"It's easy to get attached to whatever you're working on. It becomes a baby, and it's very hard to abandon your baby. The best way to make a decision is the data-driven way. That's why it's important for products to have metrics. The other thing that helps is to set time horizons. After a certain amount of time -- for example, three months -- if the project isn't going like this [points upwards] then we've done the best that we could, you make the decision to move on." -- Guangwei Yuan