The Indian Ocean Geoid Low (IOGL) appears as a prominent feature if the geoid is, as usual, shown with respect to the Earth’s reference shape. However, if it is shown relative to hydrostatic equilibrium, i.e. including excess flattening, it appears as merely a regional low on a north-south trending belt of low geoid. For a mantle viscosity structure with an increase of 2-3 orders of magnitude from asthenosphere to lower mantle, which is suitable to explain the long-wavelength geoid, a geoid low can result from both positive density anomalies in the upper mantle and negative anomalies in the lower mantle. Here we propose that the IOGL can be explained due to a linear, approximately north-south-trending high-density anomaly in the lower mantle, which is crossed by a linear, approximately West-Southwest-East-Northeast trending anomaly low-density anomaly in the upper mantle. While the former can be explained due to its location in a region of former subduction and inbetween the two Large Low Shear Velocity Provinces (LLSVPs), we propose here that the latter is due to an eastward outflow from the Kenya plume rising above the eastern edge of the African LLSVP. We show that, with realistic assumptions we can approximately match the size, shape and magnitude of the geoid low.