Configure to order is the deal a manufacturer makes with complexity: do the engineering once, up front, and let the customer do the choosing. Every option is designed, validated, costed and routed before anyone orders it. When an order arrives, nobody draws anything. The order simply picks a path through decisions that were already made.
Office chairs, commercial door hardware, HVAC units, switchgear and most trucks are sold this way. So is a pizza.
How it works
CTO has two halves that run on different clocks.
Before any order. Engineering releases the modules and options. Someone like me turns them into a product model: attributes, allowed values, constraints, a 150% BOM with a condition on every line, routings and a price build-up. This is slow, careful work, and it's where almost all the risk lives.
At order time. A dealer or salesperson configures the product in a configurator or CPQ. The rules filter the 150% BOM down to the order's 100% BOM, ERP schedules it, and the plant builds it. If the first half was done well, the second half is boring. Boring is the goal.
Where CTO sits among the strategies
| Strategy | What exists before the order | Typical lead time | Variety |
|---|---|---|---|
| Make to stock | Finished goods on a shelf | Days | Low |
| Configure to order | Engineered options and modules | Weeks | Very high |
| Make to order | A released design, built per order | Weeks | Medium |
| Engineer to order | A capability, not a design | Months | Unlimited |
Most real companies run more than one. A furniture maker can sell a configured task chair, stock a popular side table and engineer a one-off reception desk, all in the same week.
Where it goes wrong
- The model drifts from the plant. Engineering changes a part, the rule set doesn't follow, and the configurator keeps selling something that no longer exists. The result is order fallout.
- "Special" becomes a habit. Every non-standard request that gets handled by hand is a little engineer-to-order work hiding inside a CTO business. A few are healthy. Hundreds a month mean the option model is missing something customers actually want.
- Options are added and never removed. CTO makes adding variety cheap at the moment you add it, and expensive forever after. That's how zombie variants are born.
Example
A commercial lock family offers 5 functions, 3 levers, 4 finishes, 3 cylinder types, 4 handings, 2 door thicknesses, 2 fire ratings and 3 electrified options. That is 8,640 combinations on paper. Nobody stocks them. The plant stocks a few hundred parts, and the Rules Lab shows how the right ones are picked for each order.
Questions people ask
Is CTO the same as make to order?
Not quite. Make to order builds a known design once an order arrives. Configure to order also lets the customer choose the design from a predefined option space. All CTO is make to order; not all make to order is configurable.
When should a company move from engineer to order to CTO?
When the same “custom” requests keep coming back. If engineers are redrawing the same five variations every month, those variations belong in the option model.
Updated November 2026. Think I've got something wrong? Tell me.