We switched from a 30 day to a 14 day (2 weeks) trial period. We found out that if the user does not use the service within the first two weeks, he is not very likely to use it in week 3 or 4 as well. When we've split-tested sign-ups we did not find a significant difference. But the conversion rate tester -> paying customer increased. We're at a 3:1 rate now (B2B, pricing starts at 19.90, open end), no free plan offered.

The trial period is accompanied by a follow-up email sequence (day 0, 1, 3, 7, 11, 14 after sign-up) to ask how's it all going and it includes a lot of supporting material (tips, videos, links to support pages, offer to schedule a call or screen sharing session) to get interacting with them and to make sure they really ask if there are questions.

The user can unsubscribe from this sequence (unsubscribe link in every email), but nobody really ever does. We get the occasional (maybe 1 out of 1000 testers or less) complaint about how annoying the follow up sequence is. We just remove those from the list. I remember 1 guy who was really annoyed but became a user anyway :-) So usually the testers really appreciate this information because they get the feeling there's someone who cares.

In all the follow up emails we make it clear that it would be easy to extend the trial period, a short email asking about it would be enough. We had one guy in a trial period for over 160 days. He wanted to use our product but his project did not take off as fast as expected and we just kept waiting (no follow ups). He's been a user for years now.

So if I was to start a new project, I'd start with 14 days and run the occasional split test from time to time.