It is widely believed that competitive elections are required for good democratic performance. Many races, however, see little electoral competition, due to asymmetries in voters' evaluation of candidates' quality (due, for example, to incumbency) and party labels (due, for example, to ideology). We study the consequences of both types of imbalances in a unified theoretical framework building on the notion that voters are rationally ignorant and need to pay costly attention to learn about candidates. Our paper rationalizes key empirical regularities such as the existence of large incumbency spending and electoral advantages or the heterogeneous effect of incumbency. Further, we highlight that properly accounting for voter attention is critical to interpreting empirical estimates of key determinants of electoral success, the sources of the incumbency advantage, and the causal effect of incumbency status. We also show that while depressing electoral competition, imbalances nonetheless improve voter welfare.