[dageroth]

I'd find the number intersting how many users would pay, if there were no free option available - my take is that it would be much more than the one percent of users that seem to be the rule in the freemium model. Of course a free plan attracts more potential users in the first place, which means the cake is smaller, so it's hard to evaluate what yields more paying users...

The more important factor to me, though, is cost and that usually not only involves storing some data as in the case of historious, but also support requests and scaling costs - when you have a paying user base scaling just won't be that much of pain point.

So, I'll go with a free 30 Day trial period after which the user has to decide, whether he'd like to sign up and pay. That way I can easily handle a couple of thousand users on one server and have more time left to spend on development instead of suporting nonpaying customers.

Free or Freemium only makes sense to me, if the presence of more users provides value to paying users, e.g. okcupid, Skype, and other network-effect businesses.

[DJN]

I'm not a big fan of the 30 day trial because I think it adds extra friction to the signup process. Especially when a credit card is required.

Isn't it better to offer an unlimited, resource constrained plan? Note: you can still let them use all the features but with limited resources.

Advantages: 1. You get permission to send them product related emails. 2. They hit the resource limits quickly thereby encouraging them to upgrade.

We do something similar at Trafficspaces and it works well for us. Unlimited free trial but with very low resource usage.

Just my 2 cents.

[StavrosK]

Our free package is a trial. The 300 free bookmarks either run out quickly, so the user has to select whether to pay or leave, or they last long enough that the user needs no more, at which point we don't mind the cost.

It works out much better than a time-based trial, which I don't think makes very much sense when you can have other factors.