Multi-Warehouse Fulfilment Is Breaking Your Google Ads Bidding

The Complexity Nobody Models

Single-warehouse businesses have one delivery cost per order. Simple. Multi-warehouse businesses have a matrix: customer location × warehouse location × product weight × carrier rate × speed tier. A £45 order from Manchester fulfilled from a Midlands warehouse costs £3.20 to ship. The same order fulfilled from Kent costs £5.80. From Scotland, £6.50.

These differences compound. On a £45 order with 40% gross margin (£18), the fulfilment cost variance of £3.30 represents an 18% swing in contribution margin. Smart Bidding sets one bid for both scenarios because it has no visibility into which warehouse will fulfil.

For brands running three or more warehouses, this margin blindness is systematic. Every bid is wrong - the only question is which direction and by how much.

The Delivery Promise Gap

Multi-warehouse fulfilment exists to improve delivery speed. Customers near a warehouse get next-day delivery. Customers far from any warehouse get 2-3 day delivery. This delivery speed variance directly impacts conversion rate - but Smart Bidding treats all regions equally.

A customer in Birmingham (near your Midlands warehouse) sees "next-day delivery" and converts at 4.2%. A customer in rural Devon sees "2-3 business days" and converts at 2.8%. Smart Bidding doesn't distinguish between these because it can't see your warehouse map. It bids the same for both, despite the Devon customer being 33% less likely to convert.

The irony: you invested in multi-warehouse infrastructure to improve the customer experience, but your advertising doesn't leverage the improvement. The conversion rate benefit of fast delivery exists - your bidding just doesn't capture it.

Regional Bid Distortion

The maths of regional distortion are brutal. Consider a brand with warehouses in Manchester and London:

• • London customer, London fulfilment: £3.50 delivery cost, next-day, 4.5% CVR → £14.50 contribution margin
• • Scotland customer, Manchester fulfilment: £5.20 delivery cost, 2-day, 3.1% CVR → £12.80 contribution margin
• • Cornwall customer, London fulfilment: £6.80 delivery cost, 2-3 day, 2.6% CVR → £11.20 contribution margin

Smart Bidding calculates one blended target. If it optimises for the average (£12.83 margin, 3.4% CVR), it overbids for Cornwall customers by 14% and underbids for London customers by 13%. Over thousands of clicks per week, this compounds into significant waste.

The brands that solve this use geographic bid modifiers or regional campaign structures. But most don't because their agency runs a single national campaign and reports blended metrics that look fine on average.

Stock Fragmentation and Availability

Multi-warehouse means stock is split across locations. Product X might have 200 units in Manchester and 15 in London. A London customer orders it - if the London warehouse is out, it ships from Manchester at higher cost with slower delivery. The customer experience degrades, returns increase, and your margin calculation was wrong at the point of bidding.

Stock velocity compounds this. Fast-selling SKUs deplete unevenly across warehouses. Your Manchester warehouse runs out on Tuesday. London still has stock. From Tuesday to restock, all Northern orders ship from London - doubling delivery cost and adding a day to delivery time.

Google Ads keeps bidding as if nothing changed. Your feed still shows the product as "in stock." The CPC stays the same. But every sale from Tuesday to restock costs you an extra £2-4 in fulfilment. At 50 orders per day, that's £100-200 in invisible margin erosion - every day.

Smart Bidding's Structural Blind Spot

Smart Bidding optimises for conversion value. It knows the revenue of each sale. What it doesn't know:

• • Which warehouse will fulfil the order
• • What the delivery cost will be from that warehouse to that customer
• • Whether the order will trigger a split shipment (2x fulfilment cost)
• • Whether the nearest warehouse is out of stock, forcing a distant fulfilment
• • Whether the delivery promise shown to the customer matches reality

POAS bidding helps - but only if your profit signal accounts for actual fulfilment costs, not averages. Most POAS implementations use a flat fulfilment cost per order. That's better than nothing, but it's still wrong for multi-warehouse operations.

True accuracy requires feeding real fulfilment costs back into your conversion value calculations. This means connecting your order management system to your ad platform - which adds complexity but eliminates the structural blind spot.

Structural Fixes

Three approaches, from simplest to most sophisticated:

• • Regional campaigns: Split your account by geographic region aligned to warehouse coverage zones. Set different POAS targets per region based on average fulfilment cost from the primary warehouse. This captures 70-80% of the distortion.
• • Dynamic profit signals: Feed actual order-level fulfilment costs into your conversion value through server-side tracking. Each conversion reports its true margin after real fulfilment costs. Smart Bidding then optimises on accurate signals.
• • Stock-aware bid modifiers: Connect your inventory management system to bid adjustments. When a warehouse runs low on a SKU, reduce bids in that warehouse's coverage zone to avoid expensive cross-warehouse fulfilment. When restocked, bids return to normal.

Most brands should start with regional campaigns. The data infrastructure for dynamic profit signals takes 3-6 months to build properly. Stock-aware modifiers require real-time inventory feeds, which most OMS platforms can provide but few agencies know how to use.

Next Steps