One example: I live on the west coast but ran the Boston Marathon the year of the bombings. People closest to me knew I was there and we were able to get in touch eventually. Would this app know to alert my friends and family who didn't know I was in Boston?
Alternatively, we had a 6.something level earthquake near my home at a time I was traveling overseas. My grandparents eventually got in contact but they were needlessly worried since I was thousands of miles away from the event.
In fact nothing in the dead man switch is computationally expensive and all of it is event driven. Its basically just a database of (email, location, last_checked_in). Thousands of users would take up mere kilobytes of storage. You could run the entire thing severless and it would take years before it would even be worth billing you.
I would totally pay $1 a year for this. Maybe more but probably not much more. With just a few hundred users, you'd be well into respectable side project money.
BTW there is one feature that I requested from dms and they told me no. So here is your chance to differentiate your product.
Instead of every x weeks of time, I'd like it if the are-you-alive check would trigger on news events. For example a major industrial accident / terrorist attack / war breaks out in the general region of the world I happen to be in.