30 April 2006

Caterina Fake answers questions about starting Flickr

(This is a quick upload of my notes. Please give me feedback in the comments section at the end. Thanks. --Pat)

At Startup School today at Stanford, Caterina Fake answered questions about Flickr - starting it, what like is like post-acquisition, and how it came to be. I also have entries for the talks by: Joshua Schachter (del.icio.us, memepool), Paul Graham (YCombinator), and Om Malik.

Flickr grew out of the efforts of Caterina and Stewart Butterfield's company, Ludicorp. They set out to make an online multiplayer game and had a successful prototype called Game Never Ending (GNE). However, the backend development on this game fell behind the client development, and so they put together seven products in six months using the basic GNE framework. Flickr was one of these.

In its original form, it was a chat client with the ability to share photos. But as they watched their users, they realized that the chance of two people winding up on the same chat at the same time was small, so as an afterthought, they added the ability to share photos via web pages.

The new application took off. Ludicorp only had six people, so they chose to pursue this unintended direction and kill off GNE.

They announced Flickr at O'Reilly's ETech conference in 2004. By November of that year, they were talking to Yahoo, and the acquisition was finalized in March 2005. The Flickr team moved to the Bay Area in June 2005.

After talking about the history of Flickr, Caterina took questions. I've transcribed some of these here:

What did you do wrong, and what would you do differently?

When you're small, each person is critical. Early on, Ludicorp/GNE probably hired too fast and fired too slow. The backend fell behind as a result. But that led to Flickr being born, so maybe it wasn't so bad :)

"Foo"-up fast (she joked about going for a PG rating in her talk) development strategy - you should create and kill lots of projects. Learn from mistakes rapidly (mistakes are inevitable, it's trying things and learning from them that's important).

They did no formal usability testing, and instead relied on a constant dialog with users.

"Get it out early" is the best usability testing.

As an example, they redesigned the front page back when they had 25k - 50k users. Within ten minutes of pushing the new page out, they new it was a bad design. Their users were not reluctant to tell them.

Did you consider video sharing?

You need to be able to say this is what it is. GNE was a big morass, so they had to strip it back. [that said, I got the feeling they're still thinking about video --P]

Starting as a game was useful. It gave the naive optimism. Had they started on day one designing a photo site, it wouldn't have been as successful or playful.

What did they look for in their negotiations with companies

We were looking for a place that would preserve our essence, and after seeing Yahoo from the inside and out, they felt it was a good fit.

How did you get your PR? Did you hire people to talk about you?

Flickr did not have a PR budget, so they had to come up with features that would get the word out. Stewart and Caterina were bloggers since 1998 or 1999, so they deliberately designed blog-friendly features like the "blog this photo" button.

80% of Flickr's new users found Flickr through blogs.

Flickr's initial press was unsolicited from alpha geek writers.

Stewart and Caterina would give interviews to anyone who asked, even high school newspapers.

They eventually hired a PR firm not to generate publicity, but to handle it.

How did you design the UI

The user interface evolved organically. They wanted to preserve so much of why the internet was interesting -- people with interests in a culture of sharing. Back when, people would hang out on IRC channels and have long conversations about their interests (danish scholars and US techies). It was a culture of generosity.

Caterina quoted David Weinberger: "strangers are the source of fear - they lurk in the shadows - but on the internet, strangers are the source of everything good."

They wanted to make Flickr so that users were facing each other, instead of facing the website.

To encourage this, George Oates and Caterina spent almost 24/7 greeting each person who came in to the original chat client when it launched. They had an app called "the newbie spotter" which would tell them who was new and what their interest was (you filled this in on your user profile) so Caterina or George would then introduce them to other users: "I see you like ... Black Metal. User GeorgM in Berlin also likes Black Metal."

Flickr is like a party with a good host. Someone there to take your coat, get you a drink, and introduce you around. By setting up this culture of meeting people, they made Flickr a distinctive, friendly, and ultimately successful site.

No comments: