Sunday, 27 December 2015

Girl on the Train - The Sales Bump from Goodreads #1 Readers’ Choice Award

Girl on the Train - The Sales Bump from Goodreads #1 Readers’ Choice Award

Earlier this month (Dec 1, 2015), Goodreads (the social media site for book discussions, reviews, etc) announced the 2015 Goodreads Choice Awards. A couple of books that were chosen by Goodreads members as #1 in their categories happened to be books that I have been following closely, for much of this year. Those are Harper Lee’s “Go Set a Watchman” and Paula Hawkins’ “Girl on the Train”.
In a couple of earlier blogs, I looked at how this affected Harper Lee’s recently discovered (and published) novel “Go Set a Watchman”. In this blog, I will look at what has happened to the other book in question, The Girl on the Train. I have been following the ups and downs of this book on Amazon for most of the year. It’s been interesting to watch. At times it has started to fall off of its high perch, only to get new momentum and move back up the charts. It looks like the Goodreads award, and the approach of Christmas sales has helped it out, once again.

One additional feature of this analysis is the sub-analysis of sales and reviews in the various national markets in the Amazon English speaking world - the U.S., the U.K., Australia and Canada.
As you can see in the graph, there appears to be a fairly decent rankings improvement in U.S. sales, at about the same time as the Goodreads award (Dec 1), which is indicated by the different coloured point. The book was losing rank, dropping below 50th spot, until about that point, then it picked up again, to the 20 to 30 rank.

A similar effect seems to take place in the U.K. sales, though perhaps not so pronounced. However, the book had never fallen much past the mid-20’s there, so there was less room for improvement.

Sales in Australia were down, by mid-November, then picked up after that, by a considerable amount. It is hard to pin that on the Goodreads award, as the improvement began before December 1.

Like the Australian market, the Canadian market underwent some gyrations well before the Goodreads Award in early December. One could argue that there was some pickup after the award, but the case is not exactly cut and dried.

The evidence from reviews is marginal, at best. In the overall graph, we can see a substantial bump in the number of reviews early in December. However, as the graph shows, these spikes seem to be a regular feature of the Amazon review system, so it is debatable whether we can attribute that spike in reviews to the Goodreads award.

Looking at the U.S. and U.K. results separately, we see much the same pattern. There is a jump in reviews at the beginning of December, but these jumps are happen a lot over the course of the year, so one can’t make too much of this particular increase.

When it comes to the two smaller markets, the situation is even more problematic. There simply aren’t enough reviews to even notice an increase from the Goodreads award. 

The presence of data for all four markets does allow us to make a few other observations of general interest. One relates to the size of the markets. In terms of reviews, the figures as of Dec 22, 2015 for Girl on the Train were:
  • U.S. = 28,204 reviews. The U.S. population is about 322.5 million.
  • U.K. = 8,950 reviews. The U.K. population is about 64.8 million.
  • Australia = 652 reviews. Australia’s population is about 24 million.
  • Canada = 418 reviews. Canada’s population is about 36 million.
The U.K. had 32% as many reviews as the U.S., though the population is only 20% of the U.S.. The fact that the author is British and the book is set in London probably explains this over-representation of reviews (and sales, given that sales should correlate with reviews). Canada had about 2% as many reviews as the U.S., though it’s population is about 11% the size. This may be due to the fairly significant proportion of Canadian readers who are with Kobo or buy through the U.S. Amazon store. Australia had about 2% as many reviews as the U.S., though it’s population is about 7% the size. This may reflect a slower adoption of e-reading in Australia, but that’s just a guess.

I also had a look at the correlation coefficients in rank, between the countries. That’s just a way of seeing how much two variables are related – a value of 1 implies a perfect positive correlation, 0 is no correlation, and -1 is perfect negative correlation. So, a high positive number means that when the book’s rank went up in one country, it also went up in the other country. The results were:
  • U.S. and U.K. correlation was about 0.83.
  • U.S. and Canadian correlation was about 0.82.
  • U.S. and Australian correlation was about 0.83.
So, as the book’s rank changed in the U.S. store, it tended to change at a similar rate and in a similar direction in the other stores. In other words, “Girl on a Train” was popular throughout the English speaking world, and its ebbs and flows in popularity were similar throughout that language world.
The other correlations (e.g. between Canada and the U.K.) tended to be in the .60 to .65 range. A path analysis type solution might suggest that these correlations were mediated by the U.S. (CanU.S. correlation=.8 times U.S.  U.K. correlation=.8 resulting in CanU.K. correlation=.64), but that’s just a conjecture.

After all these stats, you might want to read something less quantitative, but still featuring interesting non-fiction. So, try a road trip through North America in an 18 wheeler, with “On the Road with Bronco Billy”.  You will ave a chance to pick up some trucking jargon, some trucking workplace culture, some geography, and a lot of penetrating  sociological observations :) :

No comments:

Post a Comment