Here is a chart of official fanclub member numbers (Naver/Daum, with a few exceptions, 2nd column) and sales of a single album from those artists in 2012 (3rd column) with the exception of Cho Yong Pil's Hello and SNSD's I Got a Boy which were released this year. Artists like Super Junior, SHINee, SNSD, Busker Busker and Cho Yong Pil have higher album sales than fan club members with BIGBANG also fairly close. The extra sales can easily come from overseas fans who cannot be part of official fanclubs on sites like Naver and individuals switching fanclubs to support a comeback causing fluctuations.
The real question is those artists who have a much bigger fanclub than sales. If an individual is dedicated enough to pick being in an artist's fanclub, why wouldn't they buy the album? Infinite, Xia Junsu, B2ST and DBSK all have significantly bigger fanclubs than physical sales. The simple answer might be that many fans chose to buy digitally or perhaps even did not like the specific album and did not buy it. Others have noted that some fans wait to buy repacks if the artist is known for doing them meaning a portion of the fanbase did not buy the original album but did buy the repack, but with sales also coming from overseas and fans not in the specific fanclub, the discrepancy may be even higher.
(Super Junior, BIGBANG, DBSK, SHINee, B2ST, Infnite, Xiz Junsu, Busker Busker, Cho Yong Pil)
This chart shows the same information in bar graph form with red representing album sales and green the number of fanclub members. There may be other explanations but it is certainly a very interesting look at changing sales patterns and public recognition. Perhaps if digital whole album sales were taken into account we might see that many fans simply prefer to buy digitally or that some artists have fans who are older and do not join online fanclubs but buy albums nonetheless. What do you think these numbers mean?
source of graphs: http://www.gaonchart.co.kr/main/section/special/opinion/view.gaon?idx=68
*What is written in this article is not a direct translation of Gaon's article but an assessment of the graphs.