RIDBS · Distribution Operations

How the Distribution Centre Decides What Is on Your Shelf

An execution-ready, brand-agnostic guide to supermarket distribution-centre operations — for franchisees, independents, store and area managers, cash-and-carry operators, and family-owned retail. Built for the realities of South African retail: loadshedding, cold-chain risk, cargo crime, and the move to dark stores.

3
Temp zones
Ambient · chilled · frozen
24/7
Multi-shift
operations
0
Brand names
Generic model
13
Operations
plays
A live-style snapshot of one DC shift. Green tags = on or above target, red = a shortfall, amber = watch. Real WMS systems export these metrics continuously.
DC Shift · Live LIVE
Inbound cases40,180▼ 4%
Cases picked47,210▲ 2%
Pick accuracy99.51%● watch
On-time dispatch96.4%▼ risk
Chilled zone temp3.6°C● ok
Ch 4 · DC Economics
unit cost & payback
Ch 5 · P&L Sample
5-year ROI model
Ch 9 · DC Types
cold chain to co-pack
Ch 11 · Risk Matrix
loadshedding · crime
Chapter 01 · Orientation

Why the DC Decides Your Shelf

Read left to right: a supplier delivers bulk into the Distribution Centre, which breaks the load into mixed, store-ready drops for many outlets. Concentrating volume in one place is what makes central distribution cheaper per case. From supplier to DC to stores Supplier Distribution Centre WMS / Cold Chain Stores
One supplier truck in → mixed-load deliveries out to many stores.

In one line: the building you never see — the distribution centreA centralised warehouse that receives bulk stock from suppliers and redistributes mixed, store-ready loads to retail outlets. (DC) — is the machine that decides whether your shelves are full or empty on a Monday morning. This playbook explains how it works, what it costs, and where it breaks.

Whether you run a franchise supermarket on the KwaZulu-Natal coast, an independent store in a township, or a cash-and-carry serving spaza shops, your business runs on a logistics model most owners never fully see. The shift over the last twenty years has been total: from dozens of supplier trucks fighting for space at your back door, to a handful of mega-warehouses that run twenty-four hours a day and dispatch a single mixed-load truck to your store.

This document tells that story operationally — with a realistic model of how a large DC is sized, staffed, and funded; a synthetic P&L and payback calculation; the systems that move stock and data; and the pain points (loadshedding, cargo crime, rigid delivery windows) that hit your pocket directly. It is deliberately brand-agnostic: no chain names appear anywhere, and all financial figures are clearly labelled illustrative / synthetic so you can recalibrate them to a real client.

How to read this playbook. Chapters 1–3 are the story and the floor-level reality. Chapters 4–6 are the money. Chapters 7–10 are systems, suppliers, DC types, and technology. Chapters 11–13 are risk, the future, and what to do tomorrow morning. Hover or tap any underlined termTerms with a dotted underline open a short definition on hover (desktop) or tap (mobile). Keyboard users can Tab to them. for a definition.

Operations Play — Chapter 1

  • Map every product on your top-shelf to its source: DC delivery, direct-to-store, or cash-and-carry collection.
  • Identify the single category most exposed to a DC outage (usually fresh / chilled) and write a one-page backup plan.
  • Walk your receiving bay this week and time how long a mixed-load delivery takes to check in.
Chapter 02 · History

From the Back Door to the Mega-Warehouse

The model inverted. Left — the old world: many small supplier trucks each visiting one store. Right — today: one regional warehouse supplying many stores on consolidated routes. Costs and receiving chaos both fall sharply. Evolution from direct-to-store delivery to central distribution Pre-2000s · Direct-to-store many trucks, 1 store Today · Centralised DC 1 warehouse, many stores
The inversion: from many trucks per store, to one warehouse per region.

The shift: retailers moved from direct store deliveryDirect Store Delivery (DSD): each supplier drives its own truck to each store. High control, high cost, heavy on the back door. (DSD) — where every supplier drove to every store — to centralised distribution, where suppliers drop bulk at one regional hub and a single mixed-load truck serves each store. The chains that moved early won a logistics lead the late movers are still closing.

The old world: when the back door never stopped

Through the 1970s, 80s and into the 90s, the receiving bay was the busiest part of a supermarket. The bread van, the milk truck, the soft-drink truck, the spice rep, the detergent rep — each arrived separately, each wanted a signature, each needed dock space, and each wanted their invoice checked on the spot. Stores held weeks of inventory in large stockrooms and employed receiving staff whose day was spent arguing about short dates and chasing credits for damages.

For independents and franchisees, the real cost was buying power. A single store negotiated alone, and a supplier could quietly give a bigger chain a better deal down the road. Large retailers knew centralisation was more efficient for decades, but major suppliers pushed back hard — they liked controlling the delivery schedule and the direct relationship with each store manager, and they offered their best discounts to retailers who kept delivering store-by-store1.

The turn: central distribution takes over

By the mid-2000s, the decentralised model of the largest chains was bleeding money at scale — costs rising, availability patchy, back-door chaos unsustainable. One major retailer announced a phase-in of centralised distribution, requiring suppliers to drop goods at regional hubs instead of individual stores4. A single company truck would then carry mixed loads to each store.

The flagship hubs that followed set the template for the whole industry:

Illustrative profile of large retail distribution hubs (generic, non-branded). Figures are industry-typical ranges.
Hub characteristicIllustrative scaleSource / note
Flagship inland DC footprint~50,000 sqmR600m+ build, ~280+ stores served5
Largest single-roof DC on the continent~114,000–123,000 sqmConsolidates several older DCs67
Truck doors at a mega-DC130+ doors~20,000 products from ~600 suppliers8
Regional grocery DC network (one group)6+ regional DCsPlus satellites to cut transport cost10
All figures are industry-typical ranges. Each citation links to its documented source — hover the superscript number to see it, click to open.
The cost of waiting. The chain that centralised roughly fifteen years earlier than its rival converted that lead into lower prices and larger market share. The chain that waited is still paying to close a logistics gap built decades ago. For owners: the lesson is not abstract — late investment in supply chain is paid back in lost margin for years.

