The Magic Behind Amazon's 2.7 Billion Dollar Question

Communities often often struggle to define the meaning of a user's vote.

Amazon is wise to make a separation between "I agree with the user's opinion" (via my own review) and "I found the user's opinion useful" (via my vote).

Only one paragraph ties reviews to revenue, and I'm still looking for the evidence in here:

"As we’ve watched Amazon customers make purchases on the site, we can clearly see that promoting the most helpful reviews has increased sales in these categories by 20%.(One out of every five customers decides to complete the purchase because of the strength of the reviews.) From this, we can project it has contributed to Amazon's top line by $2.7 billion."

Maybe I'm just missing something... Is there a metric for how many users base their decision to buy on the helpfulness of the review?

Even at that, we have no idea how many of those who bought an item because of the reviews would have bought it without.

I agree with the article that reviews are useful, and I definitely appreciate the community-upvote functionality. It's pretty neat. But I think the case is greatly overstated, and their estimates are pretty dodgy.