What Are Push vs Pull Inventory Systems?
Push vs pull inventory systems describe two fundamentally different philosophies for deciding when and how much to stock. In a push system, inventory moves through the supply chain based on forecasts: you produce or order ahead of demand, position inventory at stocking locations, and accept that some forecasts will be wrong. In a pull system, inventory only moves in response to actual consumption: a customer order at the downstream node triggers replenishment from the upstream node, and so on back through the chain.
Toyota's lean manufacturing made pull famous through just-in-time (JIT) inventory and kanban systems. Pull works when lead times are short, demand signals are reliable, and supplier coordination is tight. None of those conditions apply to Amazon FBA. The 80-140 day total lead time from China factory to FBA-available unit makes pull-based replenishment physically impossible for the vast majority of FBA sellers.
Understanding push vs pull inventory systems matters for FBA sellers because it explains why your inventory strategy looks "wasteful" by lean-manufacturing standards. You're not running it badly; the structure of FBA forces push.
Push vs Pull: How They Compare
The differences come down to what triggers inventory movement, how much buffer you carry, and how the system responds to forecast error.
| Dimension | Push System | Pull System |
|---|---|---|
| Trigger | Forecasted demand | Actual consumption |
| Lead time tolerance | Long lead times OK | Requires short lead times |
| Safety stock | High (forecast buffer) | Low (signal-driven) |
| Forecast accuracy needed | Critical | Less critical |
| Risk | Excess and obsolete inventory | Stockouts during demand spikes |
| FBA fit | Required (most cases) | Possible only at AWD-to-FBA layer |
The math: a push system's safety stock must cover forecast error across the entire lead time. If your lead time is 90 days and your forecast accuracy is +/- 25%, your safety stock has to absorb 22.5 days of potential overshoot. A pull system reduces this to the time between demand signal and replenishment, which is days, not months.
Worked Example: Push vs Pull for an FBA Garden Tool
You sell a $39.99 garden tool peaking in spring. Annual volume: 12,000 units. Manufacturing in Vietnam: 60-day lead time. Ocean freight: 28 days. FBA receiving: 12 days. Total lead time: 100 days.
Push approach (what FBA forces): You forecast 4,000 units for the spring quarter. You place a PO in late November for arrival by mid-February, three weeks before the spring rush. You commit to 4,000 units of finished, packed, FBA-ready inventory based on a forecast made four months ahead. If actual demand is 3,000, you have 1,000 units of excess inventory and pay aged inventory surcharges by August.
Pull approach (theoretical, not feasible): Wait for actual orders, then trigger production. First spring order arrives March 1. Production starts March 1, ships May 1 (factory + freight + receiving). The unit is FBA-available June 9, three months after the customer wanted it. The order is canceled, the customer found a competitor, and your buy box is gone.
This is why FBA sellers run push: pull is mathematically impossible when total lead time exceeds the demand cycle. Your only escape valve is buffering inventory at AWD or a 3PL, where releases to FBA can be more demand-responsive (a quasi-pull layer downstream of the deep push layer).
Why FBA Forces Push vs Pull Inventory Systems
Three structural realities of FBA make push the only viable approach for upstream inventory:
Lead time stacking: Your supply chain has 4-5 sequential stages, each with its own lead time. By the time a unit goes from "place PO with factory" to "available on Amazon," 80-140 days have passed. Pull systems break down once total lead time exceeds the customer's tolerance for delivery, which is two days for Prime customers.
Demand-supply mismatch in geography: Most FBA inventory originates in Asia. US customer demand happens 7,000+ miles away. The transit time alone makes signal-driven replenishment impossible. By the time you "see" the demand signal in your sales data, the customer order has already been fulfilled from existing FBA stock.
FBA fee structure punishes responsiveness: Even if you could produce locally with 7-day lead times, FBA's restock limits and replenishment cycle mean you can't move inventory inbound on demand. Amazon's network is itself a push system from the seller's perspective.
Common Mistakes
1. Trying to run lean/JIT inventory in FBA. Sellers who read Toyota's playbook and try to apply it to FBA end up chronically stocked out. The principles don't translate when your lead time is 90+ days. Embrace the push reality and optimize for forecast accuracy instead.
2. Confusing reactive ordering with pull. Reordering when stock gets low is not a pull system. It's just a poorly executed push system without proper reorder points. True pull requires demand signals propagating through every stage instantly, which FBA cannot do.
3. Underinvesting in safety stock because "we'll just reorder." Push systems require meaningful safety stock buffers because every reorder commits inventory based on a forecast that will be wrong. Sellers who run lean to "save on holding costs" end up paying far more in stockout costs and lost buy box.
Related Terms
Push vs Pull Inventory FAQ
Is FBA a push or pull inventory system?
FBA is fundamentally a push system. You commit inventory to Amazon's fulfillment centers weeks before customer demand materializes, based on forecasts. The combination of long supplier lead times, ocean freight transit, and FBA receiving windows makes pull-based replenishment impossible for most FBA sellers.
Can FBA sellers ever use pull inventory?
Only at the AWD-to-FBA layer. AWD acts as a buffer between your push-stocked deep inventory and FBA's customer-facing nodes. Releases from AWD to FBA can be more demand-responsive (closer to pull), but the upstream supply chain remains push because of multi-month lead times.
Why don't FBA sellers run JIT (just-in-time) inventory?
JIT requires reliable, short lead times and tight supplier coordination. FBA sellers source from manufacturers with 45-90 day production lead times, plus 20-30 days ocean freight, plus 7-21 days FBA receiving. That's 80-140 days from PO to sellable, which is the opposite of just-in-time. Pull/JIT works for Toyota; it doesn't work for Amazon FBA.
What does push vs pull mean for safety stock?
Push systems require larger safety stock because you're committing inventory based on forecast error, not actual demand. Pull systems can run leaner because each unit is triggered by real consumption. Since FBA forces a push model, your safety stock buffer must be sized for forecast variability across the entire supplier-to-FBA lead time.