Mobile Robots
Cost per Pick: Goods-to-Person Robots vs Manual Picking
What a robotic pick actually costs against a manual one, with the labour, capital and maintenance terms laid out and the volume threshold where automation wins.
Manual picking costs roughly 0.12 to 0.35 euro per line in a conventional walk-and-pick operation. A goods-to-person robot system costs 0.08 to 0.22 euro per line once amortised, but only above about 2.5 million lines per year. Below that volume the capital charge per line dominates and manual picking is cheaper, regardless of how good the robots are.
What a manual pick costs
A conventional picker walking the aisles achieves 60 to 140 lines per hour depending on travel distance, pick face density and order profile. Travel is typically 50 % to 70 % of the shift, which is the entire reason goods-to-person exists.
- At a fully loaded labour cost of 28 euro per hour and 100 lines per hour, the direct labour cost is 0.28 euro per line.
- Add supervision, equipment and errors at roughly 10 % to 20 %, giving 0.31 to 0.34 euro.
- A dense, well-slotted operation reaching 140 lines per hour drops to about 0.20 euro before overheads.
What a robotic pick costs
The picker stays at a station and the inventory comes to them, so travel disappears from the picker's time and moves into the robot fleet. Station productivity of 200 to 400 lines per hour is the normal range, so labour per line falls by a factor of two to three. The capital charge replaces it.
| Term | 1.5 M lines/yr | 3 M lines/yr | 6 M lines/yr |
|---|---|---|---|
| System capital, 39 robots plus stations and software | 3.4 M | 3.9 M | 5.2 M |
| Amortisation over 7 years | 486 k/yr | 557 k/yr | 743 k/yr |
| Capital per line | 0.324 | 0.186 | 0.124 |
| Labour at 280 lines/h, 28 euro/h | 0.100 | 0.100 | 0.100 |
| Maintenance and spares, 4 % of capital | 0.091 | 0.052 | 0.035 |
| Energy and software licence | 0.018 | 0.012 | 0.009 |
| Robotic total per line | 0.533 | 0.350 | 0.268 |
| Manual comparison | 0.310 | 0.310 | 0.310 |
| Verdict | manual wins | roughly level | robots win |
What the cost per line leaves out
- Storage density. Goods-to-person systems typically store 30 % to 60 % more stock in the same footprint, and avoided building cost can exceed the picking saving entirely.
- Accuracy. Pick error rates commonly fall from roughly 1 in 300 lines to 1 in 2,000 or better with light-directed station picking, and the cost of a wrong delivery is far above the cost of the pick.
- Labour availability. In markets where picking positions cannot be filled, the comparison is not cheaper against dearer, it is possible against impossible.
- Ergonomics and turnover. Walking 12 km a shift produces attrition. Station work does not.
- Peak flexibility. Manual operations scale with temporary staff in hours. Robotic systems scale with hardware in months.
Point five is the strongest argument against full automation in seasonal businesses, and it is why hybrid designs, robots for the stable base load and manual picking for the peak, are common in practice.
The hybrid design that usually wins
Very few distribution centres automate every line. The pattern that survives audit is to split the catalogue by movement and let each half use the cheaper method.
| Segment | Share of SKUs | Share of lines | Method |
|---|---|---|---|
| Fast movers | 5 to 15 % | 50 to 70 % | Pick face or pallet, manual, no robot needed |
| Medium movers | 25 to 35 % | 25 to 35 % | Goods-to-person, best robot case |
| Slow movers and long tail | 55 to 70 % | 5 to 15 % | Goods-to-person for density, not for speed |
The counter-intuitive result is that automating the fastest movers is often the worst investment. A SKU picked 400 times a day sits in a dedicated pick face two steps from the packing bench, so there is almost no travel to remove. The robots earn their keep on the medium and slow segments, where a picker would otherwise walk 40 m for a single line.
Frequently asked questions
What does a manual pick cost?
Roughly 0.12 to 0.35 euro per line, driven mainly by lines per hour. At 100 lines per hour and 28 euro per hour fully loaded, direct labour alone is 0.28 euro per line before supervision and errors.
At what volume does robotic picking win?
Typically above about 2.5 million lines per year. Capital per line falls with volume while labour per line stays flat, so the crossover is set by throughput rather than by the technology.
How much faster is a picker at a station?
Two to three times, typically 200 to 400 lines per hour against 60 to 140 walking the aisles, because 50 % to 70 % of a conventional picker's shift is travel.
Does cost per line capture the whole benefit?
No. Storage density gains of 30 % to 60 %, error rates falling from around 1 in 300 to 1 in 2,000, lower attrition and reduced building cost often matter more than the picking cost itself.
What about seasonal peaks?
Manual operations scale with temporary staff within hours; robotic systems scale with hardware over months. Hybrid designs, automating the stable base load and handling peaks manually, are the usual answer in seasonal businesses.
Sources
- World Robotics, service robots report seriesInternational Federation of Robotics, logistics robot deployment volumes
- Robotics at NISTNational Institute of Standards and Technology, robot performance measurement and test methods
- ISO 3691-4, safety requirements for driverless industrial trucksInternational Organization for Standardization, constraints on fleet speed and therefore throughput