In theory, setting a price should be a rational economics problem. You have a set supply of a product and there's a certain level of demand for it in the marketplace. Since demand tends to increase as prices go down, you simply adjust your price until you've maximized profits.

Reality is more complicated. Technology companies usually don't have a finite supply of a product. And while you may spend a lot to develop software or a mobile service, over time the cost to produce additional units approaches zero.

Furthermore, many startups have a new product for which there aren't competitors for customers to benchmark against.

Under these circumstances "the traditional model starts to behave in weird ways," say Michael Dearing, a professor at Stanford University's design school, who ran pricing at eBay for many years.

In order to set a price, you'll need to form a hypothesis. You can A/B test it and use other analytics to refine it. But don't rely on data alone to inform your decisions. Also take into account input from your customers and employees, what the competition is doing and your intuition.

"Pricing is not a math problem," says Dearing. "It's a judgment problem."