The Logistics and Supply Chain Toolkit

1. 1Pick one resource and one time window first (for example, inbound labour and forklifts for a day). Utilization is only measurable against a defined activity list and a defined period.
2. 2List every activity that consumes the resource, such as unloading, put-away to drive-in racking, put-away to pick locations, and replenishment. Record the volume for each in pallets or lines.
3. 3Set the labour parameters before measuring: working hours and the utilization rate. The source's plan uses 7.2 productive hours per person against 8 hours per truck per day, which implies 90% utilization for people and 100% for trucks.
4. 4Set a rate for each activity from a time-and-motion study or from staff interviews. Convert volume to hours by dividing volume by rate (124 pallets at 16 per hour = 7.75 hours).
5. 5Add ancillary work as its own line. The source uses 15% of the activity subtotal for collecting paperwork, equipment and similar tasks. Dividing total hours by productive hours per person gives headcount, and equipment hours by 8 gives trucks.
6. 6Compare standards with actuals, then revise the standards if they prove wrong (the source's labour-management flow: identify activity, set parameters, study, agree standards, compare, revise). The gap between planned and actual hours is your utilization loss.
7. 7Cut the hours before adding resource. Use task interleaving (a WMS pairs an inbound put-away with an outbound or replenishment pick so the operator travels full both ways), which the source says can reduce equipment use by up to 30%. Slot fast movers near dispatch using Pareto/ABC analysis to cut travel.
8. 8Audit remaining waste against the 7 Muda (TIMWOOD): transportation, inventory, motion, waiting, over-processing, overproduction, defects. Add the optional 'S' for skills, meaning underused worker capability. Re-run the plan after each change.

The source's in-handling resource plan converts pallet volumes to hours. Put-away to drive-in racking is 124 pallets at 16 per hour, which is 7.75 hours on a forklift (FLT). Put-away to pick locations is 26 pallets at 5 per hour, which is 5.2 hours on a powered pallet truck (PPT). The full activity list sums to 55.95 hours. Ancillary work at 15% adds 8.39 hours, for a total of 64.34 hours. Divided by 7.2 productive hours per person per day, that gives 8.94 staff. Equipment hours of 10.20 and 18.25 are divided by 8 productive hours per truck, which gives about 1.28 and 2.28 trucks. The plan therefore says to staff about nine people and about three or four trucks, with fractions as the signal for how much slack exists. Task interleaving would then be tested against the truck figures, since the source says it can cut equipment use by up to 30%.

• If the plan calls for a fractional truck or person (1.28 trucks, 8.94 staff), then check whether interleaving, re-slotting or shift timing can absorb the fraction before you buy or hire the extra unit.
• If travel dominates the hours, then re-slot with Pareto/ABC analysis (fast movers nearest dispatch) and use task interleaving before adding equipment.
• If actual hours repeatedly diverge from standards, then revise the standard or fix the process. Do not keep planning on a standard nobody achieves.
• If you are choosing equipment, then decide it together with the storage medium. Weigh extra space capacity against throughput speed, and check aisle width and lift height in the MHE comparison (a VNA truck also needs a second truck to move product to and from the racks).
• If you justify a WMS on utilization, then use the source's ranges (labour 10–35%, floor space 10–30%, rolling stock 10–20%) as upper bounds. Plan lower savings in year one because of the learning curve.

• Planning at 100% of the shift: staff hours are not discounted for productive time. The tell is a headcount below what the same workload needs at 7.2 productive hours per person.
• Leaving out ancillary work (paperwork, collecting equipment). The tell is a plan that looks efficient on paper while operators queue or wait.
• Adding trucks or labour to fix a travel problem. The tell is high equipment hours with half-empty runs, where interleaving or re-slotting would have worked.
• Treating utilization as labour only. The tell is trucks, space and vehicles never appearing in the plan, even though the source counts all four assets.
• Setting standards once and never comparing them with actuals. The tell is no recorded variance and no revision date.

• ▸Activity: name each task using labour or equipment (unload, put-away to drive-in, put-away to pick, replenishment).
• ▸Volume: pallets or lines per activity for the planning period.
• ▸Rate per hour: from a time-and-motion study or staff interviews, with the source noted.
• ▸Hours required: volume ÷ rate for each activity.
• ▸Equipment type: FLT, PPT, reach truck and so on, assigned to each activity.
• ▸Ancillary allowance: percentage applied to the subtotal (the source uses 15%).
• ▸Productive hours available: per person and per truck per day (the source uses 7.2 and 8).
• ▸Resources required: total hours ÷ productive hours, for both staff and equipment.
• ▸Actual hours and variance: log actual hours, compare with plan, and note whether the standard or the process changes.

1. 1Fix the internal gap first. The value tree passage states a company must enable internal collaboration before exposing itself to partnerships in the open market. Map where product development, procurement, manufacturing, sales and logistics each optimise only their own process, and align them to business strategy before you approach a partner.
2. 2Build a common language that links operational and financial metrics. The passage names the common language as the most important aspect of collaboration. Use a value-tree style model so one operational action (for example a change in supplier payment terms under accounts payable) shows its financial effect in terms finance and the partner both accept.
3. 3Consult the functions that feed the plan. The supply chain strategy has to be developed with R&D (new product introduction), sales and marketing (commercial plans, long-term forecasts, geographic ambitions) and finance (cash flow, profitability, tax). These inputs are what you will later share with partners, so they must be agreed internally first.
4. 4Select the collaboration form by the control-versus-risk trade-off. In the outsourcing pyramid, 'blind' and 'procurement officer' arrangements are cost-led, with the specification based on the previous 12 to 18 months. Prescriptive arrangements keep tight customer control. At the top, parties share infrastructure, assets, cash and risk ('WIIFWe'). Decide which level the relationship justifies, and do not call a cost-led contract a partnership.
5. 5Match the mechanism to the shared problem. The toolkit lists CPFR (collaborative planning, forecasting and replenishment), vendor-managed and co-managed inventory, consignment stock and factory gate pricing as separate tools. Use CPFR for joint planning and replenishment, VMI/CMI where the supplier manages stock, and FGP where you coordinate transport with suppliers. Choose one pilot partner and one mechanism.
6. 6Agree confidentiality and trust before sharing data. The passages treat trust as shared by both parties and list non-disclosure agreements as a tool. Sign the NDA, then agree which forecasts, stock positions and plans each side will share.
7. 7Coordinate execution, not just information. In FGP, store-delivery vehicles call back at suppliers on the return leg and collect product bound for the DC. Define the equivalent joint action in your pilot, such as who collects, who consolidates and who schedules.
8. 8Review against agreed measures and reset on a cycle. Record shared outcomes such as on-time delivery at the DC, visibility, logistics cost and inventory, and review at least annually. The supply chain strategy passage sets that cadence, and the partnership plan should follow it.

