In the Linux kernel, there are tradeoffs between latency and throughput. The same principal can be generalized to personal time management. When there are multiple streams of requests coming in, there is a switching cost in switching one's attention between the streams, whether you are a computer or a person. You can maximize throughput by not switching until you are ready. But if you do that, then some easy-to-fulfill requests will end up waiting a long time because if they were prioritized lower than (or simply arrived later than) some long request. You can minimize latency by switching between the streams frequently, allowing you to respond to all short requests very quickly; but this incurs greater switching costs.

In email, the maximum throughput solution means not doing your email until you are good and ready, and steadily going down the email and replying to the most important ones in order, even if this means taking a long time to get back to some less important ones that only take a minute to reply to. The minimum latency solution means constantly triaging and replying to the easy-to-deal-with emails first.