Goods-to-Person Picking

Problem class

In traditional person-to-goods picking, operators spend 50–70% of their time walking, not picking. Goods-to-person inverts this model, solving the "wasted travel time" problem by bringing inventory to the worker. This dramatically increases pick rate per operator-hour and reduces error rates.

Mechanism

Inventory is stored in a dense automated structure (cube-based grid, shuttle-served racking, or mobile-shelf pods). When an order requires an item, the system's execution software identifies the storage unit containing that SKU, dispatches a robot or shuttle to retrieve it, and delivers it to a stationary pick station. The operator (or robotic arm) picks the required quantity, confirms via scan or vision, and the storage unit is returned to the structure. Workstation sequencing logic ensures a continuous flow of totes/shelves to minimize idle time between picks. The system handles inventory allocation, storage unit prioritization (e.g., partially depleted totes first), and real-time load balancing across multiple pick stations.

Required inputs

• Order pick list with SKU, quantity, and priority
• Storage unit content map (which tote/shelf contains which SKUs)
• Pick station availability and operator assignments
• Item master data (dimensions, fragility, special handling)
• Real-time system status (robot positions, queue depths, station throughput)

Produced outputs

• Picked and confirmed order lines (scan-verified)
• Updated inventory ledger (decremented pick quantities)
• Throughput metrics per station (units/hour, orders/hour)
• Storage unit defrag triggers (consolidation of partially empty totes)
• Operator productivity and ergonomic metrics

Industries where this is standard

• E-commerce same-day fulfillment (high SKU count, each-pick)
• Online grocery fulfillment (multi-temperature zones)
• Electronics component distribution (small parts, high SKU density)
• Pharmaceutical wholesale (each-pick, high accuracy requirements)
• Fashion/apparel e-commerce (returns-heavy, high throughput needs)

Counterexamples

• Full-pallet-out operations (e.g., beverage distribution) where the picking unit is a full pallet handled by forklift — the infrastructure cost of a goods-to-person system cannot be justified when there is no "each-pick" problem to solve.
• Extremely low daily order volume (<200 orders/day) — the fixed cost and complexity of goods-to-person automation cannot reach ROI thresholds when a small team of pick-walk operators can handle the volume comfortably.

Representative implementations

• Ocado's Customer Fulfillment Centres using Hummingbird grid robots to present totes to 600+ pick stations at rates up to 1,200 bin presentations/hour per station.
• AutoStore installations across 1,100+ sites (PUMA, Gucci, Best Buy, Siemens) providing cube-based goods-to-person in facilities from 500 to 200,000+ bins.
• Amazon Robotics using Kiva-derived shelf-carrying robots (now evolved to Proteus, Sparrow, and Cardinal systems) in fulfillment centers globally.
• Exotec's Skypod system used by Decathlon, Uniqlo, and Carrefour for high-bay goods-to-person retrieval up to 12 meters.

Common tooling categories

Automated storage and retrieval system (ASRS: cube, shuttle, or pod-based) + warehouse execution system (WES) for station sequencing + pick-to-light or pick-by-vision workstation interface + storage unit content management system + station throughput balancer.