This study presents the results of an experiment introducing commercial rainfall index insurance into drought-prone farming cooperatives in Amhara Region, Ethiopia. We provided a market-priced rainfall deficit insurance product through producer cooperatives, and test a number of potential ways to kick-start private demand. Take-up of the insurance at market prices is very low, between 0.5% and 3% across seasons. When we use a randomized experiment to distribute small free insurance contracts to farmers, 39% of subsidized individuals enroll but this fails to stimulate input use, yields, or income, and nor does it enhance demand in subsequent seasons. A training and promotion on the product improves uptake and willingness to pay, but also does not improve farming outcomes.