I feel strongly that because of the very different approach to customer engagement that is being brought on by the social web, marketing and customer support teams should think seriously about creating social labs whose mission is to experiment and try new techniques and processes - and give them the funding necessarily so that they have the tools and resources to adequately test their hypothesis.
Why? Social processes cannot easily be bolted on to existing processes. David Alston of Radian6 who is one of the new generation of social CMOs recently explained his approach as "I focus on community first and then layer on more traditional marketing programs, not the other way around". That is a fundamental shift.
The other issue with this shift - which companies are currently struggling with - is that CxOs want and need to see these new approaches replace the volume they currently get with traditional methods. The problem? Current methods of marketing investment are linear - you spend X in direct response marketing and you get y in return... and there are fairly robust benchmarks to indicate whether you are on target or not. It also scales up and down predictably - if I spend twice as much I get twice the response rate. Communities look more like hockey sticks. Initially they require a lot of investment - and have disproportionately low returns. From a measurement perspective, they look like failures for a long time before they look like successes. Over time however, communities see geometric returns that you simply can't achieve from traditional methods and that is when the costs start dropping dramatically relative to the returns. But that gestation period is different for different types of communities and it requires an act of faith to invest heavily is something for a long enough period to see those returns start to happen. In mature communities, membership doubles in increasingly small increments - SAP's SDN & PBX communities have over 1.5 million members, a year ago it was 1 million... and that was three or more years in.
Which gets me to social labs. A marketing or support organization cannot switch overnight - there would be a huge drop in performance while the community is in it's incubation stage... a stage that requires a limit to the number of community members in order to consolidate and build a core of passionate advocates which will enable the hyper-growth stage later. Today, too many groups are trying to either add on 'social' to existing processes or expect the same performance out of social initiatives - often trying to skip the critical incubation phase of community development ensuring they will never reach the performance needed to replace traditional practices. So, give your social initiatives an incubator - and don't start judging right away - it doesn't work the same way that you might expect it to.