Warehouse management a complete guide to improving efficiency and minimizing costs in the modern warehouse

1. 1Set the supervision span first. The source's optimum is one supervisor to 12-15 staff. Count staff per supervisor on every shift. Above 15, supervision weakens. Below 12, cost rises and the supervisor drifts into doing tasks, which gives job insight and a progression path but lowers overall effectiveness.
2. 2Fix the process before you add technology. The source warns that automating a bad process makes it quicker, not more efficient. Document the correct procedure for each task (receiving, picking, put-away) and confirm staff follow it before you buy a system.
3. 3Assign ownership by task and by area. Under 5S, give each individual responsibility for specific tasks and for their own work area. Run 5S in the correct order, and treat safety as part of it.
4. 4Match the instruction method to the workforce. If staff come from abroad or have limited reading ability in the working language, hire bilingual supervisors and consider voice-directed processes. Confirm that every person can read and understand safety instructions.
5. 5Match hours to demand to counter Parkinson's Law (work expands to fill the time allowed). In low-volume periods, use annualized hours or zero-hour contracts so paid hours follow workload. Flexible shifts (twilight, school-hour, spouse-aligned, variable length, student workers) widen the labour pool and add motivation.
6. 6Identify training needs explicitly and tie them to accuracy. Record errors in a goods-received non-compliance report and a manual record of work. Use the error and productivity data to decide who needs retraining and on which procedure.
7. 7Measure accuracy alongside speed. Speed versus accuracy is a named trade-off, so track productivity and error rate together. Benchmark picking against the WERC best-in-class picking measures, and price errors using the cost-of-errors comparison of pick methods.
8. 8Handle discipline and retention as part of the same system. Use consistent staff discipline, safe and comfortable conditions, and compliant contracts to keep experienced operatives, since they are harder to source in an ageing, tight labour market.

Illustrative application of the source's rules; the staffing numbers are assumed, not from the book. A warehouse runs 40 operatives on one shift with one supervisor. That is 1:40, well above the 12-15 optimum. The manager moves to three supervisors, about 13 staff each, which sits inside the range. Part of the workforce is foreign staff, so the manager makes one of the new supervisors bilingual and checks that everyone can read the safety instructions. Voice-directed picking is considered for the same group. Volumes are low on some days, so work stretches to fill the shift under Parkinson's Law. The manager introduces annualized hours so paid hours follow demand, and adds a twilight shift to attract staff who need flexible hours. Each operative is assigned a named work area under 5S. Errors are logged in the non-compliance report so retraining targets the procedures that actually fail.

• If a supervisor has more than 15 staff, then add supervisory capacity or split the team. If fewer than 12, then expect higher cost and a supervisor pulled into task work, and accept that only when the aim is to develop that supervisor.
• If you want to buy automation or a new picking technology, then first confirm the manual process is correct and followed. Automating a bad process only speeds it up.
• If the workforce includes foreign staff or limited-language readers, then add bilingual supervisors, consider voice-directed work, and verify understanding of safety instructions.
• If volumes are low and work is stretching to fill the hours, then move to annualized or zero-hour contracts so labour matches demand.
• If you can't recruit or retain operatives, then offer flexible shift patterns before accepting the vacancy, since the work is seen as unglamorous and experienced staff are scarce.

• Treating technology as the fix for accuracy. The tell: error rates stay high after a system goes in because the underlying process was never corrected.
• Running supervisors at the wrong span. The tell: above 15 staff, errors go uncaught; below 12, supervisors are picking and packing and costs climb.
• Ignoring language and comprehension. The tell: foreign staff misread instructions and safety rules, and errors or incidents cluster in that group.
• Letting fixed hours absorb slack volume. The tell: output per paid hour falls on quiet days because work expands to fill the shift.
• Measuring speed only. The tell: throughput looks good while non-compliance reports and the cost of errors rise, because speed was traded against accuracy without anyone tracking the trade.

• ▸Team / shift: name the team and shift under review.
• ▸Staff-to-supervisor ratio: headcount divided by supervisors; compare with 12-15.
• ▸Procedure adherence check: list the tasks where the documented procedure is followed and where it is not.
• ▸Task and area ownership: record the named person responsible for each task and work area (5S).
• ▸Language / comprehension gaps: note who needs bilingual support, voice direction, or safety-instruction checks.
• ▸Hours versus demand: compare paid hours with workload; flag Parkinson's Law slack and consider annualized or zero-hour hours.
• ▸Error log: count and categorise entries from the non-compliance report and manual record of work.
• ▸Training needs identified: link each recurring error to a specific retraining action.
• ▸Retention levers: record the shift flexibility, conditions, or discipline actions agreed, and the owner.

1. 1Compute the OSHA total recordable incidence rate monthly: (total injuries and illnesses x 200,000) / hours worked by all employees. Use hours from payroll, not headcount estimates. Compare the result with the passage benchmarks: 5.3 per 100 employees for US warehousing versus 3.8 for the national average.
2. 2Add a second, simple measure: days since the last reportable accident. Display it on site, and reset it only on a reportable event so it never turns into a vanity number.
3. 3Classify every injury and near miss by the passage's cause categories: overexertion (45%), contact with equipment (31%), falls and trips (19%), transport (3%), harmful substances (1%). Add the OSHA-cited causes: forklift misuse, improper stacking, missing PPE, lockout/tagout failures, fire safety gaps, repetitive motion. Rank by frequency and by severity separately.
4. 4Track severity alongside frequency. Record days away from work and flag any case over 31 days (35% of US warehouse incidents reached this). Record injury type, because strains, sprains and muscle tears made up 42% of injuries.
5. 5Break incidents down by weekday and hour. The source data shows a peak on Mondays and between 8 am and 12 pm. Test whether your own site shows the same pattern, then target supervision, briefings and pacing at the peak window.
6. 6Run a risk assessment on the top-ranked causes. Separate the hazard (anything that may cause harm) from the risk (the likelihood and seriousness of harm). Apply controls unless their cost in time, effort and money is 'grossly disproportionate' to the risk reduction.
7. 7Fix the highest-ranked cause with a design control before a behavioural one. Examples are height-adjustable operator platforms and work tables, good lighting, and totes presented ergonomically. These address overexertion and repetitive motion at source.
8. 8Re-measure after each control. Compare the incidence rate, severity and cause mix before and after, and write down which control moved which category.