Operations Play — Chapter 2

  • List which of your lines still arrive direct-to-store vs via DC; DSD lines are your flexibility, DC lines are your volume.
  • Estimate the receiving-staff hours you spend per week on DSD invoice checks — that is the cost centralisation removes.
  • If you are independent or in a buying group, benchmark your delivered cost vs a franchised peer for one shared SKU.
Chapter 03 · Floor Reality

Inside a Distribution Centre

Modern pallet racking stands six levels high. A picker rarely uses paper — a voice headset directs them to aisle, bay and case count. Note the electric forklift: large floors run almost silently because both the trucks and the picking are hands-free. Interior of a large distribution centre showing tall pallet racking and an electric forklift Pallet racking — six levels high
A modern DC is a series of specialist zones, not one big room.

The reality: a major DC runs 20–24 hours a day across multiple shifts, with receiving, put-away, cross-dockCross-docking (flow-through): inbound goods are moved straight from the receiving dock to the outbound dock, with little or no storage in between. Used for fast-moving and fresh lines., voice-guided picking, route-planned dispatch, and a returns/waste stream. Every zone has its own equipment, staff and procedures.

The shift pattern

Most major DCs run three shifts, often continuously. A single mega-DC can employ on the order of 3,000–4,000 people directly, plus indirect contractors7. Typical roles: receiving clerks, put-away teams, pickers, dispatch loaders, forklift and MHEMaterial Handling Equipment: forklifts, reach trucks, pallet jacks, conveyors and automated guided vehicles used to move goods inside the DC. drivers, quality controllers, maintenance, and security.

Receiving — where it all starts

Supplier trucks are checked in at the gate and backed into dozens — sometimes over a hundred — receiving doors. One large inland DC has roughly 260 dock levelers across receiving and dispatch12. The receiving clerk checks the delivery note against the purchase order, scans goods into the WMSWarehouse Management System: the software that controls receiving, put-away, inventory, picking and dispatch — the digital brain of the DC., and flags fresh produce, meat and dairy for quality checks. Clean stock then flows one of two ways:

  1. Put-away: pallets are stored in racking — in the ambient section, stacked up to six levels high. The system chooses the location to optimise density and FEFOFirst-Expired-First-Out: stock rotation by expiry date (a date-driven version of FIFO) so the oldest-dated goods ship first..
  2. Flow-through / cross-dock: fast-moving and fresh lines never touch a shelf — they move straight from inbound to outbound dock.

The three temperature zones

A modern DC is split into specialist environments. You cannot move equipment freely between them without a temperature lock or equipment switch.

DC temperature zones (illustrative)
ZoneTypical rangeHoldsNotes
Ambient15–25 °CDry groceries, toiletries, cleaning, cansMay include a temperature-controlled sub-zone for heat-sensitive goods7
Chilled0–5 °CDairy, fresh meat, deli, prepared mealsSpeed-critical; every minute outside the chain is risk
Frozen≤ −18 °CIce cream, frozen veg, frozen meatMost energy-intensive; strict dock-seal discipline

Picking — getting your order ready

When a store orders, the WMS breaks it into picks. A picker receives instructions through a voice headset — no paper, no shouting — telling them the aisle, bay and case count8. The stock goes onto a pallet or into a tote, is scanned, and moves to dispatch. In leading DCs the floor runs in near-total silence because picking is voice-driven and the trucks are electric8.

Picking methodologies & inventory accuracy

How a DC picks decides how fast and how accurately your store gets its order. There is no single “best” method — the choice depends on order profile, volume, and whether the goods are ambient or temperature-controlled. The four methods you will hear about are:

  • Discrete (single-order) picking: one picker walks one order end to end. Simple to train and manage, but wasteful of travel time — common in smaller DCs or for unusual orders.
  • Batch picking: one picker collects lines for several orders at once into separate totes, then sorts. Cuts walking dramatically for high-volume, small-line orders.
  • Zone picking: the DC is divided into zones (often by temperature or category), and a picker owns one zone; an order is assembled as it passes through. Essential for mixed-load grocery because ambient, chilled and frozen must stay in their own environments.
  • Wave picking: orders are grouped into waves released against the shift plan, the truck schedule, and the cut-off times. Wave timing is what gets your order onto the right truck, in the right loading sequence.

Underpinning all of this is inventory accuracy — the percentage of locations where the system’s record matches what is physically on the shelf. Best-in-class DCs run at 99% or better, sustained by cycle countingCycle counting: a continuous, rolling count of a small subset of inventory each day, replacing the disruptive full annual stock-take, so accuracy is maintained year-round. rather than a single annual shutdown stock-take. Low accuracy is the hidden cause of “phantom stockouts” — the system thinks a product is available, the picker arrives at an empty location, and your store ships short. For a store owner, every short ship is a lost sale and another credit note to chase.

Dispatch — loading the trucks

This is the logistics brain. Route-planning software decides which truck carries which stock and in what order it is loaded. The last store on the route is loaded first, so when the truck arrives at the first store, that store’s stock is right at the doors8. A single truck carries a mixed loadA single delivery combining ambient, chilled and frozen product, separated by insulated curtains or compartments so one truck serves a store’s full needs. — frozen, chilled and ambient separated by curtains or compartments.

Returns, recycling and waste

DCs also reverse-flow damaged goods, expired product, recalls and unsold promotional stock to a dedicated returns area for sorting, crediting or destruction. At modern sites every used cardboard box is baled on-site into compact “snakes” and sent to recyclers8.

Example document — DC shift KPI dashboard

Illustrative / synthetic. A real WMS exports these live; this shows the shape of the data a DC manager watches each shift.
KPITargetShift actualStatus
Inbound cases received42,00040,180▼ -4%
Cases picked46,50047,210▲ +2%
Pick accuracy99.6%99.51%● on target
On-time dispatch %98.0%96.4%▼ risk
Chilled-zone temp (avg)≤ 4 °C3.6 °C● ok
Dock door utilisation85%91%● ok
Damage / shortage rate≤ 0.4%0.36%● ok
Cold-chain maths. Food-safety guidance suggests product temperature can rise by roughly 1 °C for every ~10 minutes a chilled line sits outside the chain13. Above 5 °C (chilled) or −18 °C (frozen), product is compromised. This is why dispatch speed and dock-seal discipline are measured, not assumed.

