In the last few years, apps have become an important tool to collect data. Especially in the case of data on people’s happiness, two projects have received substantial attention from both the media and the scientific world: “Track your happiness” from Killingsworth and Gilbert (Science, 330, 932-932, 2010), and “Mappiness,” from MacKerron (2012). Both happiness apps used the experience sampling method to ask people a few times per day how they feel, what they do, with whom, and where. The collected data are then displayed for the participants in simple graphs to help them understand what makes them happy and what does not. Both studies have collected considerable data without giving participants any financial rewards. But quantity is not everything that matters with respect to data collection, and thus, understanding whether nationally representative datasets can be collected using such happiness apps is crucial. To address this question, we built a new happiness app and ran a case-study with over 4000 participants of the innovation sample of the German Socio-Economic Panel (Richter and Schupp in Schmollers Jahrbuch, 135(3), 389–399, 2015). Participants were informed that the app collects data on everyday happiness after a household interview and asked whether they would like to use the app. In the first year (2015), participants did not receive any reward, and in the second year (2016), a different group of participants received a 50 Euro Amazon voucher for their participation. The results showed that our happiness app cannot generate nationally representative datasets if it is not controlled that all demographic sub-groups have access to a smartphone, are highly motivated with a sufficient reward and data is collected with quota sampling.