The passages give the US warehouse baseline: 16 deaths in 2011 and 31,200 non-fatal injuries. Overexertion caused 45% of those injuries, contact with equipment 31%, and falls and trips 19%. Thirty-five per cent kept staff away for more than 31 days, and 42% were strains, sprains or muscle tears. The rate of 5.3 incidents per 100 employees exceeded the 3.8 national average. A manager reading this would conclude that nearly half of injuries come from lifting and handling, and that most are soft-tissue injuries with long absences. The first control is therefore ergonomic: workstations like Dematic's RapidPick, with height-adjustable platforms and tables, rake-adjustable screens and totes presented at a comfortable level, which target overexertion and repetitive strain directly. The manager calculates the site's own OSHA rate, checks whether incidents cluster on Mondays and before noon, and re-measures the rate and the share of overexertion cases after the change. (The passages give no before-and-after figures for RapidPick, so the result is something you must measure yourself.)

• If your recordable incidence rate is above the 3.8 national benchmark, treat the site as a high-risk operation. Start with the top two causes by volume before spreading effort across every hazard.
• If overexertion or strains and sprains lead your data, prioritise ergonomic redesign of workstations and handling over more training.
• If incidents cluster on Mondays or in the 8 am to 12 pm window, adjust briefings, staffing and supervision to that window before changing anything else.
• If a control's cost is grossly disproportionate to the risk reduction, document the reasoning and choose a cheaper control. If it is not grossly disproportionate, implement it.
• If days since the last reportable accident is long but near misses are rising, treat the streak as unreliable and investigate the near misses.

• Tracking only the headline count of injuries, not the rate per hours worked. The tell: the figure rises with headcount or overtime and nobody can compare the site against the 3.8 or 5.3 benchmarks.
• Ignoring severity. The tell: the dashboard shows frequency only, while a third of cases (35% nationally) involve more than 31 days away.
• Spreading effort evenly across all hazards. The tell: no ranked cause list, and controls aimed at harmful substances (1%) while overexertion (45%) is untouched.
• Over-complicating the risk assessment. The tell: lengthy documents, no controls actually implemented, and simple, cheap measures left undone.
• Treating the days-since-accident counter as the outcome. The tell: incidents go unreported to keep the number running, and near misses disappear from the records.

• ▸Period and total hours worked by all employees: enter payroll hours for the month.
• ▸Total injuries and illnesses: count recordable cases.
• ▸OSHA incidence rate: (injuries x 200,000) / hours; compare to 3.8 and 5.3.
• ▸Days since last reportable accident: current count and prior record.
• ▸Cause breakdown (%): overexertion, equipment contact, falls/trips, transport, harmful substances, other.
• ▸Severity: number of cases with more than 31 days away, and share of total.
• ▸Injury type: share that are strains, sprains or muscle tears.
• ▸Timing pattern: incidents by weekday and by 8 am to 12 pm window.
• ▸Near misses logged: count and top cause.
• ▸Control actioned and re-measure date: the change made against the top cause and when you will check its effect.

1. 1Define the ratio in one line before measuring anything: output (lines picked, orders shipped) divided by input (labor hours). Pick one output unit per activity and keep it fixed. The source lists this as a 'traditional productivity measure' (Ch. 13).
2. 2Split labor hours by activity, following the book's process chain: receiving and put-away, pick preparation, picking, replenishment, value-adding services, despatch, returns. Count indirect activities (Ch. 7) in the hours. Leaving them out inflates the ratio.
3. 3Choose measures deliberately. Ch. 13 treats 'how to choose the right performance measures' as its own step. Select a small set tied to the pressures the book names: productivity, accuracy, cost, lead time, service. Pair each productivity measure with an accuracy or service measure, because speed versus accuracy is a named trade-off.
4. 4Fix the process before the technology. The book states these processes 'need to be in place in all warehouses prior to any thoughts of introducing new technology' and that 'automating a bad process might make it quicker but certainly doesn't make it more efficient.' Walk each flow (receiving, pick preparation, layout, replenishment) and remove waste first.
5. 5Work the levers the book assigns to productivity: pick preparation and pick-area layout (Ch. 4), the picking strategy (picker-to-goods versus goods-to-picker, Ch. 5), and the pick method (paper, label, voice, barcode, light, Ch. 6). Change one lever at a time so you can attribute the shift in the ratio.
6. 6When you compare picking methods, include the cost of errors (Ch. 6). A method that raises lines per hour but raises error cost has not improved productivity.
7. 7Benchmark and score. Compare against external benchmarking (Ch. 13) and carry the productivity ratio into a balanced scorecard or integrated model, so it sits next to cost and service measures instead of standing alone.
8. 8Check resourcing against the measure (Ch. 11). Use activity-based costing (Ch. 12) to see which activities drive cost, and aim effort at those.

The source gives no numeric case, so this uses its one explicit principle with hypothetical figures labeled as such. A manager sees picking productivity lag and is tempted to buy a faster picking technology. Following the book, she first measures the ratio: lines picked per picking labor hour, with replenishment and other indirect hours counted separately. Walking the process, she finds long travel from poor pick-area layout and incomplete pick preparation. She fixes those first. Say the ratio rises from 50 to 60 lines per hour (illustrative numbers, not from the source). Only then does she compare pick methods such as voice, barcode or pick-to-light, adding the cost of errors to each. The book's warning stands behind this order: automating a bad process makes it quicker, not more efficient. She also notes that technology reduces staff and improves efficiency but at a cost, and that not every operation can afford it.

• If the ratio is weak and the process is undocumented or unstable, then fix receiving, pick preparation, layout and replenishment before buying technology.
• If a method or change raises output per labor hour, then check the accuracy and cost-of-errors measures before declaring a gain, because speed trades against accuracy.
• If budget is limited, then take process and method improvements first. The book says all warehouses can become more efficient 'irrespective of budget'.
• If you propose automation, then weigh the staffing and efficiency gains against the investment, because not all operations benefit or can afford it.
• If one productivity number is driving decisions, then add paired measures (hard and soft, scorecard or integrated model) so cost and service trade-offs stay visible.

• Counting only direct picking hours. The tell: productivity looks strong while replenishment, value-adding and other indirect work is unaccounted for and total cost does not fall.
• Chasing speed at the expense of accuracy. The tell: lines per hour rises but error, return and rework volume rises with it.
• Automating before fixing the process. The tell: new equipment runs faster but travel, queuing and rework patterns are unchanged and cost per line does not improve.
• Tracking too many or poorly chosen measures. The tell: the team cannot say which number drives which action, and the measures conflict.
• Changing several levers at once. The tell: the ratio moves but nobody can say whether layout, pick method or staffing caused it.

• ▸Activity measured: name the process (e.g. picking, put-away, despatch) from the chain.
• ▸Output unit: state the single unit counted (lines picked, orders shipped).
• ▸Direct labor hours: record hours spent on the activity in the period.
• ▸Indirect labor hours: record replenishment, value-adding and other indirect hours attributed to it.
• ▸Productivity ratio: output divided by total labor hours.
• ▸Paired accuracy/service measure: record the error or service figure that sits alongside the ratio.
• ▸Cost of errors: record the estimated cost of errors in the period.
• ▸Lever changed: name the single change (layout, pick preparation, pick method, strategy).
• ▸Result and next step: record the ratio and paired-measure change, then note whether to keep, revert or benchmark.

