Warehouse logistics that work

Build 3674 scope and the logistics model

This guide describes the warehouse distribution system in Big Ambitions 1.0 build 3674, after the EA 0.9 weekly-import redesign and the EA 0.10 factory separation.

A warehouse is the distribution layer between purchased or manufactured inventory and retail businesses. In the current design, importers send goods into a warehouse, factories send finished goods into a warehouse, and a warehouse sends goods onward to stores. A factory is no longer a substitute retail-distribution warehouse: the developer explicitly confirmed in EA 0.10 that factories manufacture and may deliver to warehouses or export to importers, while warehouses distribute to stores. That distinction determines which destination options appear and prevents an obsolete direct factory-to-store design from working.

Build 3674 matters because its hotfix repaired a quit-state defect that could block imports, deliveries, training, and other systems. If a save was opened on build 3672 and every apparently valid logistics plan stopped together, advance into the next game day after updating before rebuilding the network. The hotfix note says affected systems should recover as play continues. A single missing route is still more likely to be configuration or capacity; simultaneous failure across unrelated routes is the stronger signature of the repaired defect.

The complete dependency chain

A route exists only when every physical, employee, management, inventory, and destination dependency is valid at the calculation time.

The minimum import-to-store chain is importer contract to Purchasing Agent plan to receiving warehouse pallet capacity to warehouse vehicle bay to compatible vehicle to assigned Delivery Driver to Headquarters Logistics Manager to a non-zero destination target to destination storage or display capacity. The Purchasing Agent controls inbound purchasing; the Logistics Manager controls outbound target replenishment. Confusing those jobs is a common diagnostic dead end: a driver does not collect the importer order, and a Purchasing Agent does not choose store destinations.

For factory goods, prepend raw-material procurement and production, then preserve the warehouse stage: raw-goods importer to Purchasing Agent to factory storage to configured production line and workers to factory finished inventory to factory Logistics Manager, driver, and vehicle to distribution warehouse, followed by that warehouse's separate Logistics Manager, driver, vehicle, and store routes. Each managed factory or warehouse needs its own applicable Logistics Manager plan. Treat each arrow as a measurable handoff rather than assuming the whole chain is one automation.

Prerequisites before the first automated delivery

Build the receiving and management infrastructure before committing a large weekly order.

Register the industrial building with the correct warehouse business type, then install pallet shelves. Imports directed at an empty industrial shell do not have usable storage. Hire a Delivery Driver, acquire a delivery-capable vehicle, drive or have it delivered into a warehouse bay, and assign the driver in BizMan. Create a Headquarters with a desk, chair, and computer workstation, hire a Logistics Manager, assign that employee to HQ, schedule the workstation, and create a plan for the warehouse. The historical developer explanation remains mechanically useful: the vehicle and driver enable the Logistics Manager's delivery plan; they are not a manually visible daily trip simulation.

Do a small commissioning order or manually place known inventory in the warehouse before scaling. Add one store as a destination, select one unambiguous item, and set a target large enough to create a visible shortfall. Record warehouse and store inventory shortly before and after 02:00. This isolates outbound logistics from importer timing. Only after that test succeeds should you add many products, a second vehicle bay, factory feeds, or multiple destinations.

The 02:00 target-replenishment calculation

Warehouse routes are automatic target fills, not fixed shipments and not a forecast of demand.

For an item, let T be the destination target saved in the Logistics Manager plan, D be total destination inventory immediately before calculation, S be usable source inventory, and C be free receiving capacity expressed in units of that item. The intended request is R = max(0, T - D). The delivered quantity is bounded by the chain: Q = min(R, S, C), subject to a valid route, driver, and vehicle. The developer's example is T = 500 and D = 200, producing R = 300. Total destination inventory includes goods on display as well as back-room storage, so counting only storage shelves will overstate the expected shipment.

