There is little evidence on how the large market for credit score improvement products affects consumers or credit market efficiency. A randomized encouragement design on a standard credit builder loan (CBL) identifies null average effects on whether consumers have a credit score and the score itself, with important heterogeneity: those with loans outstanding at baseline fare worse, those without fare better. Selection, treatment effect, and prediction models indicate the CBL reveals valuable information to markets, inducing positive selection and making credit histories more precise, while keeping credit scores’ predictive power intact. With modest targeting changes, CBLs could work as intended.