An often-overlooked aspect of an MCU by junior engineers is its expected and guaranteed availability. Just because a part is available for purchase at the beginning of the project does not mean it will be available for the entire time that the end product will be sold. In the case of consumer devices, this probably isn't a huge issue. This is because these devices can have extremely high volumes, but any single revision is only in production for a limited amount of time (from months to a year or two). 

Contrast the consumer electronics mindset of planned obsolescence with something on the industrial, telecom, or aerospace side. In these industries, development timelines can be measured in years and required support periods are often a decade or more. This is why part availability is a very real consideration. Be sure to investigate the manufacture guarantees on availability and weigh these against their history, reputation, and the risk to the project—it's not a pleasant experience to get 80% of a design complete only to find out that the MCU can't be sourced during a pre-production run!

Now that we've covered some of the initial considerations to be aware of, we'll move on to some of the more unique considerations for embedded processors—hardware peripherals.