Internal logistics is evaluated daily at 02:00 in the current 1.0-era developer reply. This supersedes old replies that said midnight. A target of 1,000 does not mean send 1,000 every night; it means attempt to leave the destination holding 1,000. If the store has 820, the request is 180. If it has 1,050, the request is zero. The target mechanism is already smart replenishment, but the game does not choose T for you. Demand estimation and safety stock remain player decisions.

Monday sequencing and the six-hour gap

Monday has two distinct supply events, and their order can create a false contradiction in notifications.

Warehouse and factory logistics runs at 02:00 each day. Standard importer deliveries arrive at 08:00 Monday unless an urgent order is used. Therefore Monday's 02:00 store replenishment uses warehouse inventory left from the previous week; it cannot use the inbound shipment arriving six hours later. A low-inventory message timestamped 02:00 may be correct even if the warehouse looks full after 08:00. The newly imported stock normally becomes available to the next internal run at 02:00 Tuesday.

Design safety stock around that ordering. Let d be expected daily warehouse outflow and b be additional uncertainty stock. A minimal Monday bridge is B = d + b after Sunday's retail replenishment, because the source must survive Monday 02:00 before the new shipment. In practice, use recent item-level outflow rather than store revenue. If a product averages 700 units per day and the selected buffer is 25 percent of one day, B = 700 + 175 = 875 units that should remain available before Monday's run. This is a planning formula, not a hidden game formula.

Setting defensible store targets

Targets should cover the interval until the next internal run plus uncertainty, not an arbitrary number copied across products.

For each store and item, track actual units sold on complete, in-stock days. Let A be the average daily units sold, P be the maximum observed complete-day units, L be one day between logistics runs, and z be a chosen safety multiplier. A simple target is T = ceil(max(A, P) times L times (1 + z)). If average sales are 420, the observed peak is 510, and z is 0.20, T = ceil(510 times 1.20) = 612 units. Round upward to a practical amount that fits the store. Do not use a shortage day as evidence of true low demand because lost sales are absent from the sales count.

Opening hours, promotion, neighborhood demand, pricing, competitors, and display capacity can change sales. Re-estimate after any material change and use at least one full week without stockouts where possible. If storage is tight, lower targets on slow products before starving fast products. A target that cannot physically fit causes an incomplete delivery, while a target below daily sales creates a late-day stockout even though the route worked perfectly at 02:00.

Warehouse stock and weekly inbound planning

A store target is daily; the warehouse's inbound target must cover all destinations over the weekly importer interval.

Let i index stores, A_i be expected daily sales of the item at store i, H be seven days between regular imports, F be factory or other known inbound units during the week, and B be safety stock. Expected warehouse need is W = ceil(H times sum(A_i) - F + B). For three stores consuming 220, 310, and 470 units daily, weekly demand is 7,000. If a factory reliably supplies 1,400 and the chosen buffer is 1,000, W is 6,600 units. When Smart Delivery is enabled on a recurring importer order, use W as the target inventory concept rather than treating it as a guaranteed shipment quantity.

The formula assumes all stores are served by this warehouse and that factory output is reliable. Split networks require demand allocation by source. Also account for the Monday sequence: W should not be allowed to decay to zero before Monday 02:00. Weekly importer limits, cash, and physical storage can bind below the calculated need. When any constraint binds, the answer is not to inflate store targets; add inbound capacity, another eligible destination order, factory output, storage, or a deliberate urgent order.

Vehicles, drivers, and route-slot capacity

For automated logistics, vehicle class primarily controls destination slots; manual cargo capacity is a separate statistic.

The developer has stated that automated warehouse logistics is not limited by the truck's manual box capacity; it is limited by destination count, source inventory, and destination space. Vehicle class supplies a base destination count, and Logistics Manager skill increases the available count up to the live maximum. A warehouse with two active bays can use two vehicles and drivers while still using one Logistics Manager for that warehouse. The current community reference table reports Freight Truck T1 at four standard or eight with a 100-percent manager, Vord Courier D500 at three or six, and Mersaidi Dash, UMC Nunavut, or Vord V150 at two or four. Treat those values as reference data to verify in F1/BizMan, not timeless constants.

