I completely agree about the risk of it being dismissed as a data hygiene issue. Making the metric actionable is the key.
Your point about source and age is critical. In one implementation I analyzed, we included *value* at creation as a third dimension - a lead scored as Marketing Qualified and assigned a high lead score that gets junked is a much more serious signal than a low-scoring, unqualified lead from a purchased list. It changed the conversation from "why is your junk rate high?" to "why are you dismissing this specific type of high-potential lead?"
The dashboard shift from blame to diagnosis only works if you give managers the levers. If the data shows a sourcing problem, the next click should be a filter for "leads from source X purchased in Q3" to take to the marketing team. Otherwise it's just a prettier report.
No free lunch in cloud.