It depends on what you mean by "low quality." Enterprise products often don't have shiny GUIs, extensive marketing associated with them, cool domain names, etc. But they do have to avoid permanently deleting data or being inaccessible for long periods of time. They usually come with 24/7 support contracts.
A social networking website can be down for a day for a few users and nobody will care. With enterprise software, that could be a deal-breaker. If engineers have to be woken up at night, then they will be.
You're right that enterprise is about sales, whereas consumer is about marketing. It actually affects the company as a whole more than you might think. Heavily marketed consumer-focused companies never seem to stop marketing themselves. Even in the cafeteria, you're bombarded with the marketing about how this is the best company EVAR and all the cool people are here, etc. Sales is more about relationships and building trust. There are some bad aspects to both, but on the whole I think I prefer a sales-focused company.
I fear that I sounded like I was making a value judgement. I wasn't.
Generally speaking, I think enterprise apps are simply easier to deliver than successful consumer apps (most consumer apps aren't successful and nothing is easier to deliver than an app nobody uses). There's a bunch of reasons for this:
(i) famously, the person making the purchasing decision for most enterprise apps isn't the eventual user, so to make sales and achieve usage you need please only during demos;
(ii) the bar is much lower in the enterprise because you are generally solving money problems (increasing revenue or decreasing costs/headcount); most enterprise software is clunky, and the software virtually never delights, but for the most part the line-of-business software deployments I've worked with tend to make lots of money;
(iii) the fit/finish and scale problems that HN people hyperfocus on simply tend not to be problems enterprises pay to solve; also, once a firm verifies for itself that it'll realize a return on a deployment, it'll tend to be happier to simply scale up its Oracle deployment rather than trying to be clever with how it persists data.
(iv) a six figure pilot deployment pays for a whole lots of hardware, and successive purchases, which tend to happen after the firm verifies that the software is worth something to them, are even more expensive.
(iv) relatedly: the financial metric most enterprise software is benchmarked against is "fully loaded headcount cost", which, wow, is a much nicer benchmark to have than "$5/mo" or (worse) $2.99 "premium" app store app.
I think these things tend to add up to enterprise software being much, much easier to build and deploy, even if you want to believe that there's some countervailing cost of additional testing (I actually don't buy that enterprises really demand a higher standard of testing, either).
Things like Yammer and Salesforce are practically the exception that proves the rule; the very rare app that needs to scale multitenant across every company in the world. And a gargantuan swath of the entire Fortune 500 could care less about salesforce management.
Not all is roses in the enterprise software world.
Businesses will often refuse to buy from companies they regard as likely to disappear in a few years-- unfortunately, most startups tend to get lumped into this category by default. The burden of proof is on you to prove otherwise. (This is rational behavior on their part, by the way.) Managers are naturally conservative. They know that signing a slightly too-expensive deal with, say Oracle (to pick a random example), won't get them fired-- but signing a deal with some startup that fails to deliver surely will end their career.
Finally, there's the problem that because sunk costs in existing systems are often high, it can be hard to break into a market. This is what keeps COBOL and even FORTRAN chugging away in corners of the world. Costs might be lower a few years after replacing these systems-- but then again, they might not. It depends on who you get to put in the replacement, and if you have trouble telling the good developers from the bad, maybe you are better off not touching it. Building and deploying is one thing. Selling is quite another, and it's not easy or cheap. Forget about word-of-mouth and friends-of-friends.
Do enterprises demand a higher standard of testing? I think it depends on what your product does. I work in enterprise storage, so for us there absolutely is a higher standard of testing. I can't speak for other applications, though.
Things like Yammer and Salesforce are practically the exception that proves the rule; the very rare app that needs to scale multitenant across every company in the world.
Yeah, but that's where the fun is for engineers. And in time, companies like these will reduce the need for so many companies to build custom solutions.
A social networking website can be down for a day for a few users and nobody will care. With enterprise software, that could be a deal-breaker. If engineers have to be woken up at night, then they will be.
You're right that enterprise is about sales, whereas consumer is about marketing. It actually affects the company as a whole more than you might think. Heavily marketed consumer-focused companies never seem to stop marketing themselves. Even in the cafeteria, you're bombarded with the marketing about how this is the best company EVAR and all the cool people are here, etc. Sales is more about relationships and building trust. There are some bad aspects to both, but on the whole I think I prefer a sales-focused company.