You can be a rather successful community manager or conversation manager without analyzing data.
That said, I believe that analyzing data is the next big step. An analysis of social data will help you during conversations with management, resource distribution, it will allow you to become more efficient and correct slight mistakes. Overall it will make your job more fun. (at least if you like data)
First of all, make sure you measure and keep the data. It is a real waste if you just let that pass you by. It seems like a no-brainer. It also tends to be not the case in a majority of companies that do community management.
One of the obvious uses of this data is your work schedule. It doesn’t pay to be at the job at 9 if there is almost never any real conversation going on at that point. It doesn’t pay to be offline at 7 pm if that is a time when there is a ton of it. You will most likely discover that there are certain times during the day that are responsible for a majority of the interaction. The 20/80 rule (parkinson’s law) can be valid here, meaning that 20% of your time is when 80% of the interaction happens. During those times, you need to be on top of things. Uninterrupted and without any scheduled activities. Your second line needs to know this too. Technical experts and helpdesk customer care agents need to know that during this time, helping you is priority #1.
Of course, this will also tell you what times tend to be calm. This is a very valuable thing to know. If during this time there is an outburst of activity, that usually means that something is wrong. (or very right) It also means that you know what the perfect time is to do what we do best: individual conversation. This “downtime” is when you have the time and the peace of mind to really engage with people. Why not organize a chat and announce it trough social media? Why not have a live videostream with people asking questions in a chat? (you can use tinychat.com and ustream.tv for instance).
With this in mind, you can imagine how bad poor planning can be. And you can’t plan without data.