Operations Play — Chapter 3

  • Ask your DC/wholesaler for a monthly service report: on-time %, pick accuracy, damage rate — if you can’t see it, you can’t manage it.
  • Train receiving staff to record temperature on every chilled/frozen delivery before signing — that record is your credit evidence.
  • Pre-stage your receiving area before the delivery window: dock free, cold room space reserved, scanner charged.
Chapter 04 · Economics

The DC Economic Model

Three illustrative footprints, from a small regional shed to a mega national hub. Per-case cost falls as volume rises because fixed cost spreads over more units — a half-empty DC is the most expensive one to run. DC size bands and throughput SmallMediumMega 8–15k15–40k40–120k+ sqm footprint
Three illustrative DC size bands. Numbers are industry-typical ranges.

Sizing the machine: this chapter gives a granular, brand-agnostic operational model — footprint, throughput, headcount, capex and an operating-cost breakdown — for large DCs serving multiple outlets. All figures are illustrative industry-typical ranges; recalibrate to the client’s actuals.

Facility size, throughput and headcount bands

Illustrative DC operational bands (generic). Replace ranges with documented references before publishing.
BandFootprint (sqm)Annual throughputHeadcountIndicative capex
Small / regional8,000–15,0006–12m cases/yr150–400R150–400m
Medium / multi-temp15,000–40,00015–45m cases/yr400–1,200R400m–R1.2bn
Mega / national hub40,000–120,000+50–120m+ cases/yr1,500–4,000+R1.2bn–R3.5bn+
Throughput can also be expressed in pallets: a mid-size DC may move 1.5–4m pallets/year; mega hubs multiples of that. Capex rises steeply once refrigeration, automation and specialised racking enter the build.
33–38%Labour share of opex
20–24%Transport / fleet share
8–10%Utilities / energy share

Illustrative operating-cost breakdown

Where a DC’s running cost actually goes. Percentages are typical industry blends for a multi-temperature retail DC; chilled-heavy or automated sites skew differently.

Cost line% of opex (typical)Primary driver
Labour (direct + supervisory)33–38%Headcount, shifts, wage inflation
Transport & fleet20–24%Distance to stores, fuel, tractor-trailer fleet
Facility rent / finance11–13%Capex amortisation, lease, land
Utilities & energy8–10%Refrigeration, lighting, generator diesel
Equipment amortisation & maintenance7–9%Forklifts, racking, conveyors, MHE leases
IT / WMS / telecoms3–4%Software licences, voice/data, integration
Security & risk2–4%Guards, tracking, escorts, CCTV
Other (insurance, waste, consumables)4–6%Packaging, recycling, insurance premiums

Worked example: unit economics & break-even

Illustrative arithmetic for a medium ambient-leaning DC. Numbers are round-figure placeholders to show the method, not a quote.

Take the DC above: it handles 28 million cases a year with R238m of operating cost. That works out to roughly R8.50 of operating cost per case handled. Add the recovery margin target and the cross-dock premium, and the DC’s blended handling fee lands around R9–10 per case before delivery. Delivery then adds a per-drop and per-kilometre component on top (see the fee menu in Chapter 6).

Illustrative cost-per-case build-up (synthetic).
StepPer caseBasis
LabourR 3.10~36% of opex ÷ 28m cases
Transport & fleetR 1.93allocated share
Facility & energyR 1.82rent/finance + utilities
Equipment, IT, security, otherR 1.65remaining lines
Total operating cost / caseR 8.50
Margin / recoveryR 1.10~13% on cost
Blended handling feeR 9.60before delivery

Break-even: the DC’s fixed costs (facility, equipment, core staff, IT) do not fall much if volume drops, so cost-per-case rises sharply below the design throughput. If this DC ran at 60% of plan (~17m cases), the same fixed cost spreads over far fewer units and per-case cost climbs — which is exactly why under-utilised DCs destroy value. The operational lesson is blunt: a DC is a volume machine; running it part-empty is more expensive than running it full. That single fact is why retailers fight so hard to consolidate volume into fewer, fuller hubs.

What drives cost up or down? (expand for the levers)
  • Temperature mix: frozen/chilled adds refrigeration capex and energy — a chilled DC can cost 2–3× more per sqm to build than ambient.
  • Automation: conveyors, voice and AGVs cut labour but raise capex and IT spend; the trade-off depends on wage levels vs capital cost.
  • Catchment radius: every extra kilometre to a store raises transport cost — site selection near demand clusters is decisive.
  • Energy resilience: generator/solar/battery backup is now non-negotiable in SA and adds 5–15% to energy cost1516.
The “security tax”. Cargo-crime cost has climbed from roughly 0.3% of operating cost a decade ago to around 2.3% by end-2025 — and up to ~4.1% for high-value goods19. That line item flows straight into the per-case price stores pay.

Operations Play — Chapter 4

  • Model a DC’s break-even cases/year: below it, unit cost spikes; above it, marginal cost falls — know where you sit.
  • Challenge the transport line: route optimisation and load consolidation typically cut 4–8% off fleet cost.
  • Separate “productive” energy from “resilience” energy in your P&L so backup-power cost is visible, not hidden in utilities.
Chapter 05 · Financials

Sample P&L & 5-Year Payback

Cumulative net benefit fills the building toward its build cost over five years. In this synthetic model the DC breaks even at roughly year six; ambient DCs repay faster because they carry no refrigeration capex. Payback: cumulative benefit filling toward capex Chilled DC Cumulative benefit → approaches capex by ~Year 6
Illustrative 5-year cumulative benefit ramping toward the build cost.

The money: a captive retail DC usually runs as a cost-recovery operation (low single-digit EBITDA) — its real return is captured system-wide as lower delivered cost, fewer stockouts and less spoilage. Below: a synthetic P&L, then a 5-year payback for a medium chilled DC.

