Maybe some people thought that holding the Worldcon in Japan would result in tens of thousands of people showing up? There were around 3,000 bodies on site, of which somewhat under 1,000 were gaijin like me. To me, that is really quite good, considering that there has never been a Worldcon there.
I am one of a relatively small number of people who have been studying Worldcon attendance figures over time. One thing I've concluded is that it's very dangerous to try and draw conclusions from small data sets. No one year will tell you anything meaningful. There's too much variability in the data. Worldcons are sometimes held in places that are relatively inaccessible like Australia, or that have relatively poor air connections to the rest of North America, like Winnipeg. Sometimes they are held in areas with very large "day tripper" locations, like Southern California, the Bay Area, and Boston. So don't look at any given year and say, "We're doomed!"
Another thing to consider: You can't really compare a traveling convention like Worldcon to stationary, professionally or semi-professionally run ones like ComicCon or DragonCon. The very structure of Worldcon, which is a gigantic annual one-shot convention run completely by amateurs with no permanent staff and no ability to build up local goodwill, means that the current size of the event (roughly 5500 +/- 1500 in North American and Europe, lower elsewhere) is probably the limit to growth. To get significantly larger, we'd have to fundamentally alter the convention's management structure.
I continue to wonder why people who I think should know better keep predicting the imminent death of fandom.