When describing and/or evaluating databases, developers often talk about consistency guarantees. However, database and distributed systems practitioners have put forth different definitions of consistency, which Alex DeBrie reviews: 1) the CAP Theorem definition represents consistency as the property that a reader will receive the most recent write for an item from a system. In this setting, developers must evaluate the need for consistency in the context of network failures during replication; 2) the ACID definition wherein consistency ensures that a database transaction cannot result in an invalid state; 3) database consistency models that measure linearizability and serializability. He recommends thinking of consistency not as an absolute value but in the context of tradeoffs, which may be impacted as new technologies become available.