Illustrative DC P&L (synthetic)

Figures in R millions per year. Synthetic example for a medium multi-temperature DC (~30,000 sqm, ~28m cases/yr). Recalibrate to client.
LineR m / yrNote
Per-case distribution fees (charged to stores)168.0Primary recovery mechanism
Cross-dock / flow-through handling fees39.0Fresh & fast-moving lines
Supplier services (slotting, co-pack)24.0See Chapter 8
Ancillary (returns, recycling rebate, 3PL overflow)17.0Third-party throughput
Total throughput revenue248.0
Labour(86.0)~35% of opex
Transport & fleet(54.0)Fuel, tractors, trailers, maintenance
Facility rent / finance(29.0)Capex amortisation + lease
Utilities & energy (incl. generator diesel)(22.0)Refrigeration-led
Equipment amortisation & maintenance(18.0)MHE, racking, conveyors
IT / WMS / voice / data(8.5)
Security & risk(7.5)
Consumables, insurance, waste, other(13.0)
Total operating expenses(238.0)
EBITDA10.0~4.0% margin — typical cost-recovery
Depreciation & amortisation(22.0)
Net finance / interest(11.0)
Pre-tax result(23.0)Book loss is expected for a captive DC
Why a “loss” is fine. A captive retail DC is not a standalone profit centre. The negative pre-tax result is offset system-wide by avoided DSD cost, lower procurement prices from national buying, reduced spoilage, and freed-up selling space in stores. The payback calculation below captures that system benefit.

Illustrative 5-year payback — medium chilled DC

Capex: R720m (refrigeration, insulated panels, dock seals, specialised racking, backup power). Net annual benefit = avoided DSD cost + handling revenue + spoilage reduction, less opex. R millions.
YearNet benefitCumulative% of capex recoveredPayback signal
0 (build)(720.0)(720.0)0%Capex outlay
195.0(625.0)13%Ramp-up
2118.0(507.0)30%Volume matures
3132.0(375.0)48%Steady-state
4141.0(234.0)67%Optimisation
5148.0(86.0)88%Near break-even
6150.064.0109%✓ Payback (~5.6 yrs)

Reading the result: a medium chilled DC pays back in roughly 5–6 years in this synthetic model. Ambient DCs (no refrigeration) typically pay back faster — around 3–4 years — because capex is far lower. Highly automated or fresh-meat-processing centres take longer, often 7–10 years, unless co-funding from suppliers shortens it.

Adapting these figures to a specific country / province cost base

To re-baseline this model to a target region, adjust four levers and let the rest follow:

  • Labour rate: replace the wage assumption with local cost-per-head per shift; labour is 33–38% of opex, so this moves the P&L fastest.
  • Energy tariff: set the local R/kWh and diesel price; for SA, factor loadshedding backup cost (generators run at ~5× grid cost15).
  • Land & rent: provincial land/rent per sqm sets the facility finance line.
  • Freight distance: average km to a store drives the transport line — rural networks cost more per case.

As a rule of thumb: hold throughput constant, swap these four inputs, and the P&L and payback recompute coherently. Always label the output illustrative until verified against real tenders.

Operations Play — Chapter 5

  • Present DC investment as system ROI, not standalone EBITDA — the avoided-cost line is where the case is won.
  • Stress-test payback against a +15% energy and +20% security-cost shock; if it still pays back under 8 years, it is robust.
  • Negotiate supplier co-funding (slotting/co-pack) to pull the payback year forward.
Chapter 06 · Commercial Model

How DCs Pay for Themselves

Five reinforcing levers fund the DC. No single lever carries the whole load — distribution fees and supplier money flow in from the left, while procurement power and cross-dock savings enter from the right. Five revenue/cost-recovery levers Fees Scale Slots DC cost recovery Procure X-dock
Five levers — fees, scale economies, slotting, procurement power, cross-dock savings — that fund the DC.

The funding model: a DC “pays for itself” through five reinforcing levers — distribution fees, supplier contributions, scale economies, cross-dock savings, and centralised procurement power. Rarely does one lever carry the whole load; the business case is the stack.

The five commercial levers

LeverHow it worksTypical effect
Distribution feesStores (or the group) pay a per-case / per-pallet handling & delivery feeRecovers 40–60% of operating cost
Supplier contributionsSlotting fees, delivery-window premiums, co-funding, co-pack revenue5–15% of throughput revenue
Scale economiesFull-truckload inbound, lower unit handling, shared fixed cost over more volumePer-case cost falls as volume rises
Cross-dock savingsFresh/fast lines skip storage — less labour, less spoilage, faster cycleCuts fresh-line handling cost materially
Centralised procurementOne national deal replaces many store-level negotiationsLower cost-of-goods flows to every store

Example fee structures (illustrative)

Synthetic fee menu — adapt to client commercial terms.
Fee typeIllustrative rateCharged to
Ambient case handlingR 0.85 / caseStore / group
Chilled case handlingR 1.40 / caseStore / group
Frozen case handlingR 2.10 / caseStore / group
Delivery (per drop, per km band)R 350 + R 9/kmStore / group
Slotting fee (new line listing)R 8,000–25,000 / SKUSupplier
Cross-dock / flow-throughR 0.55 / caseSupplier or store
Energy / resilience surcharge5–15% on handlingStore / supplier16
The compounding effect. A store that previously received 40 supplier trucks a week might receive 3 DC mixed-loads instead. Less dock congestion, less receiving labour, less damage from repeated handling, and more selling floor once the stockroom shrinks. Those savings are real margin — they just do not appear on the DC’s own P&L.

Operations Play — Chapter 6

  • Quantify avoided DSD cost (trucks/week, receiving hours, damage rate) — it is the strongest argument for the model.
  • Price a transparent fee schedule: per-case by temperature, per-drop, per-km. Predictability beats lowest price.
  • Ring-fence supplier contributions (slotting/co-pack) to fund capex payback, not day-to-day opex.
Chapter 07 · Integration

Stores, Systems & Integration

A data packet rides the chain end to end: Supplier → DC inbound → WMS/WCS → cross-dock/replenishment → store order → store receiving. Each arrow is an automated message or API call — the store order itself is rarely typed by hand. Systems integration flow Supplier DC inbound WMS /WCS Cross-dock /replenish Store order Receiving A data packet rides the chain end-to-end
End-to-end flow: Supplier → DC inbound → WMS/WCS → cross-dock/replenishment → store order → store receiving.