1. 1Fix the two denominators first. Pick total warehousing cost as a percentage of sales and cost per order shipped (or per pallet moved, for pallet operations like Fredericks). Compute both for the same period from the same cost base, so a change in one can be read against the other.
2. 2Split the cost base using the types-of-costs logic in the book's warehouse costs chapter. Separate fixed building and racking cost, labour, energy, equipment and shared overhead. Use activity-based costing rather than a flat traditional allocation where several products or customers share the space, so you see which activity drives each cost.
3. 3Fix the process before the technology. The source is blunt: 'Automating a bad process might make it quicker but certainly doesn't make it more efficient.' Confirm that receiving, put-away and picking are stable and measured before you price any automation or racking change.
4. 4Attack storage density against the building constraint. In the Fredericks case, a size-limited building had to hold at least 13,000 pallet spaces. Compute cost per pallet space under each racking option (mobile racking, high-bay automated storage, buffer and sequencing). Fixed building cost spreads over more spaces as density rises.
5. 5Test automation on volume and dependency. Automated stores are 'very economical for customers if volume related' and best attached to a high-volume production plant. They depend wholly on design and WMS software for capability, capacity and speed, and they demand high pallet presentation standards. Include the low manpower and energy benefits, but also the cost of failing to meet those standards.
6. 6In temperature-controlled sites, run the energy list as a cost programme. Start with product entering at the correct temperature, then improved plant design and operational management and maintenance. Then work down the store-management items: layout and temperature flows, raising refrigeration evaporating temperature (a potential 11 per cent or more saving), lowering condensing temperature, seasonal adjustment, splitting cold-store and blast-freezer systems, variable-speed drive fans, and door opening design and operation.
7. 7Work the tariff as well as the consumption. Shift energy loads to cheaper periods, accept being cut off at peak load, or be paid to have supply curtailed for a fixed timed period. Record each as a separate cost line so the saving is visible.
8. 8Check tax and cash-flow structure where goods are imported. A customs warehouse with duty and VAT deferment gives a further six weeks of cash flow benefit, and goods exported outside the EU carry no duty or VAT. Cost the duty management system and the AEO route, which waives the financial guarantee, against that benefit.
9. 9Re-measure monthly and record the trade-off. Cost versus service and storage capacity versus speed of retrieval are named trade-offs in the book. Log any service change next to each cost change so a saving is not bought by silently degrading service.

Fredericks Dairies, the UK's leading ice cream manufacturer, needed to expand its warehouse. The brief was at least 13,000 pallet spaces in a building with imposed size limits, able to move 600 pallets in an eight-hour shift at -28 °C. The throughput target works out to 75 pallets per hour (600 / 8). RediRack supplied an automated pallet storage, retrieval and sequencing solution with a pallet buffer and sequencing system, and raised storage density by up to 90 per cent. Treat the building as a fixed cost: 1.9 times the pallet spaces in the same footprint means building cost per pallet space falls to roughly 53 per cent of its former level (1 / 1.9). The source text also reports an annual running cost reduction, but the figure is cut off in the supplied passage, so do not quote one. The method to copy: set the pallet-space and throughput requirements first, express cost per pallet space and per pallet moved, and score each racking or automation option on those two numbers. The source also notes that high-density automated storage reduces air circulation and therefore energy use, which feeds the same cost lines.

• If the operation is high-volume and attached to, or contracted with, a production plant, then evaluate automated storage. If volume is low or variable, or you lack WMS and pallet-presentation discipline, then do not.
• If processes are unstable, inaccurate or unmeasured, then fix them before buying technology, because automating a bad process makes it quicker, not more efficient.
• If the building size is fixed and pallet demand is higher, then compare density options by cost per pallet space, since fixed building cost spreads over more spaces.
• If the site is a cold store, then work the no-capital energy items first (product entry temperature, door management, seasonal adjustment, evaporating temperature) before plant investment.
• If an energy tariff allows load shifting, peak cut-off or paid curtailment, then check whether the operation's cold-store thermal behaviour can tolerate it before committing.

• Reporting cost as a percentage of sales only. A sales rise masks rising cost per order; the tell is a flat percentage alongside climbing unit cost or falling throughput.
• Buying automation to fix poor processes. The tell is a new system installed on top of unchanged exceptions, rework and manual workarounds, with the cost per order not falling.
• Ignoring software and presentation dependency in automated stores. The tell is capacity or speed shortfalls traced to WMS design or pallets outside the design pallet gauge.
• Treating energy as a fixed overhead. The tell is a tariff never reviewed and refrigeration settings (evaporating or condensing temperature, seasonal adjustment) never changed, and doors left open during loading.
• Cutting cost without recording the trade-off. The tell is cost per order improving while service levels, accuracy or lead times quietly deteriorate.

• ▸Period and scope: state the dates and the site or customer set covered, so cost and sales use identical boundaries.
• ▸Total warehousing cost: sum building, labour, energy, equipment and overhead for the period.
• ▸Net sales for the period: the denominator for cost as a percentage of sales.
• ▸Orders shipped or pallets moved: the denominator for cost per order or per pallet.
• ▸Cost as % of sales and cost per order: calculate both; note the direction versus last period.
• ▸Cost breakdown by type (fixed, labour, energy, equipment, shared): show each as a share of total, using activity-based allocation where space is shared.
• ▸Pallet spaces available and cost per pallet space: record the density option under test and its effect on fixed cost per space.
• ▸Energy levers actioned: list each (entry temperature, evaporating temperature, door management, load shifting, etc.) with status and measured saving.
• ▸Process readiness check: note whether receiving, put-away and picking are stable before automation is considered.
• ▸Trade-off log: record any change in service, speed or accuracy that accompanies each cost change.