Factory gate pricing shows collaboration as coordinated execution. UK retailers introduced it in the early 2000s. Quick-response lean principles meant smaller, more frequent deliveries, which pressured the DCs. Under the old model suppliers delivered, or a nominated carrier collected. Under FGP the retailer buys at an ex-works price and organises transport itself (Potter et al, 2007). Vehicles that deliver from the DC to stores call back at nearby suppliers on the return journey and collect full or part loads, or make several collections, depending on truck capacity and driver time. That requires suppliers to share collection windows and ready-to-ship volumes, and the retailer to share its vehicle schedules. The passage reports better on-time delivery at the DCs, greater visibility, fewer vehicles at site and logistics cost reductions of 5 to 8 per cent (Potter et al, 2007; Le Blanc et al, 2004). Treat 5 to 8 per cent as a sanity-check range, not a target. The prerequisites are suppliers in close proximity and vehicles suitable for both inward and outward journeys.

• If product development, procurement, manufacturing, sales and logistics still run as silos, then fix internal alignment before inviting external partners, because internal collaboration comes first.
• If the relationship is driven by cost saving with no service or operational criteria, then you are in 'blind outsourcing' at the base of the risk pyramid. Define requirements before you describe it as collaboration.
• If you want shared infrastructure, assets, cash and risk (the WIIFWe level), then expect to give up some prescriptive control and agree how risk is split, or stay at the prescriptive level.
• If you run a vehicle fleet and have suppliers close by whose products suit your return journeys, then test FGP before wider data-sharing programmes.
• If the two sides cannot express options and decisions in a common language, such as shared operational and financial metrics, then pause the joint planning and build the shared metric set first.

• Starting with the partner before the house is aligned. The tell: your sales forecast, finance plan and procurement plan disagree when you try to share them.
• Mistaking a contract for collaboration. The tell: the specification is copied from the last 12 to 18 months' performance and the only objective is a lower price.
• Sharing information without agreed rules or trust. The tell: no NDA, and each side withholds the data the other needs.
• Using the tool name without the shared decision. The tell: CPFR or VMI is on the slide, but nobody can name the joint plan, the owner or the replenishment rule.
• Applying FGP-style coordination where the geography or vehicles do not fit. The tell: suppliers are far from the delivery routes, or trucks are unsuitable for inward collection, so backhaul runs empty.

• ▸Partner and relationship level: name the partner and place the relationship on the pyramid (blind, procurement-led, prescriptive, or shared risk/WIIFWe).
• ▸Internal alignment check: record whether R&D, sales/marketing, finance, procurement and logistics have signed off the plan inputs, and the date.
• ▸Common metric set: list the shared operational and financial metrics both sides will use.
• ▸Mechanism chosen: CPFR, VMI/CMI, consignment stock or FGP, with one sentence on why it fits.
• ▸Information to be shared: forecasts, stock levels, schedules; state who shares what and how often.
• ▸Confidentiality status: NDA signed (Y/N), date and named data owners.
• ▸Joint execution action: describe the coordinated activity, such as backhaul collection or replenishment trigger, and who owns it.
• ▸Outcome measures: on-time delivery, visibility, logistics cost, inventory, with baseline values filled in by you.
• ▸Review date and decision: set the next review (at least annually) and the criteria for scaling or stopping.

1. 1Set the baseline first. Pull total logistics cost for one period across the four buckets (transportation, warehousing, inventory holding, labor). Express it two ways: percentage of sales and cost per unit. The passages give no benchmark figures, so your own baseline is the only comparator.
2. 2Attack transport cost through fuel. Make fuel efficiency a procurement criterion for new trucks, because the higher upfront price is outweighed by later fuel savings. Choose tyres on best value (safety, longevity, fuel efficiency), not lowest price, and add tyre pressure monitoring. Retro-fit aerodynamic items such as side skirts, which can return quickly.
3. 3Protect those savings with maintenance. Fuel-efficient trucks, tyres and aero kit only deliver if kept in good condition. Schedule maintenance with the equipment suppliers, who can advise on correct use.
4. 4Cut empty running and half loads. Improve route planning and journey organization to shorten journey times, avoid congestion and reduce empty miles. Work with customers on delivery timing, depot placement outside urban centres, and consolidating so trucks run full.
5. 5Consider factory gate pricing where you run delivery vehicles and have suppliers nearby. Buy at an ex-works price and collect from suppliers on the return leg of store or customer deliveries. The studies cited report logistics cost reductions of 5 to 8 per cent. Check that vehicles suit both inward and outward journeys.
6. 6Lower warehouse labor and space cost. Match pick method to volume: label picking and barcode guns suit under 100 lines per hour, wearable scanners under 150, voice picking 100–250. Use ABC analysis to place fast movers efficiently. Choose racking and aisle width together with the forklift type, using WA + LLC + L (plus 300 mm operator clearance, and minus R for reach trucks).
7. 7Measure with telematics, and a TMS if whiteboards and planner experience are your limit. Track fuel by driver to find waste. Before buying a TMS, calculate ROI on purchase and support, and fix poor processes first so you do not automate them.
8. 8Re-run the baseline each period and credit each saving to the action that caused it. Stop actions that do not move cost per unit.

The passages offer a real pattern in factory gate pricing. UK retailers introduced FGP in the early 2000s because quick-response lean principles produced smaller, more frequent deliveries that pressured DCs and raised transport cost. Instead of suppliers delivering to the DC, the retailer's store-delivery trucks call at nearby suppliers on the return leg and collect product bound for the DC, as full loads, part loads or multiple collections, depending on truck capacity and driver time. The retailer pays an ex-works price and organizes the transport itself. The cited studies (Potter et al, 2007; Le Blanc et al, 2004) report logistics-related cost reductions of 5 to 8 per cent. Fewer vehicles visit each site, and DC throughput improves through consolidation. Apply it to your own numbers: if logistics cost is a given percentage of sales, a 5–8 per cent reduction in that cost is the plausible range to test in a pilot, not a promised result. It works only where suppliers are close and the trucks suit both inward and outward journeys.

• If a vehicle's cheapest option costs more in fuel, tyres or downtime over its life, then buy on whole-life value, not purchase price.
• If your trucks return empty and suppliers sit near your delivery routes, then test factory gate pricing and backhaul collection.
• If processes are unstable or poorly defined, then fix them before buying a TMS or WMS, since automating a poor process locks in the waste.
• If pick rates are under about 100 lines per hour and system support is thin, then use label or barcode-gun picking. If you need hands-free, higher rates (up to 250), then price voice or wearable options against their hardware and maintenance cost.
• If a WMS is needed, then buy a packaged product, since vendors amortize development over 100+ clients. Build in-house only for specialized requirements or integration.

