The way I like to look at my social CRM options is in terms of a 2x2 grid of real-time & community based conversation axes. It is all about when does a customer get a response & from whom. The medium is immaterial, could be an archaic forum or nouveau twitter. If you get your query answered within minutes its real-time. If you get it answered by someone not with the business, it is community based.
Community based conversations are the new outside-to-outside conversations that are happening and alarming the businesses. These cannot be controlled, only be influenced and that when the business has engaged itself in the conversations & built relationships with the community on trust & transparency. Look for help in this form of conversation from folks like Jive, Lithium, SocialText, etc. We merely need to integrate these systems with our traditional systems like CRM, DW/BI, etc.
Businesses need to take decisions only for non-community (which actually means a paid representative of the business) based conversations be they real-time (twitter, IRC, etc.) or not (forums, youtube, blogs, etc.).
For non real-time conversations, businesses can take some time to react to the people, thus allowing the business to gather some more data & analyze & review & approve the response(s). Here we can build very many systems using existing DW/BI technologies. Just need an NLP piece in there somewhere.
Real-time conversations call for quick thinking calling for deep experience about the business (organization, culture, financials, etc.) & its offerings. This is human intensive, for the time being, quite expensive & not too scalable. Explicit & clear guidelines are a great way to scale the conversations by roping in more people from the organization (or outsourced) who are not too deeply mired in the organization culture to handle easy conversations & routing it to the superstars for the tougher ones, usually requiring deep insights as well as authority to decide standing-on-their-feet.
Another way to make them (real-time non-community conversations) slightly more scalable is by using realtime sentiment analysis & other NLP techniques. Look at peoplebrowsr or OpenAmplify or the open source twitgraph.
Real-time-community based conversations are something that we cannot decide or control in real-time. These are only stuff that organizations can aim to influence over a period of time by building relationships with the community.
So how does all this responding affect social CRM? I guess we need to look at Mitch's post on social CRM data for more inspiration. :)
Prem,
ReplyDeleteWe are both taking the advice, and now focusing on the 'how do we get this done' versus 'what is it called'. This topic is an important one, thanks for tackling it!
One point worth mentioning, is that my take on people who comment/communicate via Social Media, want a response in a 'public' way. They do want the response in any format, but they want the world to know they received a response. So, if the response is not very public, they will make it public!
The scalability question is also a great topic which will continue. While sentiment analysis is the objective, it is very hard and I am not sure I would trust a computer to do that work for quite a while. As researchers have shown, for emotional topics, the words only represent 10% 'sentiment', the other 90% are tone and body language.
Great stuff my friend!