1. 1Fix the definition before you measure. An order is perfect only if all five attributes pass together: correct item, correct quantity, correct condition, on time, correct documentation. Write a pass/fail test for each attribute and agree it with the customer. The source passages list the five criteria but do not give formulas, so your own definitions are the standard you hold yourself to.
2. 2Score at order level, not line or unit level. One failed attribute fails the whole order. This is what makes the perfect order figure lower than any single accuracy figure. Report each attribute's failure rate beside the composite so you can see which one drags it down.
3. 3Map each attribute to the process step where it is lost. The source treats the warehouse as a chain: receive, process orders, replenish, add value, despatch. Item and quantity errors usually trace to receiving and picking. Condition errors trace to handling, storage and loading. Timeliness traces to order processing and despatch. Documentation traces to despatch paperwork and returns. Log every failed order against one step.
4. 4Use the source's evidence tools at the origin of errors. Start with the goods-received non-compliance report, so inbound errors are recorded at the dock. Then review the cost of errors and the cost-and-accuracy comparison of pick methods (paper, label, voice, barcode, RFID, pick/put to light). Compare your pick error rate with the best-in-class picking benchmark (WERC).
5. 5Pair hard measures with soft ones. Hard measures are counted, such as OTIF and perfect order percentage. Soft measures are customer perception. Review both, because a customer can be unhappy while your hard figure looks acceptable, and the reverse.
6. 6Benchmark and then place the result on a balanced scorecard. Compare your figures with WERC metrics or a peer or internal benchmark. Put service beside cost, productivity and people measures so that a service gain bought with unsustainable cost or overtime is visible.
7. 7Price each service improvement before you commit. The source includes a 'cost of service improvement' view and states that managers must improve service 'at least cost'. Estimate what each extra point of perfect order costs and stop where the marginal cost exceeds the value to the customer.
8. 8Fix the process before you automate it. The source warns that 'automating a bad process might make it quicker but certainly doesn't make it more efficient.' Stabilise the failing step manually, confirm the error rate falls, and only then consider a technology change.

The source passages give no OTIF or perfect order figures, so this example is illustrative and uses invented numbers. A warehouse audits 200 despatched orders against the five criteria. Item and quantity pass on 190, condition on 196, timeliness on 188, and documentation on 194. Most orders pass most tests, yet only 171 pass all five, so the perfect order rate is 85.5%. Each failure is logged against the process step where it occurred. Late orders cluster at despatch. Item and quantity errors cluster at picking, where paper lists are used. The manager compares the cost of errors under paper picking with the cost and accuracy comparison of other pick technologies. First, however, she checks that the paper process is being followed, since the source says a poor process should not be automated. She reports the composite rate beside the five attribute rates and a customer-perception score, so the next review shows which fix moved the figure.

• If the perfect order rate is well below the individual attribute rates, then the errors are scattered across attributes. Rank attributes by failure count and fix the largest first.
• If errors originate in receiving (goods-received non-compliance reports show a pattern), then fix inbound checks and supplier feedback before changing picking, because picking cannot correct what was received wrongly.
• If the pick error rate stays high after the process is stable and followed, then compare pick technologies on cost and accuracy. Do this only after the process is fixed.
• If hard measures look good but customers complain, then add or review soft measures such as customer perception, because the hard measure is missing something the customer values.
• If the next gain in service costs more than the customer values it, then stop at the current level and hold it, as the service/cost trade-off requires.

• Reporting accuracy at line or unit level and calling it OTIF or perfect order. The tell: the headline figure is 98–99% while customers still report problems and the order-level audit shows far lower.
• Measuring the outcome but never tracing it to a process step. The tell: the monthly figure is reviewed, yet no failed order is tied to receiving, picking, loading or despatch.
• Buying technology to fix service. The tell: a new pick system is installed while the underlying process is unstable and the error rate barely changes.
• Leaving out documentation and condition. The tell: the order is delivered correct and on time, yet invoices or paperwork are wrong or goods arrive damaged, and the scorecard shows it as a success.
• Chasing service with no cost view. The tell: service metrics rise while overtime, labour hours and cost per order climb with no cost-of-improvement analysis.

• ▸Audit period and sample size: dates and number of despatched orders reviewed.
• ▸Order ID and customer: reference for each audited order.
• ▸Item correct (Y/N): matches what was ordered.
• ▸Quantity correct (Y/N): matches the order.
• ▸Condition acceptable (Y/N): undamaged and fit for use on arrival.
• ▸On time (Y/N): delivered within the agreed window.
• ▸Documentation correct (Y/N): paperwork and invoice accurate and complete.
• ▸Perfect order (Y only if all five Y): composite result; total passes divided by sample size gives the rate.
• ▸Failure step: receiving, picking, replenishment, value-adding, despatch or returns.
• ▸Corrective action and owner: the fix, who owns it, and the date to re-audit.

1. 1Measure the whole cube, not the floor. Record usable height, rack and aisle volume, and the share filled by stock. The source points at 'the cube of the building', as in the KNAPP hanging-garment sorter, whose rails run above the work and storage areas. Empty air above racking counts as unused capacity.
2. 2Start before the dock. Compare container loading options for the space they cost. Palletizing can take up to 10 per cent of container space for the pallets alone. Pallet clearance height, and pallets that cannot be stacked, reduce utilization further. Price this against the handling benefits: less damage, fewer people, faster loading, and smaller bay space.
3. 3Test slip sheets as the alternative. They are about 2 cm thick and sized to the unit load, and they raise the loading cube. They also avoid FAO ISPM 15 wooden-pallet issues and are easier to clean. The cost is a special forklift attachment at both despatch and receiving warehouses, so confirm both ends can buy one before switching.
4. 4Decide fixed or random locations per product group. Fixed slots let pickers memorize locations and pick faster. When a product is out of stock its slot stays empty and pallet storage utilization drops significantly. Random slotting puts each pallet in the most efficient available slot. Keep fixed slots only where picking speed justifies the empty-slot cost.
5. 5Slot by velocity and ergonomics for carton storage. Put the fastest movers in the middle row of shelving, so pickers avoid bending and stretching. Put slower movers on the lowest and highest shelves. Density is only useful if the pick face stays quick to work.
6. 6Check each location against its constraints before putting stock into it. The source lists pallet location size, rack weight capacity, current pick-face status, and distance from receiving and despatch. Hazardous and high-value items need appropriate or secure areas, such as a lockable cage or secure carousel. Those areas are set aside, so do not count them as flexible space.
7. 7Group by similarity and co-occurrence. Store related parts together, such as gearbox parts. Put items that often appear on the same pick list side by side, such as a ¾-inch bolt next to its nut. This keeps density gains from adding travel.
8. 8Use task interleaving where a WMS supports it. The system directs a put-away on the way to collect a picked pallet or a replenishment pallet. It does not add storage volume, but it protects accessibility as density rises.
9. 9Price bulk buys against storage before approving them. Lower unit cost and cheaper large shipments both need extra space. Calculate whole-life cost: storage and handling, working capital interest, possible discounted sales, and disposal. Do not approve the purchase on unit price alone.