• Buying the cheapest tyres or trucks. The tell: fuel cost per mile rises while the purchase budget looks good.
• Fitting aerodynamic kit and efficient equipment but skipping maintenance. The tell: fuel consumption drifts back up after an initial improvement.
• Measuring only transport cost. The tell: transport falls while inventory holding or warehouse labor rises, and total cost as a percentage of sales does not improve.
• Automating a broken process with a TMS or WMS. The tell: new system, same errors, rework and manual workarounds.
• Over-specifying how a WMS should work in the RFI. The tell: every vendor proposes the same approach and you miss faster, cheaper alternatives.

• ▸Period covered: state the start and end dates of the baseline.
• ▸Sales for period: enter total sales so cost can be shown as a percentage of sales.
• ▸Transportation cost: enter fuel, tyres, maintenance, drivers and subcontracted freight.
• ▸Warehousing cost: enter space, racking, equipment, systems and pick-method costs.
• ▸Inventory holding cost: enter the cost of stock held across DCs and stores.
• ▸Labor cost: enter warehouse and transport labor, including supervisors.
• ▸Total logistics cost % of sales: divide the sum of the four buckets by sales.
• ▸Units shipped and cost per unit: divide total cost by units for a per-unit view.
• ▸Action and expected saving: name each initiative (e.g., side skirts, backhaul, TMS), its cost, and the saving you expect, then record the actual next period.

1. 1Write the perfect order as a checklist before choosing any metric. The source's intro gives the test: right product, correct quantity, right customer, prescribed time, good condition, acceptable cost. An order is perfect only if every condition holds, so each condition becomes a pass/fail field on the order record.
2. 2Start the clock at the customer's order, not the internal hand-off. The source warns that 'dispatch within 24 hours of the warehouse receiving the order from sales' may not be 24-hour dispatch from receipt of the customer order, because the order 'could have sat on someone's desk for a day'. Timestamp customer order receipt, release to warehouse, dispatch and delivery. Cycle time is the total from first to last.
3. 3Measure fill against what the customer ordered, not what the warehouse was told to pick. The source notes that 100 per cent dispatch of what's available is not what the customer ordered if the order was stock-adjusted before reaching the warehouse. Keep the original customer order lines and compare final delivery to them. Record every adjustment as a shortfall.
4. 4Pick a short KPI set. Start with the customer-service measures in the source's Table 7.1: on-time delivery and orders in full first time. Add order accuracy, cycle time and the perfect order as defined in step 1. The source warns that too many measures leave you 'spending too much time measuring and not enough time managing'.
5. 5Set a standard and a target for each KPI (the source's Table 7.2 pairs indicators with standards and targets). Translate them into 'bite-sized chunks' so every employee knows 'what a good job looks like'. Give each KPI a named owner, because the source stresses user ownership.
6. 6Validate the data before you act on it. The source's tests: accurate, complete, aimed at the correct audience, simple enough to maintain, and measured against historical data as well as best practice. Reconcile a sample of reported 'on-time' orders to the original customer order and delivery evidence.
7. 7Benchmark using the source's stages. Define what customers want and which process affects it, identify partners (including firms in other industries), collect the data, and determine the performance gap. Then work out how the best performers do it, not just their numbers, and target future performance.
8. 8Track cost next to service. The source says to monitor the criteria that matter to customers (service) and those that matter to you (costs). It also says the measurement effort must cost less than the savings from improved productivity. Review the KPIs when business activity changes.

The source gives two ways a service metric can look healthy while the customer is poorly served. First, a warehouse reports 100 per cent dispatch of what is available. But the order was stock-adjusted before it was sent for picking, so the customer received less than they ordered. The in-full KPI passes and the customer is still short. Second, the warehouse reports dispatch within 24 hours of receiving the order from sales. The order sat on someone's desk for a day before reaching the warehouse, so the customer waited longer than the metric says. The fix is to re-anchor both measures to the customer's original order. Compare delivered lines to ordered lines, count every stock adjustment as a miss, and run the clock from customer order receipt. The reported figures will probably fall. They now show what the customer experienced, and the gap is the improvement agenda.

• If a metric's clock or quantity baseline starts after the customer's order, then re-base it to the customer's original order before reporting it.
• If an order is stock-adjusted before picking, then count it as not in full against the customer order, and log the adjustment as a separate cause.
• If the KPI list has grown beyond what people can act on, then cut it. The source says to avoid so many measures that measuring crowds out managing.
• If measuring a KPI costs more than the productivity savings it produces, then simplify or drop it.
• If you cannot say why a best-in-class performer scores higher, then keep digging into their process before setting a target. The source says to understand how they do it, not only what they score.

• Reporting fill against the warehouse's picked order instead of the customer's order. The tell: in-full is near 100 per cent while customers still complain of shortages.
• Starting the cycle-time clock at warehouse receipt. The tell: dispatch is within target but customers report late deliveries, and there are unexplained delays between sales and the warehouse.
• KPI overload. The tell: long dashboards and heavy reporting effort, but no one can name the two or three measures that drive action.
• Measures with no owner and no definition of a good job. The tell: figures are reported but no behaviour changes, and employees cannot say what they should do differently.
• Benchmarking numbers without process understanding. The tell: you set targets from a competitor's headline score but cannot explain how they achieve it, so the gap does not close.

• ▸Customer order received (date/time): record the moment the customer's order arrived, not the hand-off to the warehouse.
• ▸Released to warehouse (date/time): record when it reached the warehouse. The gap from the line above shows internal delay.
• ▸Original customer order lines: list product and quantity exactly as the customer ordered, before any stock adjustment.
• ▸Stock adjustments before picking: note any change to lines or quantities and the reason. Each one counts as a shortfall.
• ▸Delivered lines vs ordered lines: compare what arrived to the original order, to judge in full and order accuracy.
• ▸Delivery date/time vs promised: record the on-time result and the cycle time from customer order to delivery.
• ▸Condition on delivery: mark good or damaged, covering the 'good condition' test.
• ▸Perfect order result: pass only if right product, quantity, customer, time and condition all hold. Add cost within acceptable limit.
• ▸KPI owner and corrective action: name the owner, the root cause of any miss and the action, per the source's call for user ownership and remedial action.