Driver skill gates vehicle assignment. Current reference data reports 50 percent for the Dash-class group, 75 percent for the D500, and 95 percent for Freight Truck T1; older developer replies mention lower truck thresholds and are stale for this number. Do not infer that a 95-percent driver moves more goods than another qualified driver. The demonstrated purpose of the gate is eligibility to operate the vehicle, while Logistics Manager skill and vehicle class determine destinations.

Storage capacity is part of route capacity

A valid plan cannot force inventory into a destination that has no compatible free storage.

Warehouse pallet shelves receive inbound units in boxes internally, while the order and target interfaces use individual units after the EA 0.9 redesign. Store capacity includes storage shelves and relevant display furniture. Because products have different units per box and furniture capacities, a universal units-per-shelf claim is unsafe. Inspect the current product and furniture tooltips. The automated vehicle may have unlimited shipment volume for logistics purposes, but the receiving building is never unlimited.

Diagnose capacity by separating source shortage from destination fullness. The developer explicitly distinguished the messages: not enough inventory means the source lacks goods; not enough space means the receiver cannot accept them. If the store visually appears to have an empty back-room shelf, check whether displays or other products consume the remaining compatible capacity and whether the target itself exceeds all available space. At the warehouse, allow headroom for the Monday import peak before the first post-import store run.

Scaling gates and network topology

Scale when a measured constraint is reached, not merely because the current warehouse looks busy.

The first gate is usually route destinations: a new store cannot be added to the manager plan. Upgrade the Logistics Manager, choose a vehicle with more destination capacity, or activate another bay if the building has one. The next gate is receiving or holding capacity: weekly stock cannot fit with safety headroom. Add pallet shelves, move to a larger industrial property, or split product families across warehouses. A third gate is the per-location weekly import limit; another eligible receiving location or factory production may be required. A fourth gate is operational risk: one warehouse outage affects too much revenue, justifying regional or category separation before a hard limit.

A central warehouse minimizes duplicated buffer and contracts but concentrates risk and route slots. Category warehouses simplify purchasing and storage analysis but can send multiple routes to the same store. Regional warehouses shorten the mental map but in-game deliveries are calculated rather than simulated road journeys, so geography is primarily an organizational choice unless a live mechanic says otherwise. Factory-to-distribution-warehouse specialization follows the mandatory 0.10 boundary and lets stores receive both manufactured and imported goods from one retail distribution layer.

A reproducible commissioning test

Use a controlled one-item test to prove each handoff before trusting a large network.

Choose a product already valid for both source and destination. At 01:50 on a non-Monday day, record source inventory S0, destination total inventory D0, target T, free destination capacity C, assigned manager, driver, vehicle, and build number. Predict Q with min(max(0,T-D0),S0,C). Advance past 02:00 without changing the plan. Record S1 and D1. A clean pass has approximately S0-S1 = D1-D0 = Q, allowing for any simultaneous retail consumption only if the store is open during the observation window. Testing a closed store removes that ambiguity.

If the result differs, repeat with screenshots and one changed variable at a time. Confirm the destination is the one attached to the tested manager, target is saved and non-zero, source contains the exact SKU, and receiver has compatible capacity. Reassigning or deleting everything at once destroys evidence. For a suspected build defect, reproduce after a save/reload and next-day transition, then submit the save through F2 with manager, source, destination, expected quantity, and observed quantity, following the developer's preferred bug-report detail.

Diagnostic decision tree

Start from the exact outcome and walk upstream; do not begin by replacing managers or vehicles.

If no shipment occurs, first ask whether D was already at or above T. If yes, zero is correct. If no, verify the alert time and plan value, then inspect source inventory at 02:00. If source is insufficient, trace inbound purchasing or factory output. If source is sufficient, inspect destination capacity. If capacity exists, verify route-slot availability, the warehouse selected in the Logistics Manager plan, vehicle assignment, driver eligibility, and manager assignment. A plan can look populated in an obsolete HQ after moving offices, so inspect all HQ plans for duplicates or empty remnants.

