4.2Order picking (preparation)
The essentials in 30 seconds
- Travel dominates: over half of picking time goes walking — slotting, sorted routes and batching attack it first.
- Batch by profile: order by order, by lots/waves, by zones — choose by order size and count.
- Build quality at source: clear labelling, location + item scan, proportional checks — and a measured error rate.
Running mini-case — Nadia times her picks
After tidying up, Nadia times her storekeeper for a morning in Algiers. Verdict: of ten minutes per order, six are walking and searching. Lists follow order sequence, not aisle sequence. She sorts the first list by address, tests a multi-bin cart for eight orders, and logs each error by cause.
Continued from case 4.1 — Nadia moves from slotting to routing and batching.
1. Why travel dominates
Order picking — pick, consolidate, check, pack — is the largest warehouse labour cost. Time analysis shows over 50% goes on moving around, not on picking.
Three levers: slotting (4.1), pick path (sequence lines by location address, with no zigzag) and batching orders.
Simply sorting the pick list by address already creates a clean aisle-by-aisle route — a direct dividend of addressing.
2. Batching methods
Order by order (pick by order): simple, few mix-ups, but one tour per order.
Batch/wave: one tour picks N small orders together, sorted at a dedicated station or on a multi-bin cart — the king of e-commerce where orders are many and small.
By zone: each picker stays in their zone, orders move and consolidate — suited to large high-throughput sites.
3. Quality: measure and attack errors
One error costs disproportionately (reshipment, return, trust). Track the error rate / lines (target: a few errors per 10,000 lines) and classify causes: SKU, quantity, omission.
Graduated response: unambiguous labelling facing each SKU, separation of lookalike SKUs, location + item scan, proportional final check (weighing, visual, scan).
Pick assists (pick-to-light, voice picking, goods-to-person): for volumes that justify them, after a sound organisation — automating disorder produces faster disorder. And the final packing is the last image the customer sees.
🔎 Concrete example — From single-order picking to e-commerce batch picking
An online shop prepares 80 orders/day of 2 lines: switch to batches of 8 orders per tour with an 8-bin cart, lists sorted by location.
Result: tours divided by 6, errors divided by 3 after adding item scanning.
Lesson: organisation (routing + batching) comes before technology investment.
Section vocabulary
- Picking
- Picking order lines from stock, then consolidating, checking and packing; the largest warehouse labour cost.
Simple example: Nadia picks, groups, checks then packs each imported order.
What it is not: confusing picking with simply moving cartons.
Reflex: what share of my time is pure walking?
- Pick by order / batch picking
- Order by order (simple, one tour per order) vs batch/wave: a single tour picks several orders, sorted afterwards — a major gain on many small orders.
Simple example: Nadia combines 8 small orders in one tour then sorts by bin.
What it is not: doing one tour per order when they are two lines each.
Reflex: are my orders small and many?
- Pick path
- Sequencing picks along a logical route (aisle by aisle, with no backtracking); simply sorting by location address achieves it.
Simple example: Nadia sorts her list by address: one direction, no zigzag.
What it is not: picking in line arrival order.
Reflex: is my list sorted by address?
- Pick error rate
- Errors per lines picked; cut by clear labelling, location and item scanning, and proportional final checks — each error costs disproportionately.
Simple example: Nadia counts her errors per 1,000 lines and types item, quantity, omission.
What it is not: never measuring and fixing case by case.
Reflex: what is my rate, and my top cause?
- Pick-to-light / voice picking
- Picking aids: light displays showing location and quantity, or hands-free voice guidance; for volumes that justify them, after organisation is sound.
Simple example: Nadia will test voice when her volumes justify it.
What it is not: buying the technology before sorting the lists.
Reflex: is my organisation sound before automating?
In practice — Speed up and secure your order picking
- Time a typical pick: share of time spent travelling, searching, picking, packing.
- Sort your lists by location address (no-backtrack route) — immediate gain, zero cost.
- If orders are small and many: test batching (5-10 orders per tour, multi-bin cart) and measure.
- Measure your error rate and attack the top cause: neighbour labelling, then item scanning if volume justifies it.
Ready-to-say script — launching batch + scan (1 min)
"No more one tour per order. We leave with eight bins, list sorted by address, we scan the location then the item. If in doubt, piece or carton, we stop and ask. Goal: fewer metres, zero neighbour mix-ups."
Say it at tour launch, cart in view, then observe one full tour.
Copy-paste template — batch tour card
DATE / TOUR: _______________
GROUPED ORDERS (count): _______________
LIST SORTED BY ADDRESS (yes / no): _______________
LOCATION + ITEM SCAN (yes / no): _______________
ERRORS (item / quantity / omission): _______________
TOTAL TIME: _______________VSE / small budget
No scanner: lists sorted by address, DIY compartment cart, scale to check quantities.
Algeria SME
Power cuts and temps: label with unit clearly visible, similar neighbours separated, paper lists pre-sorted.
Small-order e-commerce
Batch 5 to 10 orders, sort at a dedicated station, weight check before packing.
High volume / large site
Zone picking with consolidation, then systematic scan and proportional checks by stake.
✓ Best practices
- Show the pick unit on the label and the list: piece vs box vs carton is the No. 1 mix-up.
- Physically separate close variants (colours/sizes): never side by side.
- Weigh instead of counting for quantities: the counting scale is fast and reliable.
✗ Common pitfalls
- Checking 100% without judgement: costly and dulling — prefer proportional checks by stake.
- Investing in robots before sorting lists by location: you automate waste.
- Half-empty carton: you pay to ship air and you create breakage.
Key takeaways
- The picker spends over half their time TRAVELLING: attacking this reserve brings 30-50% with no technology.
- Single-order (simple), batch (spread over N orders), zones (large sites): choose by order profile.
- One error costs disproportionately: labelling, item scan, proportional checks and a MEASURED rate.
- Organisation first, technology next; packing is the last image given to the customer.
48h Challenge — apply tomorrow
Sort 1 list by address and test 1 batch tour of 5 orders. Measure time and errors, then note the first cause to attack (label, neighbours, scan).
- I can explain by-order, batch and zone picking with their right use.
- My lists are sorted by address, no zigzag.
- My error rate is measured and my top cause handled.
Frequently asked questions
How can I concretely cut errors?
Measure and classify (SKU/quantity/omission), fix neighbour mix-ups (facing label, separation of lookalikes), roll out item scanning, and check proportionally (systematic on high stakes, sampling elsewhere, weighing for omissions). Each error is traced to its cause and feeds improvement.
When should I automate: pick-to-light, robots?
After organisation + light computerisation (simple WMS). Heavy systems (conveyors, goods-to-person) pay off for high steady volumes, scarce labour, tight space and a stable horizon — in full TCO including rigidity and breakdowns.
What is the first source of productivity in picking?