We partner with Ecuador’s government to implement a centralized school choice system using a Deferred Acceptance algorithm in Manta. Our study evaluates the welfare impact of transitioning from a distance-based assignment system to one that incorporates families’ preferences. Results show that accounting for preferences yields substantial welfare gains. Counterfactual analyses suggest that alternative mechanisms offer limited improvements compared to the benefits of preference inclusion and coordinated assignments. Household survey data on beliefs and satisfaction support these findings, indicating that centralized school choice systems can deliver significant welfare effects in developing countries.