Pastoralist households across East Africa face major livestock losses during drought periods that can cause persistent poverty. For Kenya and southern Ethiopia, an existing index insurance scheme aims to reduce the adverse effects of such losses. The scheme insures individual households through an area-aggregated seasonal forage scarcity index derived from remotely-sensed normalized difference vegetation index (NDVI) time series. Until recently, insurance contracts covered animal losses and indemnity payouts were consequently made late in the season, based on a forage scarcity index incorporating both wet and dry season NDVI data. Season timing and duration were fixed for the whole area (March-September for long rains, October-February for short rains). Due to demand for asset protection insurance (pre-loss intervention) our aim was to identify earlier payout options by shortening the temporal integration period of the index. We used 250. m-resolution 10-day NDVI composites for 2001-2014 from the Moderate Resolution Imaging Spectroradiometer (MODIS). To better describe the period during which forage develops, we first retrieved per-pixel average season start- and end-dates using a phenological model. These dates were averaged per insurance unit to obtain unit-specific growing period definitions. With these definitions a new forage scarcity index was calculated. We then examined if shortening the temporal period further could effectively predict most (>. 90%) of the interannual variability of the new index, and assessed the effects of shortening the period on indemnity payouts. Our analysis shows that insurance payouts could be made one to three months earlier as compared to the current index definition, depending on the insurance unit. This would allow pastoralists to use indemnity payments to protect their livestock through purchase of forage, water, or medicines.