1. 1Separate agility from its neighbours before you measure anything. The SCOR passage defines five attributes: reliability (on time, complete, undamaged), responsiveness (time to react to and fulfil demand), agility (ability to increase or decrease demand within a given planned period), cost and assets. Write one sentence per attribute for your chain so you do not score a slow chain as an inflexible one, or the reverse.
2. 2Capture the configuration as-is across the SCOR processes. Plan: levels of aggregation and information sources. Source: locations and products. Make: sites and methods. Deliver: channels, inventory deployment, products. Return: locations and methods. Enable: how the chain is managed. Record each process in one line so you can see where a demand or supply change would enter and where it would stall.
3. 3Measure with the first-line metrics the passage lists: perfect order (on time, in full, damage free, documents accurate), order fulfilment cycle time, and supply chain flexibility. Compare them against internal goals and external industry benchmarks. The passages give no agility thresholds, so set your own planned-period target.
4. 4Find the binding constraint. SCOR pairs process change with theory of constraints analysis to reach root-cause issues. Ask where a change in volume or supply backs up: a picking method, a lead time, an information gap, or a supplier loading bay.
5. 5Shorten and flex the physical flow. Where products can bypass storage, use cross-docking: consolidate items on the inbound side and move them straight to dispatch, with stock on site for a maximum of 48 hours and most leaving sooner. This needs advance shipping notices, goods that are easy to identify on arrival, and barcodes or RFID tags aligned across suppliers.
6. 6Match the pick method to the speed you need. Individual order picking is time-consuming and error-prone on paper. Cluster picking takes several orders out at once and needs system help and accurate sorting. Batch picking collects large quantities in one go. Release tasks only when they can be planned alongside other work, because releasing too early or too late damages productivity and equipment and labour use.
7. 7Cut the lead time of making and changing things. Where parts are machined or outsourced, test whether 3D printing removes missed deadlines and 'just in case' stock. The passage credits it with more iterations, faster failure, early detection of design flaws and less machine downtime.
8. 8Realign and re-measure. Move from as-is to to-be through process re-engineering, lean analysis or Six Sigma, then re-run the metrics from step 3. Treat it as a cycle, because SCOR is described as cyclic.

The passages give two cases. Walmart attributes part of its success to cross-docking as much as 85 per cent of its products, working closely with suppliers and running sophisticated IT systems. Because stock does not sit in storage (maximum 48 hours on site), a change in demand reaches dispatch in hours and holding and handling costs fall. The prerequisites are ASNs and barcodes or RFID aligned across suppliers. The second case is the factory gate pricing flow in Figure 4.1. A supplier collects from another supplier, delivers to the DC, then collects from the DC and delivers to stores on the way back to its own site. This adds routing flexibility, but only if the buyer invests in an IT system to manage the whole movement. Product compatibility (cube, weight, temperature, fragility, hazardous status), supplier loading-bay capability (double-deck trailers can cause problems) and communication between the inventory and logistics teams all constrain it. In both cases the agility comes from information and compatibility, not from adding capacity.

• If you can only measure one thing, use the SCOR trio of perfect order, order fulfilment cycle time and supply chain flexibility, and compare them to an external benchmark before changing any process.
• If product can pass from inbound to dispatch without storage and your suppliers can supply ASNs and aligned barcodes or RFID, then cross-dock. If they cannot, fix the information flow first.
• If you are consolidating supplier collections on return legs, then check product compatibility and the supplier's loading bays and operating hours before you commit, and make sure an IT system can manage the whole route.
• If you are holding stock 'just in case', or missing deadlines because of machining or outsourcing lead time, then evaluate 3D printing ROI for those parts.
• If a response is slow and you cannot say why, then apply theory of constraints or root-cause analysis before buying capacity or technology.

• Treating responsiveness as agility. The tell: you report cycle time only and cannot say how far volume can rise or fall within a planned period.
• Cross-docking without data discipline. The tell: goods arrive without ASNs, labels do not match across suppliers, and staff send items to storage anyway.
• Adding supplier backhauls without checking the sites. The tell: vehicles such as double-deck trailers cannot use a supplier's loading bays, or mixed loads break compatibility rules.
• Releasing warehouse tasks too early or too late. The tell: idle labour and equipment alongside congestion and queues.
• Using speed metrics and ignoring reliability. The tell: faster cycle times while perfect-order rates (on time, in full, damage free, documents accurate) fall.

• ▸Planned period: the period within which you must raise or lower volume, in your own units.
• ▸SCOR configuration: one line each for plan, source, make, deliver, return, enable.
• ▸Perfect order rate: on time, in full, damage free, documents accurate; record the current value.
• ▸Order fulfilment cycle time: current value and the benchmark you compare against.
• ▸Supply chain flexibility: current value and internal target.
• ▸Binding constraint: the process step where change stalls, and the root cause found.
• ▸Flow-shortening option: cross-dock, supplier backhaul, pick method change or 3D printing, with its prerequisites (ASNs, barcode/RFID, IT system, loading bays).
• ▸As-is to to-be change: the process change chosen (re-engineering, lean, Six Sigma).
• ▸Re-measure result: the three metrics after the change and the date of the next cycle.

1. 1Start from the business case, not the tool. The strategy passage says a supply chain strategy involves change, so build a business case that justifies the effort and resources, and write it in the financial terms Finance cares about (cash flow, profitability, tax). A strong financial case also generates commitment towards implementation.
2. 2Consult the three functions the passage names before you put numbers in the case. R&D gives the future product portfolio and new-product plans. Sales and marketing give commercial plans, long-term forecasts and geographical ambitions. Finance gives the financial imperatives and constraints. Each one changes what the savings or revenue gain is worth.
3. 3Quantify every benefit alongside the ROI report. For a TMS, the passage says fleet efficiency, labour productivity and network design support all 'need to be quantified and presented alongside the ROI report.' Put each benefit on its own line with an owner, so none stays a vague 'efficiency gain'.
4. 4Separate capital avoided from capital spent. Count freed capacity and avoided investment as ROI. The AstraZeneca strategy used lean and demand-driven supply to create 'capacity headroom for new products without major capital investments', and used outsourcing 'to avoid significant capital investments and manage risk'.
5. 5Treat inventory as a cash lever. The AstraZeneca strategy included 'inventory optimization to free up cash for investment in licensing opportunities.' Show the cash released as a separate line, and name what the cash will fund.
6. 6Test the claimed benefits against evidence. Visit reference sites for any system you plan to buy, and speak to the operators, not only the managers who chose the software. Ask what benefits the system has brought since implementation, and compare them with the vendor's claims before you put them in the ROI report.
7. 7Protect margin through channel and dispatch choices. Where you fulfil several channels, the passage says stock visibility and the ability to choose the most cost-effective dispatch location (DC, fulfilment centre, store, supplier) will improve efficiency and safeguard margins. Score your capability with the omni-channel self-assessment to see where you stand.
8. 8Run the project, then re-review. Implementation 'will not happen by accident': set a clear plan with monitoring steps. Review the strategy at least annually and expect a revision about every three years. Re-test the ROI figures at each review.

