ShelfMind

ShelfMind — Store-level 3D shelf digital twins built from retailer planogram files. Customers upload JDA/PSA planograms (v8-v10+), product master files and store lists; the platform builds an exact spatial model of each shelf — X/Y/Z capping, orientation and volume-fill 'to 100% accuracy to the PSA file' — then computes facings, share of shelf, adjacency and compliance metrics, revenue-velocity heatmaps, and derived analytics such as the 'Halo Effect' between adjacent SKUs, 'visual suffocation' where high-potential items are blocked by low-value bulk, dormant versus high-velocity shelf zones,…

ShelfMind offers (Alternative, Fundamental) — Not published. Bounded from the product: volume scales with stores x bays under management, and a single retailer tenant can span thousands of stores with tens of bays each — so one mid-size chain yields hundreds of thousands of 3D bay models per planogram version. The platform supports multi-planogram comparison within a store (e.g. 24ft versus 4ft layouts) and version-tracked history, which multiplies stored models per bay.

3D scene reconstruction and spatial-reasoning model training using exact planogram-to-geometry pairs (a rare source of metrically correct retail 3D ground truth, versus inferred depth from photos); adjacency and contextual-influence models learning cross-SKU halo effects; shelf-zone margin attribution and space-elasticity models; planogram-compliance detection as supervised classification; retail computer-vision systems needing a canonical intended shelf geometry to register observed images against; synthetic shelf scene generation for rendering-based training and augmentation; and store-level assortment recommendation using joined demographic and velocity data

The data is with 3 years of history.

Coverage spans US, UK; Application Software, Food Retail; alternative, fundamental; equities.