If your service is compelling and your cost per customer is low enough, then I'd go with #1. If you can't convert the users to paying customers then you probabl

If your service is compelling and your cost per customer is low enough, then I'd go with #1. If you can't convert the users to paying customers then you probably need to revisit your business model. Assuming your service is useful--and the visitor decides they need/want it--you need to factor in their switching cost. If they have some time invested in configuring and using their account, they will be less likely to move to a competitor. If your service is prone to throwaway accounts, this clearly isn't a factor. Recently signed up for a 30 day trial with Pingdom and converted to a paying account ($49/month or so). Not sure what their incremental cost per customer would be (middle of the road I'd say) but they provide a useful service and once I got all my servers set up and had a significant history established, I didn't want to bother evaluating/switching to a competitor.