720411Handpick for Fashion – Implementing Narrative Search with an Attribute-Value ONtology (AVON) to Enable Sequential, Natural Language Product Navigation on Any DeviceClosedGRD Development of PrototypeInnovate UKWhen it comes to shopping online for nuanced products, such as fashion, shoppers are often unable to articulate and discover what they seek — especially on mobile, where browsing many products at once is tedious or impossible. Most major fashion retailers’ mobile websites are simply shrunken desktop sites. However, desktop search-and-filter tools frustrate shoppers when applied on mobile — as small screens curtail shoppers’ ability to quickly browse. Total online UK fashion sales are £7.1B, growing at 16% YoY (Mintel 2013). Although Mobile accounts for 23% of this total spend, the visit-to-purchase conversion rate on mobile is half that compared to Desktop (IMRG 2013). Low conversion rates on mobile cost UK fashion retailers £1.6B last year in missed sales (Verdict, 2013). To close this conversion rate gap, we are developing Handpick — a mobile sales platform with sequential, natural language search to enable low-latency, non-historical personalisation in real-time. We call the approach Narrative Search: whenever shoppers notice something they like, they can find more items like it, but better, in one or more specific ways (e.g. more casual, more appropriate for work). Narrative Search relies on AVON (Attribute-Value ONtology), our learning engine that links discrete product attributes with scalar ‘shopper values’ e.g. casual, work-appropriate, elegant, smart. Sponsored in part by a TSB Proof of Concept grant, we have already shown that Narrative Search, as compared to conventional search-and-filter tools, significantly increases shopper satisfaction and decreases product exchanges on a major retailer’s site. Handpick mobile will complete the automation of AVON for all fashion categories and enable rapidly deployment for any retailer. The Handpick / AVON infrastructure is extensible and can be used in the future to support mobile retail in other complex categories such as real estate, wine and furniture.