almost 8 years on 2009-06-21


- i want to know if people actually finish this book
- on what page/chapter does the average reader generally  abandon (segmented by sex, edu. etc) ?
- are there sections that people read faster/slower?
- how quickly do they finish the book (pages/min)?  (is  it a 'page turner')
- what time of the day do most people read this book?
- how many days does it take the average person to finish  this book?
- did my roommate actually read the book he keeps  quoting?

the kindle has all of this data - and it is the only  platform that has this data - amazon is clearly  collecting it centrally (whisper sync to my iphone proves  they have it).  so, while all i want for christmas is  access to a amazon kindle api, i suspect that amazon  recognizes how priceless the data is (value of data is  based on scarcity and they are the only ones who can  capture this stuff) and will not be giving it away any  time soon.

that said, if i had access to a kindle api (which is  really just one huge but simple reading dataset -- user,  book, page -- with time stamps on every action, page  flip, open/close, etc.) here is what i would do:

1.  make the purchase page/rankings/information about  books data driven:

this is the most benign and probably the first thing that  will happen.  instead of having a four star rating on a  book, when i am purchasing i want to know the rate at  which a book has actually been finished and the speed at  which people tend to read it segmented by age, gender,  and probably some education and social stats that would  require a mashup to facebook connect.  i don't want to  read a book that everyone clearly gets bored with and  abandons on chapter 3.

while we are at it, it is worth customizing the stats to  my own personal book reading history to pull out how  likely i am to finish the book, my expected pace ppm  (pages per minute), where i will slow down, etc.

finally, i would love to be able to search amazon by  saying, i have a four hour flight, can you give me  something that i can read in four hours and will likely  finish and enjoy?

2.  give authors tools for understanding & revising their  work:

further out, but not necessarily requiring any more data  than the kindle api would easily have, why not give the  data back to the authors and help them refine their  works.  if a book is looking people on chapter 7 as noted  by a high abandon rate and slowing pages per minute of  the average reader (or the target of 28 year old females  who like ponies and tattoos), why not go back and revise?

just like google analytics, the kindle api could give  authors powerful tools to analyze how people are reading  their books based on some easy segmentation.  since we  are talking about digital ink and not physical releases,  why not let the authors edit the section and push up a  new edition.... let's move away from book releases to  baas (books as a service)

3.  give teachers/educators a dashboard to watch their  student progress:

this one the kids will hate... again, just use the kindle  api and a few basic stats and teacher can track the  progress of students through a book.  are there sections  where the kids are getting hung up as their ppm drops  significantly (either because it is too boring or too  difficult).

is timmy clearly not doing his homework (either the books  are going unfinished, or is there a pattern of very low  ppm followed by bursts of very high ppm -- while he is  just hitting 'page next' while watching tv)?

how is timmy progressing, is his speed measured by kindle  and comprehension (probably requires tests) increasing  throughout the year?

how is the class performing against averages of students  their age?  are they reading more quickly or less  quickly. 

4.  not to mention the pure research of it all:

how do people really read, when, and how.  how does  reading evolve with age and along other segmentations.   how is reading changing over time.  what percentage of  books purchased actually get read.  what percentage of  books that are started are finished, and where are the  common abandon points.  these are all fascinating  research projects, all they need is the dataset that  amazon has. 

ultimately, this is not new thinking - but newly possible

as a highly dyslexic person who didn't learn to really  read until about 6th grade (and faked it before that) i  find this new transparency a very complex concept  socially.  as someone who loves books and believes lines  about the sanctity of literature, i am not sure that  living wikiesque books is a good idea...

but, 1.  this is all going to happen whether we like it  or not, of this i am relatively sure.  2.  the 'data  nerd' in me loves every second of it.

so, amazon, do what is right and make a whole new  industry and set of problems for us - release a kindle  api and let me mash it up with facebook, build interfaces  for teachers and authors, and generally completely  overhaul the concept of what a book is and what reading  one means.

original swl blogposts and letters 2007-2010