14/06/09

The Long Tailed Book Seer

So in downtime over the past month or so we made The Bookseer, a fun little web app, and it went live last week. It’s really simple, but we’re delighted to note that it has seen a lot of love and quite a lot of action in its first week.

Seeing as the Bookseer is about books, and data, and openness, I thought I would share some of the early stats with those of you who are interested in such things. This is all based on the first few days’ traffic up to June 13th. (Whilst launched before then, we announced in on June 9th.) As well as being fun, I think that the data is a mild demonstration of The Long Tail in action.

Visitors

• We’ve had 19,716 page views, across 4,403 page titles, with 15,181 unique page views.
• Visits are 7,123 with uniques at 6,602. Which suggests about 10% of people asked the Book Seer for advice on more than one title. (Or, if you felt generous, 90% of people got the answer they were looking for. Personally, I’m not quite that optimistic – if that were true, then we really should have set up an affiliate account on the referrals to Amazon… although I think that violates the non-commercial terms of service on the API.)

Pages

• The home page was the most popular, but only bounced 30.5% of visits. Which seeing as much of the referred traffic was from StumbleUpon, I think is pretty good – they are hard people to entertain. Having said that I find it really hard to see what StumbleUpon has said about a site or even to find a referring link.
• There are only two pages on the site- the home page, and the results page. The results page has the name of the book written into the URL (and the title) so we can tell from the logs what books have been looked at the most. If that URL had included the ISBN or something else abstract – we wouldn’t know anything. Nice work James. As a result we can tell which titles have been the most popular requests of the seer.

Books

And so the next bit is the fun bit. We have 4,406 different page titles, and 3 of those were the home page (we collected data when the site was codenamed “My Next Book”). All the other 4,403 are questions for the seer.

Here’s the top 20 – and it provides a fun overview of people’s reading habits. Note that some – not all – results are skewed by the fact that bloggers linking to the bookseer linked directly to a results page. So Stumbleupon appears to have linked directly to the Dave Eggers, What is the What page, a link which resulted in 166 of the 269 page views for that title. Still, What is the What remains a popular book among people who come to consult the Bookseer – or, looked at another way, one that people in aggregate find particularly hard to follow up to.

Top 20

Note that the links in the list below are ones I have made on the title only. Requests for an author only (e.g. Dave Eggers, without “What is the What”) are met with a demand for better information.

1. The Book Seer | What Is The What- 269 [This is slightly unfair as it is one of the pages linked to by StumbleUpon)
2. The Book Seer | Twilight - 116
3. The Book Seer | Harry Potter 74 [Note - Librarything has a lot of problems with this request]
4. The Book Seer | 1984
5. The Book Seer | The Road 55
6. The Book Seer | Neuromancer 52
7. The Book Seer | Atlas Shrugged 30
8. The Book Seer | Pride And Prejudice 27
9. The Book Seer | The Bible 26 [some great left-field recommendations for that]
10. The Book Seer | Outliers 24
11. The Book Seer | Enders Game 23
12. The Book Seer | On The Road 23 [Note the difficulty Amazon has compared to Librarything when no author is entered]
13. The Book Seer | The Catcher In The Rye 22
14. The Book Seer | ウェブ進化論 22 [No idea, sorry, but from the recommendations, looks pretty cool]
15. The Book Seer | Infinite Jest 21
16. The Book Seer | The Secret History19
17. The Book Seer | The Stand 19
18. The Book Seer | American Gods
19. The Book Seer | The Book Thief 18
20. The Book Seer | Catcher In The Rye 17

Whilst the top 20 is interesting, what I love is that the full list of titles is so broad. The vast majority of books only get one search on them; the bottom of the list is Q E D.

Observations

• I think this is pretty cool, although I don’t know why. Taking aside Neuromancer – James’ test title – is it a really interesting zeitgeist of what people are reading? Or their favourite books? Or the books they find hardest to followup? Or the first book that comes to mind? Twilight, Harry Potter – OK, these may demonstrate a paucity of imagination as much as they show a teenage visitor set. But Infinite Jest?, What is The What? Atlas Shrugged? The Road? Whilst in the news, these aren’t exactly bestsellers. When Russell* linked to Rogue Male, I think the favourites is a good guess.
• Oops. We should probably strip out “The”, “A” and other prefixes to avoid duplicates such as [The] Catcher In The Rye.
• The Bookseer works much more accurately if you put in an author as well as a title, although James does a lot of clever things in the background (stuff learned from projects such as Bkkeepr for a start) to make up the difference.
• LibraryThing handles “title only” requests much better than Amazon. Amazon gets confused by “The Bible”, or rather, the number of titles in its database that include “The Bible”.
• LibraryThing gets confused quite often. Sorry Tim – do get in touch if we can find better ways to do this, or to query Unsuggester items (although part of me thinks the beauty here is in the simplicity).

And finally apologies to anyone using Internet Explorer 6 (and to my horror, this includes about 3.5% of visits and several large UK publishers) – you may have issues. Not sure what that says – we tried really hard to support you.

Visit the Bookseer

* [Russell - I'm with you. I recently re-read Rogue Male and loved it. I also gave my copy to David Simon, fan boy-style, because I read that he is apparently making a film based on Man Hunt, the Fritz Lang movie based on RM. He hadn't heard of the book, and wasn't very impressed by my gift.]

Posted by Peter Collingridge in Apt Studio work, Librarything, Long Tail, Uncategorized, Web.

← Seamus McGarvey, Cinematographer // Apt’s links for June 17th through June 21st →