AstraZeneca, formed by the 1999 Astra–Zeneca merger, first ran an operations strategy focused on integration and launching major new products. In 2001 it rewrote this as a supply chain strategy aimed at poor customer service and high inventories. The levers were lean and demand-driven supply (shorter lead times, capacity headroom for new products without major capital investment), product lifecycle management, more outsourcing, and a shift from a manufacturing focus to customer service and supply. The results the passage reports: de-bottlenecked supply of certain products, increased sales, inventory reduced by circa $500 million, and customer service above 99 per cent. In 2006, with the new-product pipeline thinner and patents about to expire on major brands, the strategy was revised again. It added a push to reduce cost of goods through continued lean, end-to-end supply chain management with advanced planning systems, inventory optimization to free cash for licensing opportunities, and more outsourcing to avoid significant capital investment. The financial impact has three parts: higher sales, about $500 million less capital tied up in stock, and avoided capex. The passage reports no ROI percentage, so state the three effects separately rather than inventing a single return figure.

• If the strategy requires change and resources, then write a business case first; a strong financial case also builds commitment to implement.
• If a benefit cannot be quantified (fleet efficiency, labour productivity, network design), then leave it out of the headline ROI and list it as unquantified upside.
• If you can add capacity through lean or demand-driven supply instead of new assets, then price the avoided capital investment as part of the return.
• If inventory is high and cash is needed for investment, then target inventory optimization and name the investment the freed cash funds.
• If a vendor's pitch rests on features, then discount it until reference-site operators confirm the benefits delivered since implementation.

• Counting only cost cuts. The tell: the case shows no sales, service-level or capital-avoidance effects, even though the AstraZeneca example gained on all three.
• Building the ROI report from vendor claims. The tell: you have spoken only to the managers who chose the software and not to the operators at reference sites.
• Skipping Finance, Sales and R&D in the case. The tell: the forecast, product plans or cash constraints in your numbers contradict their own plans.
• Treating the strategy as set-and-forget. The tell: no annual review, no monitoring steps, and the savings case is more than three years old.
• Buying on price alone for a large purchase. The tell: the vendor was chosen through an e-auction, even though the passage says success comes down to the relationship with the people at the vendor.

• ▸Initiative and owner: name the change (eg TMS, lean programme, inventory optimization) and the accountable manager.
• ▸Functions consulted: record R&D, Sales and marketing, and Finance inputs and the date of each.
• ▸Quantified benefit lines: list each benefit (fleet efficiency, labour productivity, network design, inventory reduction, sales gain) with a value and basis.
• ▸Capital avoided: list major capital investments avoided through capacity headroom or outsourcing, with value.
• ▸Cash released: state the inventory or cash reduction and what it will fund.
• ▸Cost and resources required: list one-off and running costs, people and time.
• ▸Evidence check: record reference sites visited, operators spoken to, and benefits confirmed since implementation.
• ▸Monitoring steps and dates: name the milestones and measures used to track delivery against the case.
• ▸Review date: set the next annual review and the three-year revision point.

1. 1Start with the three audits the toolkit provides: warehouse audit (1.1), transport audit checklists (2.1) and inventory management audit (3.1). Score each function against the tools it should be using, so you see where tools are missing rather than assuming.
2. 2Build a tool register per function. Warehouse: 5S (1.2), Pareto/ABC (1.3), slotting (1.7), order-picking strategy (1.4), aisle width (1.22). Transport: load and pallet configuration (2.6), fuel efficiency (2.4), road freight charges (2.8). Inventory: ABC Pareto (3.2), cycle counting (3.5), reorder point (3.9), periodic review (3.8), safety stock (3.14), EOQ (3.11). Record for each tool whether it is in use, who owns it, and when the calculation was last run.
3. 3Run ABC/Pareto first, because it sets priorities for the other tools. Use it on warehouse items (1.3) and on stock (3.2). Then apply the heavier tools, such as EOQ and tighter counting, to the vital few. The toolkit also lists combining Pareto with EOQ for group analysis (3.12).
4. 4Choose the order-picking method from the order profile, using the comparison in Table 1.5. Check the operating conditions the table lists: system assistance for cluster and batch, near-equal activity per zone, and extra sort space for batch.
5. 5Calculate before you buy or re-rack. For a counterbalance truck, aisle width = WA + LLC + L, plus 300 mm operator clearance (Tony Sellick's suggestion). For a reach truck, aisle width = WA + LLC + L − R, plus 300 mm. Take WA, LLC and R from the specific truck's data, and ask racking and MHE suppliers to check the result.
6. 6On the transport side, test whether delivery vehicles can collect from nearby suppliers on the return leg (factory gate pricing, 4.4). Check vehicle suitability and driver time before you commit.
7. 7Close the loop with measurement and a cycle. Put warehouse KPIs (7.4) and freight performance measures (7.3) on each tool, and review with PDCA (6.7) or DMAIC (6.3). A tool counts as applied only when its output changes a decision and a measure tracks the result.

The passages give formulas and decision criteria but no site data, so this example uses the toolkit's own rules and leaves the measurements as inputs. (1) A retailer ships e-commerce orders, with hundreds of orders for single-line items. Table 1.5 says batch pick to zero is very effective for this profile. It also says the method needs system assistance to combine orders and cannot easily separate urgent orders. The retailer therefore confirms its system can consolidate orders and sets a separate route for urgent orders. (2) It is buying reach trucks for new racking. It takes WA, LLC and R from the truck specification and the load length L from its pallet. It computes WA + LLC + L − R, adds 300 mm, and sets the aisle at that minimum. It does not reuse a counterbalance figure. (3) Its delivery vehicles run from the DC to stores and return part-empty past supplier sites. Under factory gate pricing it schedules supplier collections on those return legs, subject to truck capacity and driver hours. The passage cites studies showing logistics-related costs can fall 5 to 8 per cent. The retailer treats that as a range to test against its own baseline.

• If orders are mostly single-line and arrive in hundreds (typical e-commerce), use batch pick to zero. If urgent orders must be isolated, keep individual order picking for them, because batch and cluster cannot separate them easily.
• If you choose zone picking, first confirm each zone carries near-equal activity and that put-away is accurate. Otherwise change the zone boundaries or use another method.
• If you are specifying forklifts or racking, calculate the aisle from WA, LLC and L (minus R for a reach truck) plus 300 mm. Have the supplier confirm the result before purchase.
• If delivery vehicles return from stores with spare capacity and suppliers are close by, trial factory gate pricing collections. Drop lanes where vehicles are unsuitable for inbound loads or driver time runs out.
• If a tool exists in the register but its output has not changed a decision, assign it an owner and a KPI, or retire it.

• Picking method adopted without the supporting system. Cluster and batch picking depend on system assistance to combine orders. Tell: pickers sort by hand, error rates rise, and the bottleneck moves to the sort area.
• Zone picking with unbalanced zones. Tell: one zone is permanently backed up while others idle, and orders wait for consolidation.
• Aisle width taken from a rule of thumb or the wrong truck type. Tell: the 300 mm clearance is missing, or the reach distance R is ignored, and trucks cannot turn or stack in the aisle.
• Tools run once as a project and never repeated. Tell: an ABC classification, audit or EOQ figure with an old date, and stock policies that no longer match demand.
• Applying factory gate pricing on any vehicle with spare space. Tell: collections miss slots because of driver time or trucks that cannot carry the inbound goods, and the savings never show up in the numbers.

