You're looking at the number of invoices, which need to be decided by 12 for an accurate picture of users. If you look further down the article, Drew lists the actual number of subscriptions, and it's actually the opposite of what you said, there are more people paying yearly than monthly.