The source gives the container trade-off but no single case with a full dataset. This example uses its percentages and an illustrative 100-unit container. You ship palletized loads. The pallets alone can take up to 10 of the 100 units of cubic capacity, leaving 90 for product. If the pallets cannot be stacked or need clearance height, the usable figure falls further. Switching to slip sheets of about 2 cm each recovers most of that 10 units, so the loading cube rises. The source also says that depending on load configuration, shipping costs can be 15 to 33 per cent higher with pallets. Against that, you must buy a specialist forklift attachment at both the despatch and receiving warehouses. The load is lifted off the slip sheet and placed on a pallet for storage, so rack pallet locations are unaffected. If the receiving site cannot fit the attachment, the saving is not available, and the pallet option stays. Decide by comparing the shipping saving with the attachment cost at both ends plus the handling benefits of pallets.

• If a product has a fixed slot and is frequently out of stock, then move it to random locations or share its slot, because the empty slot cuts pallet storage utilization.
• If pick speed is the constraint and the SKU range is stable, then keep fixed locations and accept lower utilization.
• If containers arrive loose-loaded or palletized and shipping cost is high, then evaluate slip sheets, provided both ends can fit the forklift attachment.
• If a purchase discount requires more storage than you have, then calculate whole-life cost (storage, handling, interest, discounting, disposal) before buying.
• If items are fast-moving cartons, then slot them at middle-row height. If slow-moving, then use the lowest and highest shelves.
• If floor space is exhausted, then look to the cube above work and storage areas, such as overhead rails, before expanding.

• Counting floor area instead of cubic volume. Tell: racking is full at low level, but vertical clearance above is unused and 'we have no space' is the claim.
• Assigning fixed slots to every SKU. Tell: rows of empty pallet locations while receiving has no free slots for inbound goods.
• Over-densifying slow and fast lines alike. Tell: pickers bend and stretch or travel further for fast movers, because they sit on high or low shelves.
• Buying in bulk for the unit discount alone. Tell: obsolete stock, discounted clearance sales, or working capital tied up in unsold units.
• Ignoring special storage needs when slotting. Tell: hazardous or high-value goods sit in general rack locations, or secure cages are treated as general capacity.
• Switching to slip sheets without checking both ends. Tell: the receiving site has no forklift attachment and must re-handle the loads.

• ▸Zone / storage type: name the area, such as pallet rack, shelving, or container loading.
• ▸Total usable cubic capacity: length × width × usable height, in your own units.
• ▸Cube occupied by stock: the volume actually filled with product.
• ▸Utilization %: occupied ÷ total capacity, for each zone.
• ▸Fixed vs random locations: note which SKUs have fixed slots and how many are empty today.
• ▸Location constraints: pallet location size, rack weight capacity, hazardous or secure requirements.
• ▸Velocity slotting check: confirm fast movers are mid-height and slow movers lowest or highest.
• ▸Inbound loading method: pallet or slip sheet, estimated space lost, shipping cost effect, and whether both ends have the forklift attachment.
• ▸Planned bulk purchases: extra storage needed and whole-life cost (storage, handling, interest, discounts, disposal).
• ▸Action and owner: the single change to make, who owns it, and the date to re-measure.

1. 1Collect the inputs before you touch the layout. The source lists what a slotting or put-away system needs: size, weight and height of palletized goods; ABC or slotting results; current order data; family product groups; actual sales combinations; current pick-face status per product; pallet location sizes; racking weight capacity; and each location's distance from receiving and despatch. Build one SKU master with these columns.
2. 2Run the ABC analysis on velocity (the book shows it in Excel, Table 4.3). Rank SKUs by movement, split them into A, B and C classes, and keep the class on the master. Fast movers go closest to despatch. Set the class cut-offs from your own data, because the passages give no fixed percentages.
3. 3Add order-profile analysis on top of velocity (the book's FMCG order analysis, Table 4.5). Use current order data and actual sales combinations to find items ordered together, and slot them near each other. Use product family groups for the same purpose.
4. 4Add cube and handling constraints. Check size, weight and height against location size and racking weight capacity. Flag special cases: hazardous items need an appropriate area, and high-value items need a lockable cage or similar.
5. 5Decide fixed versus random locations per zone. Fixed locations let pickers memorize slots and pick faster, but an empty slot stays empty and pallet storage utilization drops. Random locations put the pallet in the most efficient available slot. Choose by zone, not for the whole building.
6. 6Do a pick-face analysis (Table 4.6). Check that each fast mover's pick face is sized for its demand and that replenishment from reserve or bulk storage feeds it. Put-away can route to a pick face as replenishment, to reserve storage, or straight to despatch for cross-docked product.
7. 7Load the rules into the WMS so it allocates locations in advance and directs the operator. Without a WMS, the manager calculates the optimum location and instructs operators by hand. Bartholdi's slotting tool (Table 4.4) is one worked reference.
8. 8Test the layout against the flow pattern (U-flow versus through-flow, aisle widths) and the capacity-versus-retrieval-speed trade-off. Re-run the analysis as order profiles change.

The passages do not reproduce the book's numeric slotting tables (4.2–4.6), so this example walks the put-away logic they do describe. A pallet arrives. The WMS holds its size, weight and height, its ABC class, current order data, and the status of its pick face. If the product is flagged for cross-docking, the system sends the pallet straight to the despatch area. If the pick face is low, the pallet goes to the pick face as replenishment. Otherwise it goes to a reserve or bulk location whose size and racking weight capacity fit the pallet. When several slots qualify, the system favours the one nearest despatch for an A item and a farther one for a slower item. If the product is hazardous or high value, the system overrides this and sends it to the appropriate area or a lockable cage. If the product has a fixed slot that is empty, the manager checks whether the memorization benefit justifies the unused space. Without a WMS, the manager performs the same sequence by hand.

• If a SKU is fast-moving (A class), then slot it closest to the despatch area and keep its pick face stocked.
• If order data show items regularly sold together, then locate them adjacently, using family groups and actual sales combinations.
• If a product has stock gaps and a fixed slot, then expect empty-slot waste; consider random locations for that zone to protect utilization.
• If pickers rely on memorized locations for speed, then favour fixed locations for the stable fast movers.
• If an item is hazardous or high value, then storage rules override velocity placement.
• If storage capacity gains slow retrieval, as with taller carousels that rotate more slowly, then size the system to the pick rate you need, not the height available.

• Slotting on velocity alone and ignoring order combinations. The tell: pickers travel to far corners for orders whose lines are all fast movers.
• Static ABC classes. The tell: a product's pick face is empty or oversized relative to current demand because the data predate current order profiles.
• Fixed slots everywhere. The tell: low pallet storage utilization with many empty positions for out-of-stock products.
• Ignoring physical constraints such as size, height and racking weight capacity. The tell: put-away overrides or pallets placed in locations they do not fit.
• Treating safety classes as ordinary slots. The tell: hazardous or high-value items sit in the fast zone without appropriate storage conditions.

