Running a hosting business is hard work, I did it for over 10 years. It's under-appreciated, users complain about everything, nobody wants to trust you, and many of the customers you do remain want to bleed your support dry. Some of the most successful hosting companies have succeeded by eliminating SLAs and making support hard to reach. Others such as Rackspace make support easy to get, but make sure the customers pay for the privilege. All of the problems with trying to start a hosting company are clear from the responses here.<p>One of the problems in this space is that it's a mature market. Developers are accustomed to Amazon, Digital Ocean, Linode, Azure, and GCE. These are mature products and yes, users are skeptical of new entrants that don't do as much, don't have a complete product, or even have put effort into their website. For HackerNews, it's ironic, but clearly this audience is expecting a mature product not a lean approach. (Self-provisioning while now industry-standard requires a significant effort to prevent fraud & abuse, and slow-provisioning is completely in-line with Lean methodologies)<p>So, some advice:<p>* Don't give up just because of the bad feedback. Iterate!<p>* Focusing on devs is probably wrong if you don't have a plan to leverage individual sales to enterprise sales. Individuals pay less and you need more of them. More customers usually equals more support, they're more fickle, and less likely to be retained. Perhaps surprisingly, they're also traditionally seasonal, with dev-based buy-in being strongest during summer.<p>* Find a hook. Something you do better, or different. For my hosting company, it was virtual servers for only $6/mo, when at the time the lowest priced alternative was $20/mo. We adopted an architecture that allowed us to radically compete on price. It doesn't have to be price, and honestly, it's better that it's not. Also make sure the hook is understood by the market, one of my company's problems was that our low price had us compared to shared hosting or container-based solutions, because virtual servers at that price was too unbelievable.<p>* Hack trust. Some people here mention trust. Now, I think this is over-valued. People put data with companies they really don't know much about all the time. At some point, nobody knew who Dropbox, Docker, or Digital Ocean were. If, however, this proves to be a problem - and it could be in the current surveillance state - then look at ways to hack trust. For instance, a service such as AWS Lambda has a different threat (and trust) model than EC2, based solely on how developers use these products. If you stick with VM-based hosting, then develop partnerships, get reference customers, bring on staff that the community already knows, loves, and trusts.<p>Good luck!