If every route fails simultaneously, confirm build 3674 and advance one day because the hotfix repaired a global failure triggered by bugged quitting employees. If one SKU fails while others pass, suspect item stock, target, compatibility, or capacity rather than vehicle infrastructure. If one destination fails for all products, suspect destination capacity or route configuration. If failure happens only Monday at 02:00 and disappears Tuesday, suspect the six-hour import timing gap rather than a broken route.

Operational tracking fields

A small weekly ledger makes logistics explainable and prevents target inflation from becoming the default response.

Track one row per source-destination-item combination: build, week, source, destination, SKU, route target, destination stock immediately before 02:00, delivered units, source stock after delivery, complete-day sales average, peak sales, stockout hours, receiving capacity, warning text, and warning timestamp. For importer-fed warehouses add Sunday plan status, smart toggle, recurring toggle, weekly target, expected charge, actual Monday receipt, per-location limit, and Monday 01:59 bridge stock.

Use the ledger to distinguish demand errors from flow errors. If deliveries equal predicted shortfalls but stores still run out, targets are too low or demand changed. If targets fit but Q is below prediction, a physical or plan constraint is binding. If weekly receipts are short while internal routes work, investigate purchasing rather than logistics. Retain at least two comparable weeks after opening or repricing; a single promotional or shortage week is not a stable baseline.

Safe warehouse migration and plan change control

Moving a live distribution hub is a multi-event cutover, not a furniture exercise, because inventory, vehicles, drivers, managers, and weekly orders have separate bindings.

Commission the replacement warehouse in parallel. Register its business type, install pallet capacity, assign vehicles and drivers, create the new Logistics Manager plan, and prove a one-item 02:00 route before redirecting the weekly purchase. Keep the old warehouse funded and stocked through at least one import-to-store cycle. If the move crosses Sunday 20:00, record which Purchasing Agent plan owns the upcoming order; an empty plan left in an old HQ or a manager removed before lock can cancel Monday supply even when the employee is scheduled elsewhere.

Freeze discretionary target changes during the cutover so inventory movement can be explained. Record old and new source inventory by SKU, every destination assignment, vehicle bay, driver, route target, purchasing destination, recurring state, and smart state. Redirect one product family or store group at a time. After the new route passes, reduce the old route target to zero or remove it deliberately, then confirm that no destination is served twice and no item is stranded. Do not terminate the old building until imported goods, manually placed boxes, and active management plans have been reconciled.

A rollback point should specify the last known working save, the old plan ownership, and the inventory needed to survive the next Monday gap. If the new warehouse fails, restoring the route without restoring inbound purchasing may only postpone the shortage. If movers were used and vehicles appear unassigned, inspect each bay; EA 0.10 patch notes specifically fixed warehouse vehicles not being unassigned after movers, showing that vehicle binding has historically been migration-sensitive. Build 3672 also fixed HQs opening no days after movers, so verify the active HQ schedule rather than assuming office furniture transfer preserved execution.

Version notes and claims deliberately excluded

Several widely repeated logistics claims are obsolete, ambiguous, or unsupported for build 3674.

Do not state that warehouse delivery occurs at midnight: a May 2026 developer reply specifies 02:00, while imports arrive Monday at 08:00. Do not state that the factory may distribute directly to stores: the EA 0.10 separation intentionally requires factories to feed warehouses or exporters. Do not use manual cargo capacity as an automated nightly throughput cap: the developer explicitly described no automated quantity limit per truck, only destination, stock, and receiving constraints. Do not state one universal driver threshold for freight trucks; current reference data reports 95 percent and older replies report lower values.

Exact product box sizes, importer limits, rents, prices, vehicle prices, and warehouse layouts are omitted because they vary by product, difficulty, balance patch, or building. The current UI and F1 Help are stronger than an unversioned list. This guide also avoids claims that warehouse geography changes automated delivery time, that shelf sorting increases throughput, or that Logistics Manager work hours change the fixed calculation event; no accepted current source in this audit established those as universal 1.0 mechanics.