Make to stock (MTS) builds products in advance based on forecasts and sells them from inventory. Make to order (MTO) starts production only after a customer places an order. MTS gives fast delivery but carries forecast and inventory risk. MTO cuts finished-goods stock and allows customisation, but customers wait for production and delivery.
Most manufacturers do not pick one for everything. The real decision is where to place the point at which production switches from forecast-driven to order-driven.
How each strategy works
In an MTS system, planners forecast demand, set production plans and keep finished goods in warehouses. A customer order is filled from stock, so lead time is mainly picking and shipping. The risks are wrong forecasts: too much stock ties up cash and may become obsolete, too little means lost sales.
In an MTO system, the customer order triggers the production schedule, and often the purchase of materials. Finished-goods inventory is minimal, though raw materials and components still need planning. Delivery time is production lead time plus shipping, so it matters how long customers will wait and how reliably the plant can hit dates.
Between the two sit other strategies. Assemble to order (ATO) holds standard components in stock and assembles the final product once the order arrives. Engineer to order (ETO) designs the product for the specific customer. These sit on a spectrum, and the key idea is the customer order decoupling point: stock is held upstream of it and work is done to order downstream of it.
Comparison across the criteria that matter
| Criterion | Make to stock | Make to order |
|---|---|---|
| Trigger for production | Forecast and stock targets | Confirmed customer order |
| Customer lead time | Short, from available stock | Longer, includes production |
| Finished-goods inventory | High relative to MTO | Low or none |
| Forecast risk | High; errors become surplus or shortage | Lower at product level, but capacity and materials still need planning |
| Customisation | Limited to standard variants | High |
| Production efficiency | Longer runs and levelled schedules are possible | Shorter runs and more changeovers |
| Capacity use | Smoother, since output can be levelled | Uneven, following order arrivals |
| Typical products | Standard, high-volume, predictable items | Configured, bulky, expensive or low-volume items |
| Main weakness | Obsolescence and holding cost | Delivery delays and peak overload |
Choose make to stock when
- Customers expect immediate availability and will switch if an item is not on the shelf.
- Demand is stable enough to forecast with reasonable accuracy.
- Products are standard, with a long shelf life and low risk of becoming obsolete.
- Production benefits from long runs and level scheduling.
- Delivery speed is a main part of the competitive offer.
Consumer packaged goods and standard industrial supplies often fit this pattern. Even then, good practice is to segment: fast movers get stock targets, while slow movers deserve a second look.
Choose make to order when
- Products are customised, configured or built to specification.
- Demand per variant is low or unpredictable, so stocking each one would mostly create surplus.
- The product is expensive or bulky enough that holding finished goods is costly.
- Customers accept a quoted lead time as part of buying.
- The cost of obsolescence is high, for example when designs change often.
The discipline in MTO is honest lead-time promising. Quoting dates that the plant cannot meet damages trust faster than a longer, reliable lead time. Visibility of capacity and material availability is central.
Hybrid approaches that combine them
Most real operations are hybrid. A common design is to forecast and stock the common components or semi-finished goods (the MTS part) and finish, configure or pack to order (the MTO part). This shortens customer lead time compared with pure MTO, while cutting finished-goods variety and inventory compared with pure MTS.
Another approach is to split the catalogue. Top sellers are made to stock, the long tail is made to order, and the boundary is reviewed regularly as sales patterns change. Postponement is a related idea: delay the step that creates variety, such as labelling, colouring or final assembly, until demand is clearer.
Analytics can support these decisions by classifying items by volume and variability, testing how much safety stock a service target needs, and simulating lead times. They cannot remove uncertainty, and the output depends on the quality of demand history. Items with little history or sudden changes remain a judgement call.
A short decision checklist
- How long will customers realistically wait, and does a competitor ship faster?
- How predictable is demand at the level at which you would stock it?
- How many variants exist, and how many of them sell regularly?
- What does holding inventory cost, including space, capital and obsolescence?
- How flexible is production: changeover times, supplier lead times and spare capacity?
- Can you stock components or semi-finished goods and finish to order?
- Which items should move between strategies as their demand matures or fades?
- How will you measure the result: service level, inventory turns and order lead time together?
Sources and further reading
- Association for Supply Chain Management (ASCM)
- MIT Center for Transportation and Logistics (MIT CTL)
- Harvard Business Review: Supply Chain Management topic (Harvard Business Review)