A: At a high level we care deeply about two types of metrics: what people do and what people say. We believe both are important in understanding how well we're connecting people to the stories that matter to them.
What People Do
We try new things every week, new designs for story formats, new ranking changes, etc. What we look at with these tests primarily is whether or not these changes increase engagement. Do people spend more time on News Feed? Do they like and comment and share more? These behaviors are proxies for value. If someone is commenting on more in general that's a signal that they're finding their News Feed more interesting.
What People Say
We know that we can't fully understand if a story is interesting to someone based if if they liked or commented on it. A story might be a status update about losing a loved one, which may be very important but not something you would 'like.' So we've spent a lot of time and energy on something we call the 'Feed Quality Program' which includes tens of thousands of online surveys a day and a panel of professional raters that give us detailed feedback about their experiences. An example of a metric that comes out of these programs is 'transposition score,' which is a measure of how differently we ordered your News Feed from the order you would have put stories if you had spent a few hours doing so. Over time we work to minimize that difference.
A: We should always be solving a problem for people as we build. Balancing creating more functionality for some and keeping things simple and easy to understand for others can often be tricky, especially if you're building for a broad audience.
Personally I'm a big fan of progressive disclosure. A UI can adapt, expand, and even get more complicated for those who might benefit from it while remaining simple for those who might not. A good example of this is most email apps that support multiple accounts. If you have one account you have one inbox, one sent folder, etc. But if you add multiple accounts the experience starts to adapt, and yes, in turn, become more complicated, as it meets people's needs.
We do some of this in Facebook today. When you first sign up the experience is largely oriented around getting you connected to your friends, because Facebook isn't useful until you've found friends and followed pages. That said I think we can do more here. A lot of people use our products, and they have a huge range of interests and needs, so our experience should adapt accordingly.
A: Our mission, as a company, is to make the world more open and connected. When deciding what projects to work on we are always tying things back to how much we think they might further our mission.
We size each idea that we consider seriously, by looking at how many people we think it will benefit over the long run, and how much value we think we will be creating for them. So even a small change can be prioritized highly if we believe it will make hundreds of millions of people's experiences just a little bit better. It's worth noting that it's critical that we think about the long term, as sometimes the most impact can be had by a project that will take a long time to materialize.