As Mike Hirshland wrote, “last night was a truly epic evening at Dogpatch Labs“. Alongside Facebook, we hosted 150 people at Dogpatch Labs San Francisco to discuss Facebook Connect and demo their implementations and best practices. Over 50 companies were represented and we had teams fought the bad weather to arrive from New York, LA, DC, Montreal, Boston, etc.

The event was epic in part because the demos were diverse and terrific – but it was also epic because ten minutes before the first presentation was slated to go – there was a major power outage across SOMA… leaving 150 of us standing in the dark. After maneuvering our way through the obstacles, we arrived at the Four Seasons conference room and dove right in.

Mike has several takeaways on his blog – and we will do our best to distribute attendee feedback as well, but here are my three major takeaways:

1. Facebook Connect is not a login system. It is much more.

The best Facebook Connect implementations go beyond the login. They must give the user a reason to interact and provide value after the login – and with Connect, the content and data go far beyond registration: photos, birthdays, social graph, etc. JibJab and Personera are terrific examples: by logging in through Facebook Connect, you are enabling a better, more social site experience.

2. Think about your users, not about pageviews or actions.

If you think about the value you are trying to deliver through Facebook Connect, you will construct an experience that encourages and fosters virality. If you think of Connect as a way to drive pageviews or feed posts, you can end up missing the most important factor: ongoing engagement.

3. Test. Iterate. Test. Iterate. Test. Iterate.

You would be amazed at how small tweaks in language, style, color, etc create very significant changes. KISSmetrics gave an outstanding presentation here and folks like Hollrr showed the impact of recent design changes.

Art Chang, a Dogpatch Labs resident and phenomenal photographer, has several great photos of the event. Two of which are below:

Popularity: 1% [?]