The nervous system: a store order is rarely typed by hand anymore. It is generated from POSPoint of Sale: the in-store till system that records every sale; its sell-through data feeds replenishment and forecasting. sell-through and demand forecasts, pushed to the WMS over EDI/FTP or an API, fulfilled by the DC, and confirmed back to the store on receipt.

How a store order is created

  1. POS signals: every scan at the till updates store inventory. When stock hits a reorder point, a replenishment need is raised.
  2. Demand forecast feed: a forecasting engine adjusts the order for day-of-week, promotions, seasonality and weather.
  3. Order creation: the replenishment engine builds the order and transmits it to the DC via EDIElectronic Data Interchange: standardised B2B messages (e.g. orders, ASN advance-ship notices, invoices) exchanged between systems., SFTPSecure File Transfer Protocol: an encrypted file drop used to exchange batched order/inventory files between trading partners. file drop, or a real-time APIApplication Programming Interface: a way for two systems to talk in real time over the web, exchanging small structured requests and responses. call.
  4. DC fulfilment: the WMS converts the order into picks; the WCS orchestrates conveyors/equipment; the TMS plans the route.
  5. Store receiving: an ASNAdvance Ship Notice: an electronic message telling the receiver exactly what is on the truck, before it arrives — enabling fast, accurate check-in. tells the store what is coming; the store scans it in and reconciles.

How forecasting & replenishment actually work

Behind every auto-generated store order is a forecast. Modern demand-planning engines do not simply repeat last week’s sales — they combine historical sell-through with a stack of adjustment factors: day-of-week and day-of-month patterns, public holidays and pay-week effects, planned promotions and price changes, seasonality, and increasingly external signals such as weather and local events. The output is a predicted demand curve per SKU per store, usually looking a few days to a few weeks ahead.

The replenishment engineReplenishment engine: software that turns demand forecasts plus live POS/stock data into specific order quantities, automatically creating purchase orders for the DC. sits on top of that forecast. For each SKU it compares predicted demand against current on-hand stock, stock already on order, supplier lead times, and a safety-stock buffer. When projected stock is due to fall below the safety level inside the replenishment window, the engine raises an order line — sized to the case-pack or pallet multiple — and bundles many lines into a single store order. The result is “just-in-time” ordering rather than “just-in-case” ordering: stores hold less buffer stock because the system is trusted to top them up before they run dry.

Where this goes wrong is usually a data problem, not a maths one. If the forecast is fed stale POS data, wrong promotional flags, or a broken master-data record (a wrong case-pack, a discontinued code still listed), the engine will confidently order the wrong quantity. That is why master data managementMaster Data Management (MDM): the discipline of keeping shared core data — product dimensions, case-packs, barcodes, supplier records — accurate and consistent across every system that uses it. and clean promotional data are prerequisites, not nice-to-haves. A store that overrides auto-replenishment constantly is usually signalling a data-quality issue upstream, and the fix is rarely more manual ordering — it is fixing the forecast inputs.

Order & planogram flow (text version)

Supplier  →  DC (inbound)  →  WMS / WCS  →  Cross-dock / Replenishment  →  Store order  →  Store receiving

Which protocol when? (EDI vs SFTP vs API)
  • EDI: the retail default for high-volume, standardised messages (orders, ASNs, invoices). Reliable, batched, well-understood.
  • SFTP: cheap, encrypted batch file drop — common for smaller suppliers and nightly inventory extracts.
  • API: real-time, event-driven — used for live stock availability, on-demand (dark store) orders, and dynamic pricing.

Most groups run all three: EDI for the core, SFTP for the long tail of suppliers, and APIs where speed matters.

Operations Play — Chapter 7

  • Confirm whether your store orders are auto-replenished or manual — manual lines are where stockouts hide.
  • Make sure you receive an ASN for every delivery; it halves receiving time and improves accuracy.
  • Keep your planogram version current with HQ — stale fixtures cause phantom out-of-stocks and wrong facings.
Chapter 08 · Suppliers

Supplier Engagement

Two delivery models compared. DSD (red) keeps freshness and supplier control for a few categories; DC delivery (blue) wins on volume, cost and receiving simplicity. The split is a commercial choice made per category. DSD vs DC delivery models Direct Store Delivery (DSD) Store Via Distribution Centre DC Store
DSD keeps supplier control and freshness for a few categories; the DC model wins on volume and cost.

Who pays, who delivers: suppliers fund DCs through slotting fees, delivery-window premiums, co-funding and co-packing. Goods arrive either direct-to-store (DSD) or via the DC. The split is a commercial decision per category, not a technical one.

How suppliers support the DC

MechanismWhat it isWho benefits
Slotting feesOne-off charge to list a new SKU / secure shelf & pick-face spaceDC (capex recovery)
Delivery-window premiumsPriority or off-peak appointment slots at receivingDC (throughput smoothing)
Co-fundingSupplier contributes to equipment, racking or automation tied to their volumeBoth — payback pulled forward
Co-packing / value-addDC performs repacking, multipacks, labelling, kitting on supplier’s behalfSupplier (saves capex) + DC (revenue)
Promotional co-opShared funding of promo volume handlingBoth

DSD vs DC delivery — the trade-off

DimensionDirect Store DeliveryDC delivery
Best forFresh, short-life, local, promotional (bread, milk, soft drinks)Volume ambient, frozen, slow-movers, national brands
Supplier controlHigh — owns the shelf & merchandisingLower — DC ranges and sequences
Cost to storeHidden in list price; heavy on receivingTransparent fee; low receiving load
FreshnessExcellent for same-day categoriesGood, but adds a handling step
Barrier for small supplierLow — can deliver one storeHigh — must meet volume & appointment discipline

Common supplier SLAs (illustrative)

SLA elementIllustrative target
On-time delivery (to appointment)≥ 96%
In-full (no short ships)≥ 98%
OTIF combined≥ 95%
Lead time (order → delivery)24–72 h by category
Damage / rejected-on-receipt≤ 0.5%
Chargebacks for non-compliancePer incident, escalating
The small-supplier squeeze. Centralisation favours suppliers who can deliver full truckloads to one DC on schedule. Small regional suppliers — the kind independents used to rely on — struggle to meet volume and frequency. The result is a narrower range and weaker access for unique local products20. Independents should protect a curated set of direct local relationships precisely because the DC will not.

