Several factors contribute to the low level of improved variety use in Ethiopia. Among those, on the supply side, is the limited availability of seed in the volumes, quality, and timeliness required by farmers, which is partly a result of limited public and private investment in the sector. Beginning in 2011, the Government of Ethiopia introduced a novel experiment-the Direct Seed Marketing (DSM) approach-to reduce some of the centralized, state-run attributes of the country’s seed market and rationalize the use of public resources. DSM was designed to incentivize private and public seed producers to sell seed directly to farmers rather than through the state apparatus. This study is the first quantitative evaluation of DSM’s impact on indicators of a healthy seed system: access to quality seeds, on-farm productivity, and market participation of smallholders. Using a quasi-experimental difference-in-differences approach, the study finds that DSM led to a 26 percent increase in maize yields and a 5 percent increase in the share of maize harvest sold. DSM also led to improvements in seed availability for all three of Ethiopia’s major cereals: maize, wheat, and teff. However, DSM’s effects on yields and share of harvest sold are not statistically significant for wheat and teff. These crop-specific differences in performance are likely explainable by biological differences between hybrid maize and openly pollinated varieties of wheat and teff that incentivize private sector participation in maize seed markets over wheat and teff seed markets. These differences demand different policies and perhaps even institutional approaches to accelerating adoption between hybrids and OPVs.