What determines media coverage on cyber conflict (CC)? Media bias fostering misperception is a well-established problem in conflict reporting. Because of the secrecy and complexity surrounding cyber operations (COs), where most data moreover come from marketing publications by private sector firms, this problem is likely to be especially pronounced in reporting on cyber threats. Because media reporting shapes public perception, such bias can shape conflict dynamics and outcomes with potentially destabilizing consequences. Yet little research has examined media bias systematically. This study connects existing literature on media reporting bias with the CC literature to formulate four theoretical explanations for variation in reporting on COs based on four corresponding characteristics of a CO. We introduce a new dataset of COs reporting by the private sector, which we call the Cyber Conflict Media Coverage Dataset, and media reporting on each of these operations. Consequently, we conduct a statistical analysis to identify which of these characteristics correlate with reporting quantity. This analysis shows that the use of novel techniques, specifically zero-day exploits, is a highly significant predictor of coverage quantity. Operations targeting the military or financial sector generate less coverage. We also find that cyber effect operations tend to receive more coverage compared to espionage, but this result is not statistically significant. Nonetheless, the predictive models explain limited variation in news coverage. These findings indicate that COs are treated differently in the media than other forms of conflict, and help explain persistent threat perception among the public despite the absence of catastrophic cyberattacks.