I imagine it's a very, very hard problem to solve, since you not only need to anticipate demand, but also manage shipping flows throughout the country and place orders ahead of time.
Yep... component lead times, good production and packout lead times, consumer sales, pricing & promotion, etc etc. Some orgs call it a "flow balance" as you're trying to balance out all of the factors that impact sales.
When one forecasts this stuff, you're trying to thread the needle. You want to have enough, but not too much, components on hand, final goods being produced, goods in transit, goods in retail warehouses and goods on shelves. Have too little and you miss consumer sales. Have too much and prices are forced down, killing profitability.
But when something goes screwy, things get difficult. Consumer demand for the new boxes has been stronger than previous cycles. That caused some shortages for PS4 early, for example. But then you have to be careful... ramp production too much and exceed consumer demand... well then you have unhappy retailers and increased inventory carrying costs, forcing price cuts through promotion or MSRP changes.
Like with economic or finance forecasting, though, "bad" is relative to the rest of the industry, no?
Yep. Like Pachter says all the time, you want to be right on the summation of the entire business. If you're forecasting, and in total your forecast error is 10% or something, then you're golden. The individual components of that forecast will show a wide range of forecast error. It happens. But if you're really terrible at forecasting, and your combined forecast error is 20% or more, well, you need a new forecasting group.
Anyways.