• ▸Function (warehouse / transport / inventory): which of the three areas this row covers.
• ▸Tool and toolkit reference: for example Pareto/ABC (1.3), EOQ (3.11), aisle width (1.22), factory gate pricing (4.4).
• ▸Trigger condition: the situation that calls for the tool, taken from its 'When to use' text.
• ▸Input data and source: the measurements the calculation needs, and where each comes from (truck spec, order profile, demand history).
• ▸Current status: not used / used ad hoc / used on a schedule.
• ▸Owner and review date: the named person and the date of the last and next calculation.
• ▸Decision it changes: the choice the output feeds, such as pick method, aisle width, order quantity or collection route.
• ▸Linked KPI: the warehouse or transport measure that shows whether the decision worked.
• ▸Next action: the smallest step that moves the tool one status level up.

1. 1Pick the audit scope and the auditor first. The warehouse audit (tool 1.1) must be run by an independent person from within the company or by an external consultant. For supplier audits (Ch. 2), use auditors with an operational logistics background, health and safety and road-transport competency training, and quality-auditing training.
2. 2Build the checklist from best practice, then add your own questions. The sample form rates each item Poor / Good / Excellent / N/A with a Comments column and a header for who carried it out, location and date. Items include racking condition checked regularly, weight capacity visible on rack ends, overhanging pallets, marked escape routes, and training records kept current. The source form is not exhaustive and a full version has over 100 questions.
3. 3Tell staff why the audit is happening before it starts. The passages state the purpose must be explained to staff and that results must be shared with all of them, so they own the improvements.
4. 4Audit at the place the work happens: the premises where the contract is operated, or a similar site. Ask for health and safety, training and monitoring records on request, and score evidence rather than statements.
5. 5Score each area out of 10 and plot the result on a radar chart (Figure 2.4). Highlight significant risks separately. A single average hides a weak area.
6. 6Choose KPIs from the performance chapter by operation: freight transport indicators (Table 7.3), warehouse indicators (Table 7.4), customer service indicators with standards and targets (Table 7.2). Make each SMART (tool 7.1). Group them into the Balanced Scorecard's four perspectives (Table 7.1, Figure 7.2) so no single dimension dominates.
7. 7Benchmark using the staged approach (Table 7.5, tool 7.6). The clothing-producer example (Table 7.6) shows that some peer values will be 'NK' (not known). Record them as gaps in your data and do not guess.
8. 8Close the loop. Write a report for each audit covering every area audited and suggested actions. Put checks in place so actions are verified by the next audit at the latest, earlier if non-compliance carries higher risk. Feed recurring gaps into PDCA or DMAIC (tools 6.7, 6.3).

The source supplier audit scores seven areas out of 10: Safety 8, Compliance 8, Performance 8, Contractual 9, Dependency 8, Systems 9, Food safety 10 (Figure 2.4). Plotted on a radar chart, all seven sit in the 8–10 band, so the supplier is broadly sound. The audit questions behind 'Dependency' and 'Systems' show where to probe. Is performance reliant on one or two key individuals? Could losing this contract threaten the supplier's viability, given how dependent it is on its largest other customer? Does it report on its own systems performance, and is there a disaster recovery plan? The 8 on Dependency is the lowest-scoring area that is also a structural risk, so it gets a named follow-up action in the written report. The outsourcing company then checks that action by the next audit, or sooner if the risk is higher. The same discipline applies to the warehouse form: a 'Poor' on 'Weight capacity visible on the end of the racks' gets a comment, an owner and a re-check date.

• If the auditor runs the area being audited, then bring in an independent person or an external consultant.
• If a score is high overall but one area is low or a significant risk is flagged, then act on that area; do not rely on the average.
• If the audit finding carries high risk from non-compliance, then verify the corrective action before the next scheduled audit; otherwise verify by the next audit at the latest.
• If a peer's benchmark figure is 'not known', then record it as NK and do not estimate it; fix the data source before setting a target.
• If KPIs cluster in one area (for example only cost), then rebalance across the four Balanced Scorecard perspectives.

• Audit results are kept by management only. The tell: staff cannot name the improvements they own, and the same findings recur.
• Checklist copied unchanged. The tell: questions do not match your operation (hazardous items, high-value stock) and many items return N/A.
• Findings are written up but not followed up. The tell: the next audit repeats the previous report's actions.
• Auditors lack operational or auditing training. The tell: scores rest on documents shown rather than observed practice, and risks are missed.
• Averaging a radar chart into one score. The tell: a lone weak axis, such as dependency, never reaches the action list.

• ▸Area audited: name the area (for example Safety, Systems, Dependency, racking).
• ▸Auditor and independence: who carried it out and why they are independent.
• ▸Date and location: where the audit took place and when.
• ▸Evidence seen: records or observations (training records, rack checks), not statements.
• ▸Score out of 10 (or Poor/Good/Excellent/N/A): the rating for this area.
• ▸Significant risk flagged (Y/N): describe the risk if yes.
• ▸Suggested action: the specific improvement written into the audit report.
• ▸Owner and due date: the person responsible, with the date no later than the next audit (earlier if high risk).
• ▸Benchmark/KPI comparison: your KPI value against peer value (write NK if unknown).

1. 1Start with a baseline audit. Run the supply chain management audit (Table 4.1 template) and, for the physical network, the warehouse audit (1.1). Record current state before any redesign, so each later decision is measured against it.
2. 2Set the network strategy as a cycle, not a one-off. The source frames distribution network strategy as a cycle (Figure 4.7). Write down the criteria for warehouse location (Figure 1.11 'Warehouse location criteria', tool 1.14) before listing candidate sites, so the criteria rank the sites rather than the other way round.
3. 3Design the flow inside each node. Decide where inbound and outbound doors sit. Adjacent doors need a congestion check (Figure 1.3). Opposite-side doors give less congestion but longer forklift travel. Where the node is a hub, plan inbound and outbound coordination: a vehicle cannot depart until the last vehicle carrying goods for its area has arrived.
4. 4Profile and slot the items. Place fast movers near dispatch, conveyors and aisle ends. Put high-activity items at easy-pick levels. Co-locate items sold together (product affinity, e.g. nut with bolt, dry pasta with pasta sauce). Keep frequent, moderate-weight items between waist and shoulder height.
5. 5Decide make-or-outsource with the outsourcing tools in order. Use the outsourcing decision matrix (Figure 5.1), the control-versus-risk view (Figure 5.3), the Kraljic matrix applied to logistics services (Figure 4.4) and the risk-based approach (5.3). Then use the client questionnaire (Table 5.1) and an NDA (5.5) before sharing data with candidates.
6. 6Segment suppliers and providers. Place each on the Kraljic matrix (Figure 4.3) and add the relationship type (Figure 4.9, tool 4.11). Choose the relationship style from the quadrant, not from habit or from who the incumbent is.
7. 7Score the providers. Use the LSP criteria (Table 5.3) and a weighted decision table (Table 5.4 for a regional distribution centre, Table 6.4 for a 3PL). If a lead logistics provider is on the table, work through the 4PL decision-making process (Table 5.2).
8. 8Sequence the technology last. Justify the WMS (1.15), select it (1.16), then implement it (1.17). Do the same for the TMS using its selection process (2.9). Technology follows the network and outsourcing decisions, because those fix what the system must handle.

