A simulated 30-year climate data record of net cloud radiative effect (defined as the difference between clear- and all-sky net top-of-atmosphere radiative flux) based on the first 5 years of Clouds and the Earth's Radiant Energy System (CERES) Terra measurements is created in order to investigate how gaps in the record affect our ability to constrain cloud radiative feedback. To ensure a trend estimate with an uncertainty small enough to constrain cloud radiative feedback to 25% of anthropogenic forcing in the next few decades, the absolute calibration change across the gap must be <0.3% in the shortwave (SW) region and <0.1% in the longwave (LW) region for a 1-year gap occurring in the middle of the record. Given that current calibration accuracy of CERES is 2% in the SW and 1% in the LW (at the 95% significance level), a gap of any length anywhere in the record will significantly increase the time required in order to detect a trend above natural variability because data collected prior to and after the gap cannot be combined accurately enough to ensure trend detection at the required level. To avoid gaps, at least 6 months of global or 1 year of tropical overlapping measurements between successive instruments are needed, based on overlapping CERES Terra and Aqua data.