• ▸SKU / description: identify the item and its product family group.
• ▸Velocity and ABC class: record movement from the ABC analysis and assign A, B or C.
• ▸Dimensions and weight: enter size, weight and height of the pallet or carton.
• ▸Order affinity: list the SKUs it is most often ordered with, from sales combinations.
• ▸Special storage: note hazardous, high-value or other conditions and the area or cage required.
• ▸Location type: choose fixed or random, with a one-line reason.
• ▸Assigned location and distance: record the slot and its distance from receiving and despatch.
• ▸Location fit check: confirm that location size and racking weight capacity accommodate the load.
• ▸Pick-face and replenishment status: record current pick-face level and the reserve source.

1. 1Fix the process before you buy technology. The source states that these basic processes must be in place in every warehouse before any thought of new technology: receive goods, process orders, replenish, add value, despatch. Write down each step as it runs today and mark every step that adds no value. 'Automating a bad process might make it quicker but certainly doesn't make it more efficient.'
2. 2Apply lean thinking to receiving and put-away first. Work through the sequence the source gives: pre-receipt, in-handling, preparation, offloading, checking, cross docking, recording, quality control, put-away. Aim to treat the warehouse as a trans-shipment point where goods received are despatched as quickly and efficiently as possible. Look for stock that can cross dock instead of being put away.
3. 3Fix packaging and pallet fit upstream. The source says outer packaging should fit the pallets used for transport and storage with no overhang and minimal unused space. Check incorrectly sized cartons at pre-receipt, because they undermine every step after it.
4. 4Analyse the order and SKU profile before choosing a picking strategy. Run a Pareto (80/20) analysis of orders and SKUs, an ABC analysis of quantity and frequency of sales, and a view of product cube picked and despatched per month. Use the results to set the layout: fast movers in the most accessible pick locations, slow movers elsewhere.
5. 5Choose between picker-to-goods and goods-to-picker from the data. The source lays out the options: picker to goods, goods to picker, and types of automated picking. Do not choose from equipment brochures. Choose from the order profile, the budget, and the fact that, as the source notes, not all operations can benefit from or afford large technology investments.
6. 6Match the pick method to the profile and compare the options on cost of errors. The source lists paper pick lists, pick by label, voice, barcode scanning, RFID, pick to light and put to light. Compare them on accuracy and the cost of errors as well as on speed. Pick to light suits dense forward-pick areas, as at Yankee Candle.
7. 7Use the WMS to balance work and sequence orders. Look for zone picking with workload balancing, wave-based zone sizing, multiple-order consolidation, carton-size calculation with label printing in advance, and shipping detail transfer to carriers. These functions come from the Yankee Candle system.
8. 8Measure before and after. Record picking efficiency, shipping volume and error counts before the change, then repeat the measurement after go-live. Yankee Candle reported exactly these three: efficiency, volume and errors.

Yankee Candle (source case) holds forward-picking stock in over 1,800 locations, in pick zones laid out in a U-shape with a conveyor through the centre. It started with paper-based picking on KDR flow racking. SSI Schafer added pick to light and new software, which made the process paperless. The software provides flexible zone picking to balance workload, and for each pick wave it looks at the distribution of picks and calculates a zone size for each operator. All zones can start picking at once, even when a single large order is split across zones. Operators pick in the printing sequence of the labels, which keeps order progress synchronized. The system calculates carton sizes and numbers from a database of packing parameters, prints labels in advance, and consolidates multiple orders into one shipment. Operations manager Bruce Mitchell reports that picking efficiency rose by 50 per cent, shipping volumes were up by more than 40 per cent on the previous year, and errors were down by over 50 per cent.

• If a process is unstable, wasteful or undocumented, fix it with lean methods before you automate. Automation only makes a bad process quicker.
• If the Pareto/ABC analysis shows a small share of SKUs driving most orders, put those SKUs in the most accessible pick positions and base the technology choice on that fast-moving group.
• If errors and their downstream cost (missed or low-level shipments) are the main problem, favour methods with built-in verification, such as pick to light, voice or scanning, and compare the options on the cost of errors rather than speed alone.
• If order lines are spread across zones and the workload is uneven, use a WMS with flexible zone picking and wave-based zone sizing instead of adding labour.
• If the operation cannot afford a large investment, or would not benefit from one, do not copy automated sites. Make the basic processes lean and add targeted technology.

• Automating a broken process. The tell: the new system runs faster, yet cost per order, error rates and rework do not improve.
• Choosing equipment before analysing the data. The tell: no Pareto, ABC or cube analysis exists in the business case, and the pick method was picked from a supplier demonstration.
• Ignoring upstream packaging and receiving. The tell: oversized or incorrectly sized cartons and pallet overhang keep appearing at put-away and pack, and storage space is wasted.
• Judging a pick technology on speed alone. The tell: the business case has no cost-of-errors line, and shipping errors and short shipments persist after go-live.
• Measuring no baseline. The tell: after go-live nobody can state the change in picking efficiency, volume or errors, so the investment cannot be defended.

• ▸Process steps mapped: list receiving, put-away, replenishment, picking, packing and despatch as they run today.
• ▸Non-value-adding steps: record each wasteful step and the lean fix proposed.
• ▸Receiving/packaging issues: note incorrectly sized cartons, pallet overhang and cross-docking opportunities.
• ▸Pareto/ABC findings: state the share of SKUs driving most orders and the layout implication.
• ▸Current pick method and pick strategy: for example paper, voice, light or scanning, and picker-to-goods or goods-to-picker.
• ▸Error rate and cost of errors: record the current figures and the cost per error.
• ▸Options compared: list candidate methods or WMS functions with the main trade-off for each.
• ▸Affordability check: state the investment, who benefits from it, and whether the operation can justify it.
• ▸Baseline and target metrics: record picking efficiency, shipping volume and error count, before and after.

1. 1Write the people accountabilities of the warehouse manager role before anything else. The job description in the source asks for people management skills and the ability to motivate and lead large numbers of employees through communication and engagement, alongside negotiation, IT, and finance skills. Turn each into an observable duty, such as how and how often the manager communicates with the team.
2. 2Test the manager-selection decision. Nick Weetman warns that in much warehouse management 'people have just been promoted and don't really understand how to run warehouses'. For each current or incoming manager, record whether they were promoted for operational performance alone, and add a development plan for the people-leadership gaps.
3. 3Audit the system against the book's own people-management headings: people challenges, attracting and retaining employees, an ageing and constantly changing workforce, operating hours, training, warehouse audit, quality systems. Give each a one-line current-state note and an owner. A blank line is a gap.
4. 4Build training from the actual hazards of the site, not from a generic induction. Walk the risk assessment and list the specific hazards for each role. Then write the instruction that answers each one, including what clothing to wear, what equipment to use, and what to do in an emergency.
5. 5Design operating methods and instructions with the care of people and product as the first priority. Check that housekeeping gives a safe, clean, clear, unobstructed floor or work space for the movement of goods, vehicles, equipment, and people.
6. 6Check how your hours of operation and your workforce profile (ageing, constantly changing) affect staffing and training. Re-run the training plan whenever either changes, because a changing workforce means repeated onboarding rather than a one-off event.
7. 7Measure people outcomes with both hard and soft measures, as the performance-management chapter distinguishes them. Pair productivity figures with indicators of engagement and communication, so you do not read only the cost and productivity numbers a manager is 'hunched over a laptop' deciphering.
8. 8Close the loop through the warehouse audit and quality systems. Compare what the people system says should happen (training given, instructions followed) with what an audit observes on the floor, and record corrective actions.

