The original promise of on demand computing, software as a service, cloud computing or whatever catch phrase you want to use was that the companies would be easier and cheaper to build than traditional enterprise software companies and they’d be far more customer friendly.

The former wasn’t quite right, as we discovered in the early-to-mid-2000s during the funding wave that brought us Salesforce.com, Netsuite, Taleo, Omniture, RightNow and others. Beneath those hits– of dramatically varying size– there was a lot of carnage. It turned out you still need massive sales and marketing budgets to sell this software. It wasn’t going to just sell itself.

Meanwhile, on the revenue side there were no more big quarter-making, up-front multi-million deals. Even the biggest winner of that wave, Salesforce, has been under a constant cloud of shareholder griping that it’s not profitable enough as it continually– and wisely– plows its revenues back into sales, marketing and growing the business.

Like Web 2.0, ecommerce 2.0 and nouveaux hardware companies have all learned, the fact that it’s easier to get going, doesn’t mean building a large, scalable company is necessarily any cheaper, faster or easier. In each of these corners of the industry, there are still hard and fast rules of gravity.

The latter promise of software as a service– that it would be more customer friendly– has held up a lot better. I mean, it’d be hard to do much worse than traditional on premise software vendors. They designed software that no one wanted to use, sold it for millions of dollars, forced companies to spend millions more on consultants to get it to work, arm-wrestled their employees into using it, and continued to pay maintenance and licensing fees until the end of time for the privilege.

In contrast, software-as-a-service vendors all took a very consumer Web UI approach, falling all over themselves to make the software easy and intuitive for anyone to use. And because you are running on their servers, they can carefully watch how you use the software and where you get stuck, making tweaks on the fly and pushing new versions all the time. No more crazy upgrade cycles.

There were no more crazy consultant bills, either, because the whole point of software as a service was that it’s hosted by the software vendors. Most allow you to try before you buy, or even stick with a free version forever. And those who pay, usually pay by the month with no lock in. If it doesn’t work anymore? Just stop paying.

It’s an improvement in customer service that would make Tony Hsieh jealous. So, far be it from me to point out the fly in the ointment of business software customer nirvana…

But years into practice of actually using this stuff, the idea that customers aren’t locked in is silly. Sure you aren’t locked in because of sunk cost of millions of dollars or kludgy IT configurations, but there’s a far more powerful lock-in: Switching to something new would utterly wreck productivity. And I don’t know a business in the world that can afford that.

The more mission critical the piece of software is to your business, the more trapped you are. It may even just be the specter of lost productivity. Maybe if you pulled all of your sales reps out of Salesforce, it’d be fine. But why on earth would you take that risk if making your quarter depended on it?

I’ve experienced this as an entrepreneur directly. Everyone at TechCrunch always bitched about running the newsroom on Yammer, but it was what I knew and it worked, so we used it here. It works. It works really well. But it’s buggy and not particularly feature-rich. It has annoying idiosyncrasies like any software. It needed work. When Yammer sold to Microsoft, I had no confidence the software would get any better. If anything, it would only get worse. So I publicly said we were going to switch. As a consumer, I love nothing more than declaring boycotts and following up on them doggedly for years.

Guess what? It’s some five months later, and we’re still on Yammer. It’s still buggy. I saw that a new Yammer competitor called Unison is launching today. I might look at it. But I doubt we switch.

It’s not because I don’t want to. And it’s certainly not because of sunk costs, because we don’t pay a dime for Yammer. It’s because rapid, pushed communication in a group is mission critical to our newsroom functioning, and I can’t risk the loss of productivity now that our growth is really starting to take off. Particularly when something works good enough. And Yammer most definitely still works good enough.

See those pretty, cloudy things around your legs? Those are shackles. They may hurt less and give you more room to move around. But they’re shackles nonetheless.

While frustrating for business owners, this is clearly good news for cloud vendors. Particularly those with hopes of building a $1 billion company off of a freemium model. Chalk up another reason investors are bullish on business software over consumer these days. At least these new companies have kept one business advantage from the on-premise world.