Category: books

For the past year or so I’ve been using Bookmooch, a great free book swapping service. After signing up, you list books you are willing to give away, gaining 0.1 points per book.  Once you have a full point, you can spend that point by requesting a book from another member.  If you send a book to someone else, you gain a point.  You also earn 0.1 points by posting feedback about the swapping experience.  Finally, if you send a book to a foreign country, you earn three points (the requester spends two points).

I love the service, but lately it is getting harder to find books to spend my points on.  I have a feeling it’s related to the points system, but I’m not sure how to best demonstrate that.  Over the long term, the points system doesn’t seem to be sustainable because of inflation.  For example, the transaction nets +1.1 points for the giver and -0.9 points for the receiver, for a net gain of 0.2 points; this gain is due to the 0.1 for adding the book and the 0.1 for posting feedback.  The foreign country book exchange results in even more inflation, +1.2 for the transaction.

I decided to look through some of the publicly available stats to see if there were any trends to back up my hunch.  There were a few metrics I found interesting: total number of points in the system, total # of books saved in wishlists, and total # of books available in the system (inventory).  I normalized all these over the total number of users and graphed the result below (x-axis is time from 8/18/06 to today).

What does this imply?

• The number of books in wishlists is growing while inventory is falling.  With the price of a book fixed at one point, I think this is creating a supply / demand imbalance.
• Note that the inventory curve started positive and seemed to outpace the wishlist curve, but then the inventory peaked and started declining.  Part of this may be explained by users signing up to check out Bookmooch, but not adding any inventory.  I think this relationship also shows that point inflation is taking effect; users are finding themselves with more points and are slowly stripping away inventory in the system.
• All the while, points per user is continuing to climb.

What can we expect if we fast-forward a couple of years?  Existing users will have a large number of points to spend, dozens (hundreds?) of books in their wishlist, but very little inventory to meet that demand.  When a new user joins and posts their valuable inventory, existing users will snap it up very quickly, thanks to wishlist notifications.  That new user will find themselves flush with points but very little to spend those points on.  They proceed to add dozens of books to the wishlist.  The cycle continues.

I don’t even know if the system truly is broken; an economist could do a much better job of analyzing the data and would probably prove me wrong on many points.  But, I have a few ideas that may improve the points system.

• Eliminate the 0.1 points for adding a new book.  Either require new users to give a book before they can receive one, or give them one point for free after signing up to encourage future trading.  Or, give the 0.1 point for listing the book and the remaining 0.9 points when the book is mooched.
• Likewise, remove the 0.1 points for providing feedback.  Users will want to give feedback for free in order to clear out their “Waiting to Receive” listing.  This combined with the first idea will help eliminate point inflation.
• Continue to reward users for sending books internationally, but give only two points instead of three.  For the receiver, continue to deduct two points; I’d consider raising this to three points to create some downward pressure on # of total points.
• For wishlist books, introduce the concept of advanced notification for the price of one point.  For example, two points will give you the opportunity to view and mooch a wishlist book 24 hours before other users who are willing to spend only one point.  As the transaction happens, make that extra point vanish rather than transferring it to the giver.  I think reducing the total number of points will help encourage existing users to list more books.
• To create a true market without an artificially set price of one point per book, allow users to enter bids for a book.  Some of the extra points could be transferred to the giver, but bleed off at least one point per transaction.  This type of system would make it nearly impossible for new users to obtain in-demand books, so I don’t think this would match with the founder’s goals for Bookmooch.

My gut tells me there are too many points in the system, with too many users chasing too few books.  Removing point inflation and introducing some mechanisms for burning up points may help, but there may be some long term effects that I’m not considering.

As I said before, I love the site and these are just some thoughts about the underlying market dynamics and not criticisms by any means.  I’m really interesting if anyone else out there has thought about these issues and would care to refine / correct some of my thoughts.

Update – 9/23/08: As expected, there have been a few discussions on the Bookmooch site about point inflation.  This post recognizes that inflation exists in the system and the end result may be that “everyone will eventually end up with a bunch of points they can’t use.”  And, this comment captures the inflation issue as well; lots of other good ideas in that thread, including limiting the 0.1 point gain for the first 20 books added.

Update – 9/24/08:  John Buckman from Bookmooch posted a comment with some good insights into the system.  Based on the feedback in his first point, I’ve updated the chart to normalize all values over total users who have listed books to give, whereas previously the values were over all total users.  This method should remove users with inactive accounts, but will ignore those users who are active but have no inventory.  Still not perfect, but maybe slightly improved.  The total # of points line (blue) is still skewed higher because points of inactive users are still being counted.

Looks like the biggest change is with the inventory / member line, which shows a plateau with gradual decline rather than the peak and drop off in the original graph.

I also added in a new measurement, number of current mooches per 100 users who are listing books to give (green line).  This validates point #4, that mooching activity is on the rise, not just in total, but also on a per user basis and contradicts my hypothesis that users are finding less mooching opportunities out there.  It will be interesting to see if this line ever crosses the 100 mooches / 100 users mark.