It's very difficult to value knowledge. It's hard to know what a piece of knowledge is worth. It's much easier to see the value lost when knowledge is missing.
When conducting end-of-project Retrospects, it can be illuminating to ask "what would it have been worth to have had this knowledge at the start of the project?" Here are some real example answers from past Retrospects
“The 8 months hiatus may have cost $30 million”
“The documentation issue may have cost the project about $0.5 million”.
"The impact was a delay in the order of 3 months".
In none of these cases did the knowledge deliver a saving; instead a lack of knowledge incurred a cost or a delay. Nor can we assume that KM will necessarily deliver that value in future. However what these statements do, is start to put a monetary or time value on knowledge, by showing the value lost when knowledge is missing. It also starts to show the size of the prize which KM might make possible, by ensuring critical knowledge is available at the start of projects.