The source's temperature-controlled (chilled and cold store) section shows people management expressed as hazard-specific training. Start from the listed hazards: accidental lock-in, the effect of cold on people, accidental refrigerant (ammonia) release, slippery floors with ice build-up near door openings, ice falling from panels, product falling from racks, and working at height. For lock-in, the control is alarms and quick-release equipment, so staff training covers where and how to use them. For cold, the source calls for specific advice and training on thermal clothing, drinking lots of water, protecting bare skin (fingers, noses, ears), and taking greater care if smokers or drinkers. For working at height, mechanized elevating working platforms (MEWPs) are obligatory, but they are not equipped for sub-zero temperatures. That makes the sub-zero practice one that is closely scrutinized, and the industry and the Health and Safety Executive produced Supplementary Guidance (PM 28) in 2010 for it. The people system therefore needs a named person to keep that guidance current and brief staff on it. Risk assessments and operating instructions are written with people first, and the training record shows who was briefed on which hazard.

• If a manager was promoted for operational results and has not been developed in leading people, then treat them as a people-skills gap and add communication and engagement development before adding responsibility.
• If a role faces a site-specific hazard (cold, ice, refrigerant, working at height), then training for that role must name the hazard and its control. General induction does not satisfy it.
• If the workforce is ageing or changes constantly, or operating hours change, then re-plan training and staffing rather than assuming the previous plan still holds.
• If your performance review relies only on productivity and cost figures, then add soft measures so engagement and communication problems become visible.
• If an audit finds practice differing from written instructions, then fix the instruction or the training, and do not just reprimand the individual.

• Promoting the best operator without people-leadership development. The tell is a manager who understands the tasks but cannot motivate or communicate with the team.
• Generic induction in a hazardous environment. The tell is staff who cannot say where the lock-in alarm is, why they should drink water in the cold, or which areas ice up.
• Managing only from the laptop. The tell is a manager who knows the cost and productivity figures but is rarely on the floor, so communication with the team is thin.
• Treating training as a one-time event in a constantly changing workforce. The tell is that newer or temporary staff are the ones involved in incidents and errors.
• Writing risk assessments and operating methods that put throughput ahead of people. The tell is a layout or process with obstructed floors, unmanaged spills, or ice around door openings.

• ▸Manager accountabilities: list the people duties (leading the team, communication, engagement) from the job description.
• ▸Manager development gaps: note where the manager was promoted without people-leadership grounding and the planned fix.
• ▸Attraction and retention: current approach and the main reason people leave or stay.
• ▸Workforce profile: note ageing, turnover, and how often the workforce changes.
• ▸Operating hours: shifts worked and how training and communication reach every shift.
• ▸Site hazards by role: list each hazard (cold, ice, lock-in, refrigerant, height) against the role exposed.
• ▸Training delivered: specific instruction given for each hazard, with date and who was trained.
• ▸Hard and soft measures: productivity or cost figures plus engagement or communication indicators in use.
• ▸Audit findings and actions: gaps between written instructions and observed practice, with owner and due date.

1. 1Build the assessment around the hazard categories a warehouse actually contains. The source's health and safety chapter runs: risk assessments, layout and design, fire safety, slips and trips, manual handling, working at height, vehicles, forklift trucks, warehouse equipment legislation, first aid. Use these ten as your checklist headings so no category is skipped.
2. 2Start with the risk assessment, because every other category feeds it. Walk each area (receiving, storage, pick, despatch) and for each hazard record who is exposed, how they could be hurt, and the control in place. Assess the specific facility and operation, since hazards vary by type and size of facility and the operations undertaken.
3. 3Treat layout and design as a control, not a backdrop. Check where pedestrians, forklifts and vehicles cross, and where door openings, racking and storage sit relative to those routes. Redesign the flow before you rely on signs or PPE.
4. 4Write a specific control for each hazard, and name the equipment or training behind it. Use the cold-store pattern: 'accidental lock-in' is answered by alarms and quick-release equipment; 'effect of cold on people' is answered by thermal clothing, water and skin protection, delivered through specific advice and training.
5. 5Give working at height and mechanical handling their own rules. Decide which lifting platforms are permitted (the source notes non-integrated platforms on forklift trucks have effectively been eliminated except in sub-zero stores) and confirm that the forklift and warehouse equipment legislation applies to each machine in use.
6. 6Set a fast response for spills and displaced product. Spills are to be cleaned quickly so crushed product and packaging do not spread across the working space or into the working parts of equipment. Displaced product on racks is a falling-product hazard, so include it in the rack inspection.
7. 7Put monitoring on a routine. For temperature-controlled stores, equipment is checked each shift or am/pm and records are kept for regulatory and operational analysis. Apply the same routine to safety checks: a named checker, a fixed frequency, and a retained record.
8. 8Train against the assessment and confirm first aid cover. Each training item should map to a hazard you recorded (cold exposure, manual handling, forklift operation). Review the whole assessment when the operation, equipment or facility changes.

The source's cold-chain section is a ready-made assessment of a chilled/frozen store. Walking it hazard by hazard: (1) Accidental lock-in. Control: alarms and quick-release equipment on cold-room doors. (2) Cold on people. Control: specific advice and training on thermal clothing, drinking water and protecting fingers, noses and ears, with greater care for smokers and drinkers. (3) Accidental refrigerant release, particularly ammonia. This needs its own emergency procedure and is not covered by general PPE rules. (4) Forklifts on slippery floors where ice builds up, particularly around door openings. The risk is skidding and overturning, so control it at the door openings through floor condition and travel rules. (5) Ice build-up on panels. This is an ice-fall hazard and, if left, can bring roof panels down on operators below, so it needs a scheduled inspection and removal. (6) Product falling from pallet racks because of displaced product, controlled by rack checks and prompt correction. (7) Working at height: non-integrated forklift platforms are effectively eliminated except in sub-zero stores, so you confirm which store you are in before allowing any. Each line has a hazard, an exposed person, a control and a check, which is the form every warehouse area should take.

