I like to think small but significant.
Any decrease worldwide would be significant. But I guess if that happens the numbers will be spun further. They already use Peak publishers and have dropped "average publishers". I believe they have done this only because the "average publishers" looks worse - possibly already in decline?
When the "Peak publishers" declines then that statistic may well disappear.
So either a published decline = genuine decline.
No figures published = genuine decline
Spun figures = likely genuine decline "recounted" to make a tiny increase.
Watch out for "love of the greater number cooling off" being trumpeted as a "good thing".