Having a large number of ongoing occurrences beyond the initial spike (high standard deviation).
A typical product life cycle involves early adopters driving mass market demand, where sales peak, then fall-off, and eventually stabilize at a very low level. A long-tail strategy to game development involves something a bit different.
The goal of a long-tail strategy is to minimize inputs like time and money, and maximize outputs over a long-period of time. In contrast, a traditional sales and marketing strategy is to execute min-max efforts towards the goal of high initial sales after release: While most games seek to earn the majority of their sales in the first years after launch, a long-tail game wants to earn small, consistent sales over a long-period of time.
A long-tail oriented game is not going to be positioned as a market leader. Instead aim for the middle or upper middle end of the market. Build a marketable game with fun mechanics and a few hours of gameplay, but don’t go overboard. Release your game and put some time into marketing, but successful long-tail efforts aren’t going to be marketed as much as other titles. This is because the idea is for word of mouth to largely drive sales over time. Sales will trickle in, but once you get multiple titles, all providing a steady stream of income, that revenue can become quite significant. Also the time you don’t spend on extra marketing can be put towards your next title (opportunity cost). A long-tail strategy is not for every developer, or every title, but it is an alternate way to earn money through multiple titles with limited marketing effort.