1. Products which failed because they were targeting their niche poorly (e.g. Juicero, OLPC)
2. Products which failed because they were outright scams (e.g. Theranos, Coolest)
However, I do think the failure of OLPC is still interesting enough to deserve a mention on this list. With OLPC in particular they were trying to be like three different things all at once and they succeeded at none of them.
- OLPC wanted to sell computers to the educational market, so they were selling lots of cheap machines to schools. This is actually not a bad business model; Google and Apple today have basically cornered that market with Chromebooks and iPads. Had OLPC succeeded, we could have had schools covered in cheap XO-1s instead. However...
- OLPC wanted to be a philanthropic mission to the third world, so they were designing the laptop with all sorts of weird features that only made sense for that market. Some of them are reasonable, like they actually intended it to be ruggedized and easy to repair. But they also had some weird features - like, it was designed around being able to mesh-network an entire village together with them. And at one point they were even going to have a handcrank on the thing so you could charge the laptop by hand if necessary. All of these features are unnecessary if you're just selling laptops to schools and they ultimately wound up getting dropped.
- OLPC also was built around it's founder's pet theories about constructionist learning. If you don't know what that means, well, then you're probably not a Hacker News user; because constructonism is how most techies learn a new system. We play around with it. The laptop was designed around a Linux-based UI called Sugar and was specifically intended to push Free Software onto the third world. The idea was that kids could learn how the computer actually worked by modifying the software onboard. It even had a view source button on the keyboard! All of that is catnip to people like us and completely useless for educators.
Ultimately they wound up delivering a very weird product that didn't actually do the thing it promised to do. That's very much in the spirit of a failed startup.