Where do you start with KM? Where do you focus your efforts? Especially if you want to demonstrate quick wins?
We can look at this issue through the lens of Knowledge Push v Knowledge Pull.
Knowledge pull is where you start with a problem, and share knowledge to solve the problem. Pull start with someone who is interested in acquiring knowledge. It starts with Demand. Pull is also easy to address, through Peer Assist (the closest thing to a KM silver bullet). The knowledge shared through the Peer Assist will find a willing audience, instant application and short-term payback. There should be little or no "Not Invented Here".
Knowledge Push is where you start with a solution, and share knowledge of that solution in case anyone does anything similar. It starts with Supply. Imagine you conduct some Retrospects and document some best practices on mergers, or on outsourcing, or on implementing ISO. It may be a long time before another merger, or another outsourcing, or another ISO implementation. Maybe nobody is ready to adopt these Best Practices right now. And even if they are, there is the Not Invented Here barrier to deal with. The audience may be unwilling, and the payback will be delayed. However in the longer term, you will progressively build a better and better process for mergers, for outsourcing or for implementing ISO.
My recommendation is always to start KM with Pull, if you want the quick wins, and bring in Push later, for the longer term benefits. Pull reaps instant benefit, but maybe only once. It solves an instant problem, but leaves no trace. Push reaps benefits over the long term. Capture knowledge once, re-use it twenty times.