For several years now, many of us doing research into designing technology for health and wellbeing have leveraged mobile, ambient and wearable devices in order to provide feedback and insight into users' mental and physical states. We have designed our technology, in partnership with caregivers and clinicians, in order to complement and extend clinical care so that it reaches those that really need it. It seems like a good time to stand back and reflect on what has actually worked in terms of motivating our users to make healthier lifestyle decisions, which in turn steer them towards long-term behavioral change, if needed. Specific to our research domain, emotion sensing has become ubiquitous in the physiological sensing and affective computing communities. While we leverage these methods in our research, we have found that the truly difficult problem is "what you do about it" once you have identified a user's emotional state. This keynote will describe various lessons learned from several efforts in this space, as well as traps to avoid, if you want to design engaging and life-changing interventions to help users cope positively with stress, depression, diet, exercise, sleep, and productivity.