Drawing a distinction between delivery metrics and value metrics may not be helpful, given that Scrum actually focuses on delivering value.
What matters is being able to close the validated learning loop by releasing that value under market conditions. There aren't really any good or bad metrics per se. Rather, the measures we use can be either "actionable metrics" (good) or "vanity metrics" (bad) depending on context. As the article says, a success factor or hypothesis should be decided at the start of an initiative. That frames the context. Metrics should test the hypothesis as rigorously as possible so the smartest actions can be taken.