In scenario 1)
- you build a car for inventory and bank 85% for a while.
- the customer waits "a little" but is guaranteed to pay (hardly anyone cancels an order due to locked-in $2500 or higher deposit)
- you can sell that inventory car soon after through lot-sales or through inventory browsers and is "available" for sale even before the customer's car is built. It may take a discount due to test-drive miles or age, but so be it.
If you build customer car first, then inventory, the opportunity costs may end up being larger than the 15%+ delta. This is why auto dealerships stock an ungodly amount of on-site inventory rather than having buyers browse the lot and then put in a custom order. How many Ford or Chevy buyers now, including $50,000 or higher SUVs and trucks, are custom-ordering? 5% maybe? To do inventory right, you actually need to build inventory cars ahead of custom orders. The problem then becomes what happens if the economy doesn't support the inventory base. Inventory on the lots of sales locations is there for satisfying new, interested customers. Yes, there are millionaires who kick-tires on weekends and may write a check for a car on a Saturday afternoon. And with easy financing and lower-rates, even someone with a $70k income can get into a Tesla MS 60 or 75 on a lot with 2-year lease with high residual (Q3 2016). The chickens come home to roost on those in late 2018 when the residuals are under pressure. But in order to sell more cars, you need more cars to sell - hence inventory build.