Context: Many if not most planetary nebulae (PNe) are now thought to be the outcome of binary evolutionary scenarios. However only a few percent of PNe in the Milky Way are known to host binary systems. The high precision repeated observing and long time baseline of Gaia make it well suited to detect new close binaries through photometric variability. Aims: We aim to find new close binary central stars of PNe (CSPNe) using data from the Gaia mission, building towards a statistically significant sample of post common envelope, close binary CSPNe. Methods: As the vast majority of Gaia sources do not have published epoch photometry, we use the uncertainty in the mean photometry as a proxy for determining the variability of our CSPN sample in the second Gaia data release. We derive a quantity that expresses the significance of the variability, and consider what is necessary to build a clean sample of genuine variable sources. Results: Our selection recovers a large fraction of the known close binary CSPN population, while other CSPNe lying in the same region of the parameter space likely represent low-hanging fruit for ground-based confirmatory followup observations. Gaia epoch photometry for four of the newly identified variable sources confirms that the variability is genuine and consistent with binarity