The source's own design example is a parcel or pallet hub. Items arrive in time to meet a departing vehicle returning to its geographic area, and no vehicle leaves until the last vehicle carrying goods for its area has arrived. This makes door placement a strategic choice. With inbound and outbound doors side by side (Figure 1.3), travel is short but congestion is the risk. With doors on opposite sides, forklift trucks cross the whole building to load out, so congestion falls and travel distance rises. You choose by which cost binds: if the hub's departure cut-offs are the constraint, the extra travel may be acceptable; if throughput at the doors is the constraint, separate them. Inside the building, slotting then applies the same trade-off at item level. Fast movers sit near dispatch to cut picker travel, and affinity pairs such as nut and bolt or pasta and sauce sit together. The excerpt gives no figures for these trade-offs, so you must measure travel and congestion in your own site.

• If inbound and outbound doors are adjacent, then test for congestion before approving the layout; if they are opposite, then accept and cost the longer forklift travel.
• If a service is outsourced, then place it on the Kraljic matrix (Figure 4.4) and on the control-versus-risk view (Figure 5.3) before choosing the provider type.
• If you are comparing two or more providers, then use a weighted decision table built from the LSP criteria (Table 5.3), not a ranking from opinion.
• If you have not yet justified the WMS (1.15), then do not start vendor selection (1.16); if selection is incomplete, then do not plan implementation (1.17).
• If items are often sold together, then slot them adjacent to cut travel and separate look-alike parts.

• Choosing technology before the network and outsourcing decisions. The tell: WMS or TMS requirements are written around current layouts and suppliers rather than the designed ones.
• Picking a warehouse site without written criteria. The tell: the shortlist is justified after the fact, and the criteria list (Figure 1.11) was filled in once a favourite existed.
• Treating every supplier or provider alike. The tell: one contract style and one review cadence for all, with no Kraljic quadrant recorded.
• Ignoring door and flow layout effects. The tell: congestion at shared doors, or long forklift runs, show up only after go-live.
• Outsourcing without checking risk and control. The tell: no questionnaire, no NDA and no risk assessment on file before provider talks begin.

• ▸Baseline audit reference: Which audit (Table 4.1 or 1.1) and date establish the current state.
• ▸Location criteria and weights: List the warehouse location criteria (Figure 1.11) and the weight of each, before naming any sites.
• ▸Door configuration: Adjacent or opposite; the congestion risk or travel-distance penalty you accept.
• ▸Slotting rules: Fast-mover placement, affinity pairs, and the pick-height rule you apply.
• ▸Outsourcing position: In-house or outsourced; the Kraljic quadrant (Figure 4.4) and the control-versus-risk position (Figure 5.3).
• ▸Supplier/provider segment: Quadrant and relationship type (Figure 4.9) for each key supplier or provider.
• ▸Provider scoring: Criteria (Table 5.3), weights, scores and total for each candidate.
• ▸Technology gate status: Justified (1.15), selected (1.16) and implementation planned (1.17) for the WMS; the same for the TMS (2.9).
• ▸Review trigger: The condition that sends you back through the network strategy cycle (Figure 4.7).

1. 1Pick one process to standardize first (picking, loading, goods receiving or ordering). Write one sentence naming where it starts and stops. The source tools (flow chart for goods receiving, time-based process map) both work on a single bounded process, not a whole warehouse.
2. 2Draw the current process as it is actually run, using a flow chart (tool 6.4, with standard flow-chart symbols). Walk the floor and record what people do, not what the procedure says. Then add elapsed time to each step in a time-based process map (tool 4.16) so waiting and queuing become visible.
3. 3Build a problem matrix with the four categories from the WMS justification method: inventory accuracy, resource management, customer service, and visibility/information. List every observed problem under a category. Examples from the source: lost product, mis-picks, time wasted looking for material, inefficient pick paths, ship errors, delayed shipments, data entry errors.
4. 4Estimate the cost of each listed problem, as the source's Step 2 does. Rank by cost so you work on the expensive waste first, not the most visible.
5. 5Find root causes before choosing a fix. Use a 5 Whys chart or fishbone diagram (chapter 9) on the top-ranked problems. Then match the fix to the cause: slotting/item profiling and ABC analysis of pick list frequency for travel and search waste, location numbering for finding stock, task interleaving for empty movements, and the 5S/6S tools for workplace order.
6. 6Document the new method in a revised flow chart. Make it the single written standard for that process, and train everyone on that version.
7. 7Run the change as a PDCA or DMAIC cycle (tools 6.7, 6.3) with the 5S/6S audit as the check. Re-audit on a fixed schedule so the standard does not drift.
8. 8Choose KPIs from the warehouse KPI list (7.4) that match the problems you ranked. Record a baseline before the change and re-measure after, so each fix shows a before/after result.

The source's WMS justification method (Red Prairie, now JDA) is a standardization diagnostic in miniature. Step 1 is to build a matrix on inventory accuracy, resource management, customer service and visibility, then list the site's problems under each. A warehouse might record excess inventory, lost product and mis-picks under accuracy, and time spent looking for material, inefficient pick paths and no performance measurement under resource management. It would add ship errors and delayed shipments under customer service, and stock-outs, false stock-outs, transaction updates delayed by hours or days and data entry errors under visibility. Step 2 puts a cost against every line, which turns a list of irritations into a ranked set of wastes. The manager then maps the process behind the costliest line, for example the time-based map of goods receiving, and finds where the delay sits. Fixes are chosen by cause: location numbering or slotting for search time, ABC analysis of pick frequency for pick paths. The source does not give cost figures for this example, so you supply them from your own site.

• If people spend time looking for material, fix location numbering and slotting before buying any system.
• If travel and pick paths are the waste, run an ABC analysis of pick list frequency and re-slot the fast movers.
• If you cannot say where the delay sits, draw a time-based process map before proposing any fix.
• If a problem keeps returning after being fixed, run 5 Whys or a fishbone to find the root cause, then lock the fix into the documented flow chart and the 5S/6S audit.
• If a proposed change needs investment, cost it with the engineered approach (time-and-motion studies) or time-driven ABC. Proceed only if the result beats the cost of the problem you ranked.

