Good support is expensive and not terribly amenable to scripting, unless you want to do it badly (or not at all). It's certainly possible to differentiate by providing great support, it's just convenient (and very profitable in the short term) to not bother. Great support is hard!
I think the public have got far too accepting of little or no support. Yes it's easy to switch, but most people don't unless it's a commodity service - you have to risk the unknown, maybe migrate years of data with little certainty that support in the new home will be better. Sales support tends to be great (and unscripted from more aware staff).
There's a magic size, let us imagine it at 100 people where nearly all human endeavours become crap. Try getting support out of your electric utility, car maker, government, phone service, national charity, hard drive maker etc. Personally I think it's connected with the size where they will still hurt at, or at least talk about, your loss.
It's certainly possible to get far beyond this size and still give great support - look at Amazon and Apple. Amazon got far beyond that size before starting to become a bit rubbish. Essentially until they started ruining the brand trying to move into food, tablets, 3rd party merchants and everything. Apple have the advantage of their restricted OSX hardware, iOS app store and so on that they have a fairly known quantity to provide support for. I think the key to be in this category is what you think your brand stands for.
I think in many, particularly social spaces, it's easy to forget we're not the customer. eg, you can actually get responsive support out of Google adwords or if you buy a Nexus from them (not always great support, but they answer promptly enough). So we're never getting support for search, G+, Facebook or Twitter. I'm sure there;s scope to do it much better, even for the free services.
Lastly, I think specifically in the startup space when you're aiming for 20% growth a month, or even a quarter is going to make it far more difficult to do well. Especially if your offering is evolving as fast - even your staff will struggle to stay aware. So it becomes about great user stories, and outreach (PR) and devcons. Good for growth, but chances are individual customers get lost in the gaps. I think the tendency to shoot for an IPO or buyout is unhealthy here - individual experience matters less if you don't need to care about ongoing trading profit whilst getting big fast. That said there's plenty of services I don't think would have happened, or be great without being big. Small Spotify doesn't get music licences, small Uber doesn't have availability when you need it, etc.
EDIT: Didn't realise I'd written that much!