That article is using a horrible methodology. The reason the average tenure is low is because Amazon is growing at around 60% year-over-year.

At that growth rate, you'd expect the tenure numbers to be dominated by hiring rather than turnover, a point the article misses completely.

(Per their annual reports, Amazon had 88,400 employees on 12/31/2012, 56,200 on 12/31/2011, and 33,700 on 12/31/2010)