First off, I really enjoyed SXSW and Austin. It was a crazy, busy time. I can’t believe how much it has grown since 2000, when it also felt big and crazy.

A recurring problem, other than getting between the third and fourth floors, was that many panels weren’t what they seemed in the program guide.

I was confused about how that could be and then Sean reminded me that people voted for two-thirds of the content. That sounds utopian, but it has a major flaw:

The crowd has imperfect information.

It’s no wonder I was disappointed with many of the panels. The title and description was all I had to make my choice. The people who chose the panels had only the same information.

I’ve never been involved with putting a conference together, but I imagine it’s a tough job. You receive proposals, look them over, and ask questions of the possible presenters. When I wrote for Webmonkey, my editors would sometimes ask for a longer outline of a promising story. This helped flesh out the idea and let them know whether what I was thinking was the same as what they were thinking.

Provocative titles are important. They grab our attention. Then the content needs to live up to expectations set by the headline.

When the crowd doesn’t know the full story, they’re bound to make really bad decisions. SXSW seems determined to use the Panel Picker voting again next year. I would encourage Hugh Forrest and the other SXSW crew to do something to get better information to the crowd, or include some form of vetting. The conference will be better for it.

The key to being really good at something is terribly simple. Just do it a lot.

This story about an art assignment makes two great points to explain why:

1. Don’t drown in the details
2. Quality improves with each iteration

Thanks to Josh for sending it to my del.icio.us.

Adam Savage of MythBusters is interviewed in a podcast where he talks about how the show decides how to test myths.

“The first test we always look for is, ‘what would two guys in their garage on a lazy Sunday do?’

Then we start to try and bring more equipment to bear on figuring out what’s really going on. But we really always try and start from ‘what’s the dumbest way you could approach this? What’s the simplest possible thing?'”

It’s all about finding the core idea first. Subtract the obvious and add the meaningful.

Listen to the full podcast (23 minutes) to find out about some upcoming shows and how Savage used the community site MetaFilter to research when he needed more ideas.

A lot of us Web people do it. We know we need an obvious feature, so we implement something quickly, but it only gets us halfway there. Maybe we realize it’s not a complete feature, maybe we don’t. Regardless, we leave it half done.

It’s pretty easy to pick on MySpace, so I’ll praise them first. For a service of immense popularity, it is bound to be a favorite place for spammers. They have become very good at finding these bad guys and wiping them out almost instantly. MySpace doesn’t even send me a real email to tell me about this MySpace email. Pretty cool.

So, how do I know they are catching so many spammers?

For each message from a deleted spammer, I get an extra message in my box, cleansed of its spammy contents. Instead, each message explains…

The profile you received this message from has been deleted. Either this user deleted their own profile or MySpace deleted it for spamming and/or content violations.

To remove this message click the delete button below.

Thank you.

P.S. If you’re seeing a lot of old messages in your inbox like this, it’s because someone you know who has written you messages has been deleted. If the messages are new, it’s probably a spammer that MySpace caught in the act.

It’s like a mouse that the cat brings to the doorstep. “See what I did!”

This is a halfway there approach. Catching spammers is important to MySpace, so they don’t clutter my inbox. Yet, they clutter my inbox with their trophies.

A few ways to make this all the way there:

1. Delete the messages, too
2. Create a spam folder and move them there
3. Keep a McDonald’s-like tally: over one billion spammers foiled

Whatever they do, they should go the whole way. Halfway isn’t far enough.

Which of these Amazon homepage designs is better?

It turns out it depends what we mean by “better.” My gut would say the one on the right, because the design is cleaner, simpler. Amazon put it to an A/B test and it turns out the one on the left performed better. Orders dropped significantly when Amazon tested the simpler design.

Greg Linden points to a deck of PDF slides by Amazon employees. It’s a couple years old, but has some more good examples of lessons they learned with hard data to back it up.

For a good overview of A/B testing, something most of us can do, check out Jakob Nielsen’s Alertbox.