage, unclear size, blurry, brand visible from a competitor, low quality.
• Incentive, $10 store credit for an approved photo. Increases UGC volume 5-8x vs unrewarded.

Tooling: Foursixty for Instagram-sourced UGC (auto-pulls tagged posts, you moderate, they auto-render on PDP). Yotpo Visual UGC for direct uploads + Instagram. LoudCrowd for ambassador-program-driven UGC with attribution.

Architectural note: customer photos must be served via CDN (Cloudflare Images / Cloudinary) with size + body-type tags in the metadata so Hyvä PDP can filter the gallery client-side without bloated payloads. Self-hosted images break LCP on mobile.