Why personalisation is a bad thing for most business

A few years ago, personalisation was the next big next big thing. I worked on a personalisation project for my previous employer that sucked up about a million pounds of software licensing, and heavens knows how many people's salary, lives for six months, etc etc. And when they got to the end of it, what did they have? Not personalisation, by any stretch. Maybe the ability to segment some of the emails they sent to their customers, but that was about it. (And now it doesn't look like they segment at all, or at least not to any of the things like purchase history that are meant to segment mailing lists. They leave it up to the customer to say what newsletter they want to receive, which, in all honesty, isn't personalisation except in the most trivial sense. It's good marketing, but that's a different beast entirely...)
But there are a number of reasons why personalisation isn't a very good goal for most companies. Consider some of the fallacies:
We've got to be like Amazon OK, that's great. Amazon hold a lot of data on things different people purchase, and then tailor the home page and recommendation lists to accord with your purchase history. But they're Amazon. They got there first, and unless you're also a huge internet retailer, you won't have either the computational power to match them, or the flow of data to analyse in the first place. Amazon get lots of low ticket items with high frequency, which is a good place to be in, but it's occupied by people like them and FMCG retailers. An awful lot of the rest of the market just isn't like that.

We need to be a community Yes, because people spend a lot of time on community sites like Flickr and Friendster (remember them?) or Myspace, or ...
Yes, but they're sharing photos, or glomming contacts from old friends for future business networking, or being teenagers. They're not buying your product, and unless it's a product people want to talk about a lot, with other likeminded souls, chances are that you're not going to build much of a community up.

We need eyeballs! Yes, you're right. The more impressions you get, the more money you'll make. And it's still 1997, right?
Sorry, this just doesn't stand up against any rigorous mental effort. Sure, if you got 1 purchase for every 1,000 people that looked at your home page, if the conversion rate stayed the same then you'd double your purchases if you had 2,000 people. But the conversion rate won't stay the same if you just grab impressions from everywhere. Quantity does not always have a quality of its own. That bloke that stands on Piccadilly Circus with a megaphone proselytising for Jesus sure has a wide audience, but does he get many converts? And what if you've got 10 million visitors every day, but only a thousand of them purchase? You've still got all that extra bandwidth that you have to service, servers to pay for, people to look after them ... Oh, and why did you need personalisation to get that volume anyway? Why didn't you just get some compelling content? People don't go to the cinema because they get a personal experience, they go because there's something good to watch.

I'll come back and edit this post a bit later. There's a few bits of statistics to talk about, and maybe a rant about the long tail. Then I'll try and be positive and say what I think is more worthy than personalisation, or at least personalisation per se.