You've created an awesome product and you're confident people will buy it in droves. Here's the problem: How do you determine just how much of that product you should order for your first run? You don't want to run out-of-stock (though there are worse things in the world seeing as Apple, Nintendo, and other popular brands do so all the time), but you also don't want to see stock piling up in storage because you manufactured too much.
Properly estimating demand means becoming an expert at demand forecasting. It's a bit like forecasting the weather except no one will be mad at you for predicting sunshine on a rainy day. Instead you just run the risk of losing money, which is not as embarrassing but is probably less enjoyable overall.
To expertly forecast demand for a new product, start by asking yourself:
What is the market for my product?
Make a rough estimate of the maximum number of people who might legitimately want your product. Let's use a real product as an example of how you might approach this: The Whiskey Ball.
The Whiskey Ball is simple. It's a silicon ball that enables anyone with access to cold enough air to make large, round ice balls. Who might want to buy this product?
In the not-too-terribly-distant past, only Ice Wizards and people in Iceland (it's right there in both names) had access to the kind of cold needed to generate ice year round. But ever since the invention of the freezer, most people on earth have been able to make ice whenever they feel like it.
That being the case, couldn't you argue that most people on earth are potential customers? Potentially, but the real value of a round, larger ice cube is that it takes significantly more time to melt, helping ensure that whatever you're drinking isn't getting watered down. That makes it most appropriate for a sippable drink. What kinds of drinks are often sipped? Alcoholic ones. To be even more specific: fancy, expensive alcoholic ones. No one is sipping a red bull and vodka.
So to find the potential market for the Whiskey Ball, you'd want to visit organizations like the Distilled Spirits Council to get an estimate of how many people in the U.S. (or your targeted area) regularly buy higher end liquor. That'd represent your potential audience that you might someday reach. The larger this number, and the fewer current competitors in the space, the higher potential value your product is likely to have.
However, your potential audience is certainly not going to be your actual audience, especially not at first. This has to be narrowed down for accurate demand forecasting, so you should next ask yourself:
What is my potential initial market penetration?
If you're already selling products or services, then you should already have a pretty good grasp on your sales and growth. Hopefully you'll even have statistics on your average click and conversion rates from your advertising. If you were planning to support your new product with an email and FB advertising campaign, you could take your averages and work out a rough sales number for your first run:
If your average email campaign sees a 1.5% conversion rate and you're going to email 5,000 people over the first month then you could roughly expect sales of 75 products.
If your average FB ad sees a 1% conversion rate and you're going to advertise to 30,000 people that fit your potential market then you could expect to sell 300 products.
That'd give you a rough potential total of 375 products sold from your initial run. You could round up or down from this depending on your own feelings, but this would you a base to start from.
What if my business is new?
This is trickier in that you don't have past information that can help you predict how people will respond to your product. Here it'd be a good idea to get some feel for potential demand by allowing pre-orders on your website or by running a crowdfunding campaign through a service like Kickstarter or Indiegogo.
While neither of the options are necessarily predictive of how the product will perform, they will give you an idea if there is an interest in your product as well as initial funding that will help you get it manufactured.
How do I calculate demand after the first run?
A successful first run might have you chomping at the bit to manufacture twice as many the second time around, but you can't be confident that growth will continue or happen in a linear pattern. Is your product seasonal, meaning people are more likely to buy it at certain times of year? If your product is outdoorsy, for example, you'll almost certainly sell more leading up to summer than moving away from it.
Does your business see sales increases at specific times of the year? All businesses and products have peak times (often around Christmas) and slow times (often shortly after Christmas). Don't let sales at peak and non-peak times mislead you. Look for patterns year-over-year and take advantage of them.
What's important is that you don't make assumptions when demand forecasting: consider your own history, consider the potential market and make estimates. Your estimates won't be exactly right, but you've put all this time into planning your great product — selling that product deserves an equal amount of planning time.