Latency is the delay between doing something and seeing the result. In tech, that usually means the time from when a message or transaction is sent until a system responds.
In blockchains, latency is the time between sending a transaction and having it confirmed in a block. Once you send a transaction, it moves through different nodes, is checked, and then waits to be added to a block. This whole process is what people usually mean by blockchain latency.
Propagation latency is the time it takes for a message to reach other nodes. Validation latency is how long nodes spend checking if a transaction follows the rules. Both types add to the total time you wait for a transaction.
Latency is about how long one operation takes. Throughput is about how many operations a network can handle each second. A network might process lots of transactions per second but still take time to confirm each one. This trade-off is common in blockchains and affects what they can be used for.
Delays happen when there is too much network traffic, when nodes are far apart, or because of the type of consensus method used. If many people move coins at the same time or nodes are spread out globally, messages take longer to reach everyone.
Long confirmation times can make simple actions feel slow and create risks for traders who need current prices. In fast markets, even small delays can let prices change before an order finishes, which can affect a trade’s profit.
Developers and operators try to lower latency by making the network layout better, using faster ways to agree on transactions, and moving some activities off the main chain. Off-chain channels and payment networks allow many small exchanges to happen quickly without waiting for each on-chain confirmation.
People usually measure latency by looking at the average time it takes for a transaction to be confirmed or the round-trip time for messages between nodes. Transactions per second is another measure that shows how a network handles heavy use, but it does not replace confirmation time as a key metric for users.