Observations from SxSW Day 1
March 14th, 2009I thought it might be interesting to document each day that I am at the SxSW Interactive conference. I think the easiest way to do this will be to list a few takeaways from each session I attend:
We arrived in time to attend 1 session on Friday, so I’ll include that with Saturday’s sessions.
- Because of the economy, newspapers will make a transition in 2 years that otherwise would have occurred over the next 10 years
- Newspapers should stop paying for content the web is already producing for free; do what you do best and link to the rest
- News will become a lot more distributed
- Generate ideas in moderation, quickly kill ideas
- Share ideas liberally, if they are stolen, they weren’t unique enough
- Creativity + Organization = Impact
- Achievement = Actions Taken (not ideas generated)
- Require constraints from clients – budget, colors, etc. they are there even if neither you or the client acknowledge them
- Recognize, measure and celebrate actions taken
- And, the always popular – find time to focus on the important not always the urgent
Microformats: A Quiet Revolution
- Microformats are a set of simple, open data formats built upon existing and widely adopted standards
- Microformats are being used on a lot more sites than most think – google, yahoo, Microsoft, ebay, WordPress
Well, in the interest of honesty, I skipped the opening remarks to watch the Ohio State Buckeyes beat Michigan State in the Big 10 Men’s college basketball tournament. I will watch the YouTube video as soon as it is released. (Don’t forget, it was a Saturday!)
Kicking Ass With Controlled Metadata
- In many cases, user generated tags are not enough, experts and/or paid professionals are needed
- Pandora uses music professionals and tags songs in over 400 dimensions
- It can take a lot of tags to ensure relevance
- Some site are now sharing tags with and API
Collaborative Filters: The Evolution of Recommendation Engines
- Collaborative Filters combine input from many different people to filter information better than it otherwise would be filtered
- Recommendation Engines personalize these results
- 10 years ago, everyone saw the same Netflix site. Now each user has a unique site
- 2 types of data; 1) entered by users; 2) learned from user actions
- When last.fm enabled users to share their song lists, it changed their listening to hide “guilty pleasures”
Time to figure out my schedule for tomorrow and enjoy downtown Austin.