• Mapping the process as designed rather than as performed. Tell: the flow chart is clean but staff say 'we don't do it like that'.
• Listing problems without costing them. Tell: effort goes to the most annoying issue, not the most expensive one.
• Treating 5S/6S as a one-off clean-up. Tell: the area looks right in the first month, with no scheduled audit scores after that.
• Fixing symptoms, such as re-picking mis-picks, without a root-cause tool. Tell: the same error type reappears in the next period.
• No baseline measure, so improvement cannot be shown. Tell: 'no means of measuring performance' still appears in your problem matrix after the project.

• ▸Process and boundaries: name the process (e.g. goods receiving) and its start and stop points.
• ▸Problem category: inventory accuracy, resource management, customer service or visibility.
• ▸Observed problem: one line describing what actually happens (e.g. mis-pick, searching for material).
• ▸Time-in-step / wait observed: elapsed time from the time-based process map for the affected step.
• ▸Estimated cost per period: your estimate for this problem; use it to rank.
• ▸Root cause (5 Whys / fishbone): the cause behind the symptom, in one line.
• ▸Countermeasure: the chosen fix (slotting, location numbering, 5S/6S, task interleaving, etc.).
• ▸Documented standard: the revised flow chart or procedure reference, plus owner and date.
• ▸KPI and baseline vs result: the warehouse KPI, its value before, and its value after, with the next audit date.

1. 1Build a problem matrix first. Use the four WMS benefit categories as rows: inventory accuracy, resource management, customer service, and visibility/information management. List the real problems under each, such as lost product, mis-picks, time wasted looking for material, ship errors, stock-outs, false stock-outs, and transaction update delays of hours or days.
2. 2Put a cost on each problem. Use standard equations to estimate the cost of each one, as in the typical cost penalty if no WMS (Table 1.10). Run ship errors at two accuracy levels (98.5% and 96%) to see how sensitive the cost is. Rank the problems by cost so the biggest cost factor gets fixed first.
3. 3Treat latency as a visibility defect. Record how long each transaction (receipt, put-away, pick, dispatch) takes to reach the stock record. Paper and label picking is 'not real time' and needs a two-stage update. Barcode guns, wearable scanners, voice and finger scanning give real-time stock update. Match the capture method to your required update speed.
4. 4Choose the capture technology against the pick-rate and location facts. Lines per hour run: paper and label under 100, gun under 100, wearable under 150, voice 100–250. Check the constraints that break visibility: a barcode on every product, a system interface, wireless coverage across the whole warehouse for real-time, and the voice problem with multi-SKU locations and serial number capture.
5. 5Write the RFI from your operational specification. Include loading bays, reserve locations, pick face locations, pick-and-pack stations, and user counts by role (admin, forklift drivers, pickers, packers). Add daily transaction volumes (receipts, put-aways, picks, dispatches) and the peaks by day, week or month. Describe goals, not how the system should work, and ask each vendor for budget cost and implementation timescale.
6. 6Screen vendors on evidence of control. Ask for company and financial history, number of sites on the current WMS, who owns the source code, client list in your market, daily rates and support charges, support cover, development plans and track record. Do not build in-house unless requirements are specialized or need specialized integration with existing systems.
7. 7Calculate ROI, get board acceptance, then select the supplier. Budget the real 'cost' too: commitment from the whole warehousing team and senior management to set the system up correctly, use it properly and regularly optimize it.
8. 8Use the WMS as an execution system. Wave pick needs a WMS to manage allocation against departing trucks. Automatic replenishment, labour management and remote data visibility are the controls that make the visibility actionable.

A warehouse manager makes the case for a WMS using the five-step justification approach. In the problem matrix, the site lists excess inventory, lost product and mis-picks under inventory accuracy. Under resource management it lists time wasted looking for material and no means of measuring performance. Under customer service it lists ship errors and delayed shipments, and under information management it lists stock-outs, false stock-outs and update delays of hours or days. For ship errors the manager runs the cost calculation twice, once at 98.5% ship accuracy and once at 96%, and compares the two costs. The biggest cost line becomes the headline of the ROI calculation presented to the board. After approval, the operational specification feeds the RFI, which states bays, locations, user counts and daily transaction peaks. Vendors are then compared on budget cost, timescale, number of live sites, source-code ownership and support cover.

• If stock records lag by hours or days and you pick on paper or labels, then move to scanning or voice, because those methods update stock in real time.
• If your picking needs hands-free, eyes-free work in temperature-controlled areas or with heavy, awkward items, then consider voice picking (100–250 lines per hour). If you have multi-SKU locations or need serial number capture, add a scanner or choose another method.
• If you release orders in timed waves to meet vehicle departures, then you need a WMS to manage the allocation. Do not try it manually.
• If your requirements are standard, then buy a packaged WMS, because vendors spread development costs over 100+ clients who also proved the product. Build in-house only for specialized requirements or specialized integration with existing in-house systems.
• If you cannot show quantified benefits or secure team and senior management commitment, then do not start a WMS project, because it is not a quick fix.

• Buying real-time hardware without the infrastructure. The tell: real-time scanning is specified but there are no wireless receivers throughout the warehouse, or products lack barcodes, so stock updates stay delayed.
• Justifying the WMS on intuition. The tell: the business case lists no cost per problem, so the board cannot see which problem is the biggest cost factor and the ROI is not accepted.
• Dictating how the system should work in the RFI. The tell: the document is long and prescriptive, vendors just confirm it, and no faster, better or cheaper alternatives are offered.
• Treating the WMS as a one-time install. The tell: after go-live nobody optimizes it, usage drifts, and the original problems (mis-picks, false stock-outs) return.
• Choosing a pick technology that ignores your SKU and location structure. The tell: voice picking is deployed with multi-SKU locations and picks go wrong, or accuracy suffers when product sits in an incorrect location.

• ▸Problem category: inventory accuracy, resource management, customer service, or visibility/information management.
• ▸Specific problem observed: for example mis-picks, lost product, false stock-outs, ship errors, or update delays of hours or days.
• ▸Estimated cost of the problem: from the standard equation, with the inputs stated (for ship errors, run at 98.5% and at 96%).
• ▸Cost rank: order the problems from the biggest cost factor down.
• ▸Current data capture method: paper, pick by label, barcode gun, wearable scanner, or voice.
• ▸Stock update timing today: real time or delayed, and by how long.
• ▸Infrastructure gaps: barcodes on every product, system interface, wireless coverage, noise level, and multi-SKU or serial number issues.
• ▸RFI inputs: bays, reserve and pick locations, user counts by role, daily transactions and peaks.
• ▸Vendor evidence: live sites, source-code owner, support cover and rates, development plans, budget cost and timescale.