The folks running the SIAMPI Project could perhaps use this blog as evidence of what they are looking for to indicate impact: behavior change. @Stephen_Curry describes his attempt to negotiate Open Access for a review he was invited to write — and how his opening of these negotiations was different from what he would have done before the recent flap about the Research Works Act.
Now, if we were going to document impact in this case, the question would not be what the impact was or who was impacted. Instead, the question is who had the impact (who was the causal agent?). I wonder whether an Aristotelean account of the Four Causes might prove useful in documenting impact?