In 2011, the New Media Consortium declared that, although speculative and not yet widely used, learning analytics would see widespread adoption among college and university campuses by 2016. In that same year, Siemens and Long famously argued that our increasing ability to harness the power of institutional data through analytics was essential to ‘penetrate the fog’ that had fallen over the higher education landscape. Looking back at how colleges and universities have invested in data over the past five years, and how rewarding these investments have been for students, it is indeed time for us to celebrate how far we have come. With maturity and experience, however, comes wisdom. After years of capital investment, training, research, and many failed experiments, we now have a breadth of perspective that we could not have had in 2011. More than ever before, we understand the opportunities that educational data can bring, but we also understand the challenges. What began as an extension of business intelligence and decision support systems is now leveraging sophisticated machine learning algorithms to forecast student outcomes. We are learning that more information is not as important as good information. We are learning that powerful predictive engines are useless in the absence of the expertise, effective practices, and cultural transformation necessary to support it. We have come a long way, but we still have a long way to go. It is time to celebrate, reflect, and look forward to how we can and should mobilize educational data in support of students in the 21st century.