Operations Play — Chapter 8

  • Negotiate OTIF SLAs with measurable targets and chargeback teeth — soft SLAs are free failures.
  • Keep a short, defended list of DSD partners for fresh/local lines the DC will not range.
  • Use slotting/co-pack conversations to co-fund automation rather than just paying listing fees.
Chapter 09 · Facility Types

Distribution Centre Types

Two temperature regimes side by side. Chilled holds 0–5 °C; frozen sits at −18 °C or colder. Energy, capex and handling discipline all rise sharply as temperature falls — a frozen chamber can cost two to three times as much per square metre to build. Chilled temperature-controlled cold-room storage with insulated panels and pallet racking CHILLED 0–5°C FROZEN ≤ −18°C
Temperature defines everything: infrastructure, cost, and handling discipline.

Not all DCs are equal: a dry-goods shed and a fresh-meat processing centre share a name but almost nothing else. Each type has distinct infrastructure needs (HACCPHazard Analysis and Critical Control Points: a food-safety management system identifying and controlling biological, chemical and physical hazards across the chain., GMPGood Manufacturing Practice: minimum standards for hygiene, process and documentation in food/production facilities., plant access, specialised racking), cost drivers and throughput profiles.

DC types — operational notes (illustrative ranges).
DC typeTempKey infrastructureCost driverThroughput note
Dry goods / general merchandiseAmbientHigh-bay racking, wide aisles, MHEFloor space, labourHighest case volume; lowest unit cost
Ambient groceries15–25 °CRacking, voice pick, cross-dock baysLabour, transportBackbone of any grocery DC network
Frozen / cold chain≤ −18 °CInsulated panels, dock seals, blast freezing, backup powerEnergy, capex2–3× ambient capex per sqm
Chilled / multi-temp0–5 °CRefrigerated chambers, ante-rooms, temp locksEnergy, speedDairy, meat, deli, prepared meals
Fresh produceVariedRipening rooms, grading, HACCP, wash-down areasShrink/spoilageFast flow-through; high waste %
Fresh meat / poultry processing≤ 4 °CProcessing plant, GMP, cold chain, sanitationPlant & complianceValue-add: cut, pack, date-code
Fish / seafood cold chain−1 to 2 °CIce/brine, super-chilling, ultra-tight temp controlCompliance, shrinkShortest shelf life; premium risk
Co-packing / private-label centreMixedPacking lines, labelling, kitting, QCLabour, line capexMargin engine: house-brand value-add

Handling notes for fresh & perishable DCs

The further a DC sits toward fresh and perishable, the more its economics are governed by shrink (lost stock to spoilage) rather than by throughput alone. Fresh produce centres add ripening rooms (bananas, avocados), grading and quality lines, and wash-down hygiene zones; their waste percentage runs far higher than ambient, and a single delayed truck can turn a day’s perishable intake into write-off. Meat and poultry processing centres are part warehouse, part factory: they receive carcasses or primal cuts, break them down, portion, tray-pack, gas-flush or vacuum-seal, apply date codes, and dispatch under continuous chilling — all under GMP and HACCP controls with sanitation cycles built into the shift. Fish and seafood cold chains are the tightest of all, often relying on ice or brine and super-chilling to stretch a shelf life measured in hours, not days.

The commercial implication: perishable DCs cannot be run as pure cost centres the way ambient sheds can. Their value is measured in freshness delivered and spoilage avoided, and they are typically located as close to the demand cluster as possible to shrink transit time. For a store owner, this is why your fresh delivery’s arrival temperature and remaining shelf life matter more than the invoice price — a cheaper pallet that loses a day of life costs more in the end.

What changes between an ambient and a frozen DC? (expand)
  • Capex: insulated envelope, refrigeration plant, dock seals and backup generation push frozen capex to multiples of ambient.
  • Energy: refrigeration can be 50–70% of a cold-site’s electricity bill; every degree of warming avoided is money saved.
  • Discipline: dock doors must seal instantly, doors stay closed, and product moves fast — a single propped door can compromise a chamber.
  • Compliance: cold and fresh sites carry HACCP/GMP audit load ambient sites do not.

Operations Play — Chapter 9

  • Match the DC type to the category’s shelf life — over-chilling ambient goods wastes energy; under-chilling fresh risks lives.
  • For any cold/fresh site, prioritise backup power and door-seal discipline before automation — they prevent catastrophic loss.
  • Use a co-pack/private-label line to lift margin on house brands rather than adding pure handling volume.
Chapter 10 · Technology

Technology Stack & Integrations

Specialist systems orbit a central integration layer. Each node owns one job — WMS, TMS, ERP, forecasting — and they only work well when clean data flows freely between them. A weak integration layer is the most common cause of phantom stockouts. Integration hub connecting all systems Integration WMS TMS ERP Forecast Planogram EDI/API Voice
A central integration layer ties specialised systems together. Brands vary; capabilities do not.

The stack: a modern DC runs on a family of specialised systems — WMS, WCS, TMS, ERP, demand forecasting, planogram tools, replenishment engines, and EDI/API gateways — joined by an integration layer. Below are capabilities only (no vendor names); evaluate tools against these, not against logos.