• If a hazard can lock a person in, trap them or release a hazardous substance (cold-room doors, ammonia refrigerant), then require a physical control such as an alarm or quick-release before relying on training alone.
• If forklifts or other mechanical handling run on a floor that can become slippery (ice near door openings, product spills), then treat skidding and overturning as live risks. Fix the floor condition and travel routes, not just driver instructions.
• If people must work at height, then use integrated or mechanized access by default and permit non-integrated forklift platforms only in the narrow case the source identifies (sub-zero stores).
• If the facility type, size or operation changes (for example chilled to sub-zero), then reassess, because hazards vary by facility and operation.
• If a hazard appears in a regulated product environment (food cold chain), then keep the monitoring records, since temperature and check records serve both regulatory and operational analysis.

• Generic assessment: one template copied across sites. The tell is that cold, ammonia or ice hazards are missing from a cold store's file, or the assessment never names the specific facility and operation.
• Relying on PPE and training for hazards that need engineering controls. The tell is that lock-in or refrigerant release is handled by 'staff briefed' with no alarm, quick-release or emergency procedure.
• Ignoring slow-building hazards such as ice on panels or door-opening ice. The tell is that incidents happen only after the build-up is visible and no inspection schedule exists.
• Letting spills and displaced product wait. The tell is crushed product spreading across the working space, packaging in equipment moving parts, or product falling from racks.
• Records kept for compliance but not used. The tell is shift or am/pm check logs that are filled in but never reviewed for trends or acted on.

• ▸Area/operation: name the zone (e.g. chill store, door openings, racking aisle, despatch dock).
• ▸Hazard category: choose from risk assessment, layout and design, fire, slips/trips, manual handling, working at height, vehicles, forklift, equipment legislation, first aid.
• ▸Specific hazard: describe it concretely (e.g. ice build-up around door openings).
• ▸Who is exposed: operators, drivers, pedestrians, visitors; note anyone working below or alone.
• ▸Control in place: the physical or procedural control (alarm, quick-release, PPE, access equipment, floor treatment).
• ▸Training/advice required: what staff must be told or taught for this hazard, and who has completed it.
• ▸Check method and frequency: who checks, how, and how often (per shift, am/pm, scheduled).
• ▸Record location: where check results are kept and who reviews them.
• ▸Review trigger: the change that forces reassessment (new equipment, new temperature zone, layout change, incident).

1. 1Split picker time into three buckets before changing anything: travel between locations, waiting (for equipment or for a free pick face), and the pick itself. The source frames travel and wait as non-value-added, and the definition puts them at over 50% of picking time. Measure your own split rather than assuming it.
2. 2Apply Goldratt's first step, identify the constraint. Walk the pick area and note where people stand waiting their turn to pick. Congestion at a pick face is the signature bottleneck, so record the location, the SKU and the time of day.
3. 3Run an ABC/Pareto analysis of orders and SKUs, using order lines and frequency of sale. The book's pick-preparation chapter pairs this with a basic layout based on ABC classification. Fast movers belong where travel is shortest, and slow movers go further out.
4. 4Check the current slotting against that ABC ranking. List every high-frequency SKU sitting far from the despatch or start point, because each of these is a repeated long walk. Re-slot the worst offenders first.
5. 5Exploit the constraint (Goldratt step 2) without spending money. Do not send too many pickers to the same location at once, and stagger or re-sequence the pick waves or lists so that two pickers do not need the same hot SKU at the same moment.
6. 6Subordinate everything to the constraint (step 3). Do not release more picking work than the congested face can handle, even if the rest of the area is idle.
7. 7Elevate the constraint (step 4) by increasing the number of pick locations for the congested product, so more than one picker can pick it simultaneously.
8. 8Return to step 1. Once the face clears, find the next constraint, which may be travel distance or waiting for handling equipment. Repeat the walk-and-record cycle.
9. 9Only after layout and slotting are tuned, compare picking strategies (picker-to-goods versus goods-to-picker, cluster pick carts, pick-to-light, put-to-light) against the travel and wait share you measured. Choose the one that attacks your largest bucket.

The source's own illustration is congestion at a pick face. Step 1: you identify the bottleneck by seeing people wait in turn to pick the same item. Step 2: you maximise productivity under the current situation, and the source's instruction is not to overload the system by sending too many pickers to the same location. Step 3: you stop releasing more work to that face than it can absorb. Step 4: you elevate the constraint by adding pick locations for that product, so several pickers can pick it in parallel. Step 5: you return to the start and look for the next constraint. In practice, suppose a fast-moving SKU has one face and three pickers queue at it each wave. You re-sequence the lists so no more than one picker needs it at a time, then open a second location for it. The queue disappears, and the next walk-through shows whether long travel to far-slotted SKUs is now the largest bucket. The passages give no figures for this case, so your before-and-after numbers come from your own log.

• If you see pickers queuing at one face, treat it as a constraint. First re-sequence the work so fewer pickers arrive there together, then add pick locations for that product.
• If the SKU has high order frequency but sits far from the start or despatch point, re-slot it closer using ABC analysis, since travel scales with how often the walk repeats.
• If one location is overloaded, do not respond by adding more pickers to the same area. Subordinate the release of work to what the face can handle.
• If fixing one bottleneck exposes another, repeat the five steps. A cured constraint always returns you to step 1.
• If layout and slotting are already tuned and the travel share is still high, evaluate a different picking strategy, such as goods-to-picker, cluster carts or pick-to-light, rather than more layout tweaks.

• Adding pickers to a congested face. The tell is more people standing in line and no rise in lines picked per hour.
• Slotting by storage convenience instead of order frequency. The tell is top Pareto SKUs located at the far end of the pick area.
• Treating wait time as unavoidable and counting only walking. The tell is a time study that shows no waiting category.
• Fixing the first bottleneck and stopping. The tell is that congestion reappears at a different face or equipment point soon after.
• Buying automation or a new picking method before measuring the travel/wait split. The tell is that no one can state which bucket the investment is meant to shrink.

• ▸Date/shift and wave: record when the observation was made so peaks can be compared.
• ▸Pick location and SKU: name the face where queuing or long walking was observed.
• ▸ABC class of SKU: A, B or C from your Pareto/ABC analysis.
• ▸Travel time observed: time spent walking between locations on the sampled pick list.
• ▸Wait time observed: time spent waiting for a free face or for equipment, with the cause.
• ▸Pick time observed: time spent actually picking, used to compute the share of non-value-added time.
• ▸Number of pickers needing this face at once: shows whether work release is overloading it.
• ▸Action taken (exploit / subordinate / elevate): record which Goldratt step you applied, such as re-sequencing, holding work or adding pick locations.
• ▸Next constraint identified: note what limits the area after the fix, to restart the cycle.