I think there is a balance between code, marketing, and customer service: if no one knows about your product, you have no customers (marketing); if your product is buggy, you lose customers (code); unless you apply the third and fix bugs for your customers at the drop of a hat (customer service) in which case you'll probably not only retain customers you would have lost, but gain customers from word-of-mouth.
Early on? Marketing is the most important thing you need, since you need the users. You can get away with a crappy code since people are willing to give you the benefit of the doubt. Customer service is more or less useless too, since you don't really have any users to customer serve.
Mid-Stage? Coding is the most important. Best example is twitter. At this stage you are growing more or less organically, so you need to have great code in order to scale properly. You also need that good code to release more and more features to keep up with the competition.
Late-Stage? Customer service is the most important. At this stage your product is more or less polished, and you pretty much got as much market share as you can get. At this stage you need great customer service to keep the users you already got.
I disagree strongly. If you have ever looked at 'enterprise' software, it's generally crap. have you ever tried to use remedy? it looks like something I wrote when I got my hands on my dad's copy of FoxPro as a 14 year old.
People tolerate bugs if you are a big guy marketing to other big guys. middle management would rather do what everyone else is doing (and thus not get fired when it breaks) even if what everyone else is doing is clearly worse than some other product. they have no incentive to choose the better product, and a strong incentive to do what everyone else is doing.
Heck, even in the consumer market, crap is tolerated if it's big. look at the downtime twitter has.
On the other hand, when you are first starting out, people expect to write you off as incompetent. And they will, quickly, if they see a misstep.
I would argue that people expect more out of small and or new companies. Maybe they should; there are many inefficiencies inherent to the standard corporation model.