SystemCapabilityWhat good looks like
WMS (Warehouse Management)Receiving, put-away, inventory, picking, dispatchReal-time stock accuracy > 99%; slot optimisation
WCS (Warehouse Control)Orchestrates conveyors, sorters, automated equipmentSmooth hand-offs between humans & machines
TMS (Transport Management)Route planning, load building, delivery schedulingOptimised multi-drop routes; live ETA
ERPFinance, procurement, master data, the system of recordSingle source of truth for cost & inventory value
Demand forecastingPredicts store demand from sales, promo, seasonalityReduces both stockouts and overstock
Replenishment engineAuto-generates store orders from forecasts & POSOrders arrive “just in time”, not “just in case”
Planogram toolsDesign & distribute shelf layouts to storesConsistent, data-driven range & facings
EDI / API gatewayExchanges orders, ASNs, invoices with partnersReal-time where it matters; batch where it doesn’t
Voice picking / automationHands-free pick instructions; AGVs, conveyorsFaster, more accurate picks; quieter floor
How much automation makes sense? (expand)

Automation’s payback depends on the ratio of wage cost to capital cost. Where labour is relatively affordable, voice picking, route optimisation and conveyor sortation typically pay back fastest. Full robotic automation (AS/RS, shuttle systems, goods-to-person robots) is harder to justify until volumes are very high or labour markets tighten. The pragmatic path in most SA networks: automate the repetitive, error-prone steps first; keep humans on judgement and exception handling26.

Operations Play — Chapter 10

  • Specify requirements by capability, then evaluate vendors — never start with a brand shortlist.
  • Insist on open APIs/EDI; a closed system is a future cost trap.
  • Automate the boring, error-prone tasks (picking accuracy, routing) before chasing headline robotics.
Chapter 11 · Risk

Pain Points & Risk Mitigation

Four forces stress every South African DC: power loss, cargo crime, system outage and cold-chain failure. Mitigation here is operational, not theoretical — every one of these has a concrete, tested countermeasure. Risk shield ! Power 🚚Crime 💻Outage ❄️Cold chain
Four forces stress every SA DC. Mitigation is operational, not theoretical.

What breaks: loadshedding and the cold chain, cargo crime, supplier consolidation, and rigid cut-off times are the live SA pain points. Below is a risk-and-mitigation matrix plus the SA-specific numbers that make them urgent.

SA-specific risk signals (illustrative, to be verified)

  • Loadshedding cost: one major retailer reported a ~R200m spoilage loss in a single year14; another’s monthly generator diesel spend climbed from ~R16m to ~R150m at peak15. Cold-chain operators added 15–30% to operating cost and apply 5–15% energy surcharges16.
  • Cargo crime: SAPS recorded 420 truck hijackings in Q2 2025 alone17; over 18 months to mid-2022, ~2,670 incidents and ~R577m in losses were logged across all provinces17. The logistics “security tax” rose from ~0.3% to ~2.3% of operating cost (up to ~4.1% for high-value goods)19.
  • Cold-chain physics: roughly +1 °C per ~10 minutes outside the chain13; chilled > 5 °C or frozen > −18 °C = compromised.

Risk & mitigation matrix

RiskLikelihoodImpactMitigation
Cold-chain failure / power lossHighSevereRedundant refrigeration, generator + solar/battery, temp alarms, IoT monitoring, door-seal discipline
Cargo theft / hijackingHighSevereRoute variation, armed escorts/convoy, real-time tracking, anti-jamming, vetted drivers, no fixed schedule leaks
LoadsheddingHighHighBackup power sized for full load, solar + storage, demand-side management, schedule-aligned ops
Labour strike / unrestMedSevereCross-trained staff, contingency roster, safety stock buffer, secure site access
System outage (WMS/ERP down)Low–MedSevereFailover/HA, tested DR, paper-based fallback procedures, offline picking mode
Stock obsolescence / expiryMedMedFEFO enforcement, demand sensing, markdown engine, tight promo planning
Transport disruption (breakdown, port, road)MedHighSpare fleet capacity, multiple carriers, buffer stock at distant stores, contingency routing
The dependency trap. In the old DSD world, if one supplier failed, the others still came. In a centralised model, a single DC fire, systems crash, strike or major loadshedding event can darken every store in a region at once. Continuity planning is not optional — it is the price of scale.

Operations Play — Chapter 11

  • Test backup power under full load this month, not next — and align shifts to the loadshedding schedule.
  • Keep a vetted local fallback supplier for bread, milk, eggs in case a DC truck is hijacked or delayed.
  • Log fridge/freezer temps every 2 hours — that log is your defence if a customer gets sick.
  • Run a tabletop drill: “WMS is down for 6 hours” — can your floor still pick and dispatch on paper?
Chapter 12 · Future

Dark Stores & the Next Decade

A dark store has no walk-in customers. Staff pick and pack online orders, and a rider dispatches within the hour. This is the fastest-growing node in the network and removes the conflict between online pickers and shoppers fighting for the same shelf. On-demand dark-store fulfilment Dark store (no walk-ins) pick → pack → dispatch Customer 60-minute grocery, on demand
A dark store is a shop customers never enter — it exists only to pick and pack online orders.

The next layer: on-demand grocery is the fastest-growing part of the logistics picture. A dark storeA retail-format facility closed to walk-in shoppers, optimised purely for picking and packing online orders for rapid delivery. removes the conflict between online pickers and in-store shoppers fighting for the same shelf. Expect more micro-fulfilment, automation and township reach over the next decade.

Where the numbers are going (generic, SA)

  • SA online retail reached ~R71bn in 2023 and is on track to exceed ~R130bn by end-2025 — approaching ~10% of total retail25. On-demand grocery is the fastest segment within that.
  • One large discounter’s sub-hour service now runs multiple dedicated dark stores and fulfils from hundreds of regular stores turned into micro-fulfilment nodes; that one service alone grew revenue by ~35% in a recent half-year and now represents roughly a tenth of group sales21.
  • A premium grocer’s on-demand app grew ~70% year-on-year and is adding more dark stores2223; a value retailer’s online business grew nearly 50% and is now profitable on a fully-costed basis24.
  • Estimates put an untapped township economy on the order of ~R900bn25 — both a threat (more competition) and an opportunity (better supply access than ever).

