"The biggest risk is probably that somebody will accidentally publish something they shouldn't, which is unpleasant but not calamitous. A failed enterprise collaboration project, unlike a failed ERP system implementation, isn't going to sink the whole company," says Mann.
"In fact, social software and collaboration projects should be designed precisely so that the organisation can backtrack and change its mind if things don't work out. These are not systems you will be stuck with for a decade. Doing a number of pilots, so long as you manage user expectations and don't cancel a system they've learned to rely on, can be a very healthy approach."
Mann says that of the enterprises he has studied between 10% and 20% are stuck at the stage of still seeing social software as a triviality and a danger. Another 10% to 20% are in the real adoption phase where they are seeing how it can be used to drive transformation and get closer to customers, and the majority are reacting to demand and waiting to see where the real value lies.
Mann stresses that the single biggest key to success, as with any software implementation, is figuring out purpose. "Why are you doing it, and what is the major business benefit or outcome? That's especially important when it comes to introducing the second and third generation of users who aren't moved by how exciting or cool something is.
"You may identify that your document review process is inefficient, which is a problem in many organisations. Conflicting edits and recommendations can lead to a chaotic situation with several competing versions of reality. In this context, using a wiki as a collaborative editing tool can solve a real business problem.
"Once a clear purpose is identified, take an iterative approach. Do pilots, learn, change direction if you have to. It's increasingly a good idea to build that kind of flexibility into all your projects, so this is a great opportunity to practise. Put the risks into perspective, but don't use them as a shield or an excuse to do nothing," Mann concludes.
