Interesting little aside in this post, where he talks about hitting a license threshold on the number of projects and this triggering an evaluation of whether they really like Basecamp enough to pay more for it. I'm wondering if this would have happened if the pricing model was a much smaller increase per project. If adding a project is just a small additional cost there is no "threshold" you hit where you have to jump to a noticeably more expensive license.
Congrats on finding a way to move off Basecamp. Despite the 37 Signals halo- Basecamp was, in my opinion, a pretty crummy user experience from the start.
I have to admit Basecamp is really underfeatured, no matter how they want to sell it. But the cost having my own hardware installing and managing it, even if its just 1-2 hours of my time is more expense then a year of Basecamp or one its competitors as a hosted. service.
I absolutely understand the principal there, but the reality is that it took me less than 30 minutes to be fully up and running with aC, using hardware that was already online running our subversion server. A light bit of tinkering for a Sunday evening - I don't anticipate any further ongoing investment in maintaining it. It's just PHP, so it should run and run.
That said, this wasn't directly a cost-saving exercise for us. That's just a nice by-product.
Yeah, only 30 mins. And just pray that you wouldn't need backup, that you did not set up. Or that nobody breaks in through hole that you did not patch in php, web server and whatever else you had to install.
Self-hosted solutions requires maintenance even if you don't anticipate it. It's not "just PHP". It feels like a zero cost (you have got the spare capacity server after all), but only until you hit a problem.
ActiveCollab has built-in daily backup to dump everything to disk in a recoverable form, and then the server itself is backed up. If I wanted to go crazy I could spend another minute adding the files to the cron job that syncs data up to S3.
Of course there's a chance that it'll involve a bit of maintenance at some point, and that'll soak up a bit of my time. My experience with hosted solutions is that they have far more frequent hosting issues resulting in lack of access to data (be that scaling problems, being offline for maintenance, bad deploys etc) which soak up lots of time for everyone on the project.
When I'm in control of the situation I know that at least in a bad case scenario I can spin up another server, grab the last good backup and be up and running again within the hour. That has a lot of value.
