Make to stock (MTS) produces finished goods to a forecast and ships orders from inventory; make to order (MTO) starts production only after a customer order arrives. Assemble to order (ATO) stocks components and assembles on order; engineer to order (ETO) designs and builds after the order. The four differ in one variable: where the customer order enters the process.
That entry point, the decoupling point, determines your lead time, your inventory bill, and how much forecasting risk you carry. Get it wrong per product and you either sit on stock nobody ordered or quote lead times nobody accepts. This post walks through the four strategies, the decoupling point that unifies them, the lead-time-versus-inventory tradeoff, and how real plants mix strategies across their catalog.
What Are the Four Fulfillment Strategies?
The four strategies are make to stock, assemble to order, make to order, and engineer to order, ordered here from most forecast-driven to most order-driven.
| Make to stock (MTS) | Assemble to order (ATO) | Make to order (MTO) | Engineer to order (ETO) | |
|---|---|---|---|---|
| Order triggers | Shipment only | Final assembly | Production | Design, then production |
| What you stock | Finished goods | Subassemblies and components | Raw materials | Little beyond standard stock |
| Customer lead time | Shortest, ship today | Short, assembly time | Medium, full production time | Longest, design + build |
| Forecast risk | Highest, wrong forecast = wrong inventory | Moderate, components are shared | Low, you build what sold | Minimal |
| Fits when | Standard products, steady demand | Many variants from common parts | Custom specs, tolerable lead times | One-of-a-kind engineered products |
| Typical examples | Consumer packaged goods | Configured equipment, computers | Custom components, labels | Capital machinery, tooling |
What Is the Decoupling Point?
The customer order decoupling point is the position in your material flow where forecast-driven production ends and order-driven production begins, equivalently, the last place you hold inventory in an unassigned state. Upstream of the point, you build on speculation; downstream, everything has a customer's name on it.
Thinking in decoupling points beats thinking in labels, because it turns a four-way category debate into a single question: how far downstream can we afford to hold inventory, and how much lead time will the customer tolerate? Push the point toward finished goods and lead time shrinks while inventory and forecast risk grow. Pull it back toward raw material and inventory shrinks while the customer waits longer.
What Is the Lead Time vs. Inventory Tradeoff?
Every move of the decoupling point trades customer lead time against inventory investment: stock more, ship faster; stock less, quote longer. Neither end is free, and the cost of the stock side is bigger than most income statements make it look. The U.S. Census Bureau's M3 survey put total manufacturers' inventories near $962 billion in May 2026, against an inventories-to-shipments ratio of 1.47, meaning the average manufacturer holds roughly a month and a half of shipments as inventory (Census Bureau, Manufacturers' Shipments, Inventories, and Orders). Some of that is deliberate decoupling stock. Plenty of it is forecast error wearing a strategy costume.
Two tools manage each end of the tradeoff. On the stock side, safety stock math sizes the buffer at the decoupling point against demand variability, so the buffer is a calculation rather than a comfort blanket. On the schedule side, the master production schedule is where the strategy becomes operational: MTS items are scheduled from forecast and reorder logic, MTO items from the order backlog, ATO items via a two-level schedule, forecast the options, assemble to the order.
How Do Forecasting and Safety Stock Change Across the Strategies?
Forecast risk lives upstream of the decoupling point, so where you place the point decides what you have to forecast, and how badly a miss hurts. Under MTS you forecast finished-goods demand SKU by SKU, and every unit of forecast error becomes either a stockout or a pallet of the wrong product. Under ATO you forecast at the component level, which is structurally easier: ten end-item variants built from three subassemblies means forecasting three demand streams instead of ten, and component demand aggregates more smoothly than end-item demand. Under MTO and ETO the forecast shifts from products to capacity, you are no longer guessing what to build, only how much labor, machine time, and long-lead material to have ready.
The same logic sizes the buffers. Safety stock belongs at the decoupling point and is sized against the variability of what sits upstream of it: finished-goods buffers for MTS items, component buffers for ATO, raw material and capacity headroom for MTO. A common and expensive mistake is carrying finished-goods safety stock on items the plant claims are make to order, that is not a hybrid strategy, it is an unmanaged one. If an MTO item keeps needing a finished buffer to hit its promised dates, the honest fixes are either shortening production lead time or admitting the item has become MTS and planning it that way.
How Do You Choose a Strategy for Each Product?
Choose per product family, not plant-wide, by comparing what the customer will wait against what the product costs to hold. The working sequence:
- Measure tolerable lead time per product family. What the market actually accepts, from won and lost quotes, not sales folklore.
- Measure your cumulative lead time. Raw material procurement through ship. If tolerable lead time is longer, MTO is available; if shorter, some stage must be pre-built.
- Score demand predictability. Stable, forecastable SKUs can carry MTS economics. Lumpy or customer-specific demand cannot, forecast error becomes dead stock.
- Check variant structure. Many end items from common subassemblies is the classic ATO signal: forecast the components, assemble the combination the order asks for.
- Place the decoupling point and size its buffer. Hold inventory at the last point of commonality, sized with safety stock logic, not gut feel.
- Encode it in the master schedule and revisit yearly. Products migrate, a custom item that standardizes may earn MTS treatment; a declining SKU should lose it. The MPS is where the decision lives day to day.
Why Do Most Plants End Up Hybrid?
Because catalogs are not uniform. A typical mid-size plant ships a handful of high-volume standards that deserve MTS, a long tail of variants that only make sense MTO, and configured products in between that fit ATO. Running one strategy across all of them means either warehouses of slow movers or lost orders on fast movers. The hybrid is not indecision; it is the correct answer, managed SKU by SKU. Production modes interact with this too, batch production plants often run MTS on A-items to fill efficient batch sizes while quoting MTO on specials.
The catch with hybrids is visibility. When MTS replenishment and MTO backlog compete for the same lines, the daily fight is priorities, and it gets decided well only when planners can see real-time order status, inventory positions, and line capacity in one place instead of three systems and a spreadsheet. That is the class of problem a shared operational layer with live scheduling and shortage intelligence exists to solve: the strategy is set annually, but it is executed hourly.