A tiny effect can be statistically significant
A large enough study can find statistical significance even when an effect is tiny. A p-value tells you neither how large the effect is nor whether it matters in real life. A confidence interval can show both the estimated size and the uncertainty around it.