Confused by the amateur hour at the ISM data manipulation office? Then this table summarizing the first May data release, the revised May data and the revised revised May data (which still missed expectations of a 55.5 increase, printing at 55.4 instead), should explain it all.
[2]
Remember: when manipulating data for "seasonal adjustments" you may at least want to fudge all of it, not just those cherry-picked components that help you goalseek a number that does not anger the gods of centrally-planned stock markets.