What this means inside the DC and on your floor

  • Automation creeps in: conveyor sortation and automated guided vehicles move totes; IoT sensors flag the moment a truck’s temperature drifts26.
  • Labour still dominates: because local wages are lower than in Europe/US, full automation is hard to justify — but voice picking, route optimisation and WMS are already standard.
  • Your store may become a node: staff could be asked to pick online orders from your shelves while serving walk-in customers — plan dedicated pickers and clear processes.
  • Resilience goes green: solar + battery backup at DCs to cut diesel dependence is the coming default.
The township & rural play. Discount and value formats are still building DCs for low-income and rural markets. If you operate there, the expanding DC network is both competition and the best supply access you have ever had. The winners will combine chain efficiency with local relationships the chains cannot replicate.

Operations Play — Chapter 12

  • If your group launches online fulfilment from your store, secure dedicated pick staff and a staging area before go-live.
  • Identify whether a dark store is planned in your catchment — it changes your online opportunity and your walk-in risk.
  • For independents in townships: build the local-list relationships now, before chain on-demand arrives.
Chapter 13 · Action

Practical Playbooks by Role

Expand the panel for your role. Each is a short, actionable checklist you can act on this week.

For franchisee supermarket owners
  • Your franchise agreement likely forces most stock through the franchisor’s DC — know your receiving windows and build the roster around them.
  • Invest in backup power; if the DC truck arrives during loadshedding and your cold room is down, you lose stock before you sell it.
  • Chase shortage/damage credits like a hawk — the DC handles thousands of cases; do not sign until you have checked.
  • Prepare for online fulfilment: dedicated pickers, clear processes, so floor staff aren’t fighting pickers for the last loaf.
For independent supermarket owners
  • If you’re not tied to a chain DC, consider a buying group / voluntary chain to get DC pricing without full franchise control.
  • Your edge is flexibility: stock the local products the big chains ignore — bakeries, butcheries, produce suppliers who can’t supply a mega-DC.
  • Watch transport cost: every kilometre to a wholesaler eats margin. Consolidate collection runs.
For store managers
  • Train receiving staff to catch short dates, damage and temperature breaches before they hit the shelf.
  • Keep the cold chain intact: log fridge/freezer temps every 2 hours. That log is your defence.
  • Plan for late deliveries: have a local fallback for bread, milk and eggs if the DC truck is hijacked or breaks down.
For area managers
  • Track DC service level by region; if one DC is consistently late or short, escalate it.
  • Share crime intelligence: if hijackings rise on a route, push for route changes, escorts or earlier delivery times.
  • Keep every store current on the loadshedding schedule; confirm generators were tested this week.
For cash-and-carry operators
  • Your customers collect, but your own receiving faces the same DC risks if you buy from chain DCs or importers.
  • Decide whether on-demand/dark-store fulfilment is a threat or an opportunity — could you serve spaza shops before the chains do?
For family-owned supermarket businesses
  • The next generation must understand logistics, not just the shop floor. Your edge is now knowing your supply chain by the numbers.
  • If selling or franchising, business value is tied to DC integration — clean receiving records, low wastage and good DC relationships add real value.

Appendix

Glossary, References & Developer Notes

Glossary

TermMeaning
DCDistribution Centre — central warehouse that receives bulk and redistributes mixed, store-ready loads.
DSDDirect Store Delivery — supplier delivers straight to each store.
Cross-dock / flow-throughGoods move inbound-to-outbound with little/no storage.
WMS / WCSWarehouse Management / Control System — runs and orchestrates DC operations.
TMSTransport Management System — route planning, load building, scheduling.
ERPEnterprise Resource Planning — finance, procurement, master data, system of record.
EDI / SFTP / APIB2B message exchange methods (standard messages / encrypted file drop / real-time calls).
ASNAdvance Ship Notice — tells the receiver what’s on the truck before it arrives.
POSPoint of Sale — the till; sell-through data feeds replenishment.
MHEMaterial Handling Equipment — forklifts, reach trucks, conveyors, AGVs.
FEFO / FIFOFirst-Expired / First-In-First-Out stock rotation.
HACCP / GMPFood-safety management / Good Manufacturing Practice standards.
SLA / OTIFService-Level Agreement / On-Time-In-Full delivery metric.
Slotting feeCharge to list a SKU / secure pick-face space.
EBITDA / Capex / Opex / ROIEarnings before interest, tax, D&A / capital expenditure / operating expense / return on investment.
Dark store / micro-fulfilmentShopper-free pick facility / small-format fulfilment node.
SKUStock-Keeping Unit — a unique sellable product line.

References

Every citation in this playbook is a live link to its source. Hover a number in the text to see the source; click to open it.

#Source
1Leader.co.za — retailer history: why a late move to central distribution was costly
4IOL Business Report — a major retailer phases in centralised distribution (2007)
5Supply Chain Nuggets — flagship inland DC: build cost, footprint, stores served
6Safal Steel — largest single-roof distribution centre on the continent
7News24 Fin24 — inside a mega-DC: consolidation, sections, employment scale
8Supermarket.co.za — DC operations: dock doors, product/supplier counts, voice picking
10The Spar Group — regional grocery DC network & satellite hubs (PDF)
12WBHO — dock-leveler count at a large inland DC
13Shipping & Freight Resource — cold-chain temperature rise outside the chain
14EcoFlow — retailer spoilage loss attributed to loadshedding
15M&G Investments — generator diesel spend trend; generator vs grid cost multiple
16Cold Chain SA — cold-chain operating-cost increase & energy surcharges
17Logistics Business Africa — truck hijacking & cargo-theft statistics (SAPS / industry)
19Channelwise — the logistics ‘security tax’ trend over time
20Wiley Online Library — small-supplier disadvantage & distance from a DC
21MyBroadband — sub-hour on-demand service: dark stores, growth, % of group sales
22Business Day — premium grocer dark-store expansion
23Centro Property Solutions — premium grocer on-demand app growth
24Daily Investor — value retailer online growth & profitability
25OpenPR / IMARC — SA online-retail size/growth; untapped township economy
26ITWeb — automation & labour in smart warehouses
Scroll to Top