I think you raise a good point about the snowball effect - people not receiving requests > have no reason to log in > don’t log in and stop showing up on map > worsen chances of receiving request, etc - and it makes me wonder if, rather than just having a fixed timespan of inactivity remove you from the map, it would be better to include requests received and not responded to as part of the calculation (if that is possible).
In this case, if you receive a hosting request and a certain amount of time passes (let’s say a full week) without you logging in or checking the request, the time period until deactivation will be shortened. With each request you receive, the time period will be further shortened until you hit a max number (let’s say 10) after which you will be deactivated so that other surfers are not sending you requests.
For example: let’s say 1 year is the time span needed of inactivity to normally be deactivated. It’s been a month since your last login, so you have 11 months left. Someone sends you a couch request, but a full week passes and you did not log in still. Your amount of time until deactivation then gets reduced by a month, so you now only have ten months (minus a week) to log in or you will be deactivated.
Then let’s say there are a lot of travelers coming through and you are receiving a lot of requests. It’s now been 3 months since you last logged in and you receive your tenth coucher request of that time period. A full week (or whatever amount of time is determined best) passes without you logging in, so your account is immediately deactivated until you log in again so that so many travelers will stop messaging your inactive account.
The actual factors, like amount of time needing to pass after a request before any effect takes place, how many requests you can receive before immediate deactivation (if at all), and how much your time before deactivation gets shortened after each request, would be, of course, all up for debate.