My team tried it several times over the first 1 - 2 years it was out. I have not tried it recently. That said, at that time, we could never get it to stick. It was good for high level stuff, but like all project management software, we ended up constantly working to keep it in sync with reality. This was especially acute when using github because for any stuff that was code-related, we were basically duplicating tickets and tracking in both places. So github tickets were the real source of truth and Trello always lagged. Once that happens, the non-code milestones start to suffer as you start wondering why you're doing all this in the first place.
Tracking issues in github (and we use FogBugz as well) has the benefit of being where the "rubber meets the road", so it's effortless and transparent to keep up to date. Merging non-code tasks (e.g. design work) with code is still something we struggle to track and keep tabs on.
I definitely gave this some thought, I think it's not so straightforward though. There is a bit of an impedance mismatch between the things that a bug tracking system emphasizes and captures and the things a project management system captures. Simply porting bugs over isn't a clean solution and can lead to confusion among non-developers when the bugs are highly technical. Github makes it especially hard because there is no support for dependencies so you can't roll up a bunch of individual technical tasks under one umbrella without lots of extra effort.
To be clear: while I can describe the problem, I honestly don't have any ideas what the most elegant or clean solution is. I have never found a satisfactory solution. I end up keeping lots of stuff in my head and writing summary emails.
We implemented Trello + Github sync with Waffle.io and it was as close to perfect we could get without spending a bunch of time writing our own code to do it. It still wasn't perfect however and I spent a lot of time making sure things were in-sync, which on a normal schedule wasn't a huge deal, but when fires popped up would become a very small priority.
I'm surprised no one has built a better way to do this. We tried syncing with Zapier first (a company I absolutely love), but if I recall correctly it was a one-way sync, and we needed it to go both ways.
Yes, we (at https://codeable.io) use Trello for pretty much all internal (async) communications, from Sprints (user stories) to features, to ideas,...
No learning curve, exactly the right features.
And also there's something I like in Trello that Basecamp (I know comparing apples to oranges, but both can be used for similar purposes) doesn't support: closure. When something is no longer relevant, you archive it and it's gone, no more mental burden. Quite important when you have 100s of tickets.
My research lab uses Trello to coordinate on protocols, results, and publications. Each project gets a board, with lists for each step in the process. Cards are used to describe what was done in the processes like a typical lab notebook. Add on comments and file attachments and we can collaborate remotely.
The itch I needed scratched was that Trello lists sometimes disintegrate into junk-drawers. Their API let me build something to pull out a list, let the team rank the cards and then reimport it into Trello. Yay for killer APIs! http://blog.forcerank.it/tame-your-trello-backlog
I've used it when needed (i.e. someone picks that to get started), but honestly Pivotal Tracker is the same thing with more features and even crusty old JIRA has an agile board view option it nowadays where you can drag card versions of the JIRAs around. So Trello is kind of an also-ran that doesn't compare well with the current leader or the industry standard.
I use it for personal project managing but I knew it was popular when my girlfriend's sister was showing it to her Dad. She does R&D (chemistry) at a water proofing sealant company. The company has 30 employees and they use it to track projects. I need to get out of my software bubble and realize that a lot of tools can be applied to other industries.
Yes. It works well for us because you build your process on it rather than having a process dictated to you. A lot of the "agile" project management tools are too opinionated about how you should work, and hence are difficult to use unless your process fits exactly.