Amazon’s numbers get fuzzier?

Amazon’s numbers get fuzzier?

I love Amazon for a lot of reasons.

However, I’ve always wished they would be more forthcoming with numbers…for example, sales figures. I get that data is super valuable now, and that hoarding it can make some competitive sense.

It’s also possible, I suppose, that they’ve either decided to reduce resources on producing numbers, or discovered that they were inaccurate. Inaccurate numbers can be worse than no numbers. I understand that.

Here’s a number which is fuzzier this month than last.

For many years, I’ve run numbers on the first of the month, and shared them with you in my

Snapshots

There have been 106 posts in that category.

I was running those today, but getting a very different result (using the same technique).

For example, for total titles in the USA Kindle store, I’ve gotten these numbers in some recent past months:

January 1, 2018: 5,942,033
December 1, 2017: 5,856,598
November 1, 2017: 5,758,762
October 1, 2017: 5,608,214
September 1, 2017: 5,526,342
August 1, 2017: 5,454,920

Those precise numbers allow me to graph growth, and using that as a top level number, do a number of analyses (perception of titles which fit certain categories, for example).

Today, all it will tell me is that it is “over 6,000,000” titles.

How much over? Is it 6,000,001? Is it seventy kajilion? No way to tell, so I can’t do my analyses (I hear some of you saying, “Whew!”). 😉

I ran it in three different browsers (Chrome, Edge, Maxthon) and in different ways.

So, I chatted with Amazon support.

This is part of our conversation:

Hi Bufo.
Me: Hi!
Sa—: If your search gives a high volume of results, it will just give an arpproximation .
However, may I please know if you are looking for something in particular.
The keywords you use for what you are looking , if are specific the result would be more accurate.
Me: I guess that’s a recent change. I’ve been checking a number of categories in the USA Kindle store on the first of the month for several years (I’m a blogger). In the past, I’ve always gotten a specific number. For example, on January 1, 2018, the number I got for titles in the USA Kindle store was 5,942,033. Today, it just says, “over 6,000,000”, so I can’t do a very good comparison.
In this case, I use no keywords. I use the Advanced Search in the Book section, and just specify the Kindle format.
Sa—: Okay, Thank you for the insight.
I understand now.
Me: Newsstand has given me a precise number of titles: on January 1st it was 2,270. Today, it just says, “over 2,000”.
Sa—: Yes, our amazon web team keeps working on site changes and this must be a new change
Besides, the kindle books on the site, are being added and removed must frequently , so a precise number would be hard to account .
I will however take this as a feedback and forward it to our internal teamm.
team*
most*
Me: Thanks for confirming! I did run a much narrower search, by searching for “Doc Savage” in the Kindle store, and got a precise number (258). It doesn’t seem like the process of accounting would have gotten more difficult between last month and this, but perhaps the publishing process has changed complicating that. I appreciate your response.


I don’t quite know what I’m going to do with this yet…I’ll let you know later.

If Amazon has done this deliberately (and it could still be a glitch…something like this happened once before), what do you think? Is it okay for Amazon to do? Does it suggest any trends? Would you miss the monthly Snapshot? Feel free to let me and my readers know by commenting on this post.


You can be part of my next book, Because of the Kindle!


Join thousands of readers and try the free ILMK magazine at Flipboard!

All aboard The Measured Circle’s Geek Time Trip at The History Project!

* I am linking to the same thing at the regular Amazon site, and at AmazonSmile. When you shop at AmazonSmile, half a percent of your purchase price on eligible items goes to a non-profit you choose. It will feel just like shopping at Amazon: you’ll be using your same account. The one thing for you that is different is that you pick a non-profit the first time you go (which you can change whenever you want)…and the good feeling you’ll get. :) Shop ’til you help! :) 

This post by Bufo Calvin originally appeared in the I Love My Kindle blog. To support this or other blogs/organizations, buy  Amazon Gift Cards from a link on the site, then use those to buy your items. There will be no cost to you, and a benefit to them.

 

Advertisements

8 Responses to “Amazon’s numbers get fuzzier?”

  1. alanchurch Says:

    I wouldn’t miss it. Those numbers are of no interest to me

    • Bufo Calvin Says:

      Thanks for writing, alanchurch!

      I appreciate that feedback! I’ve always enjoyed doing them, because I like crunching data and the insights it can give me, but my sense has always been that most people don’t care about it. I suspect some industry people read it, and some other writers…it will certainly save me a couple of hours a month not to do it. It’s a takeaway for me, and that can be frustrating, but I’m thinking about how I might do some other monthly analysis…

  2. Lady Galaxy Says:

    I won’t miss all the numbers mainly because they took up so much room and made it take forever to page down to the previous article or to see comments, but I will miss your analysis of the data.

  3. A random Prime Reading book (February 2018): The Revised Vault of Walt: Unofficial Disney Stories Never Told (The Vault of Walt Book 1) | I Love My Kindle Says:

    […] Amazon fuzzied up their search results, I can’t do the monthly Snapshots I’ve been doing for years. I wanted to think of […]

  4. Edward Boyhan Says:

    I can confirm that a non-specific generic advanced search for Kindle titles returns a result of “over 6 million”. Intrigued, I decided to see what other results might be of the “over” kind, and what that might say about why this change has been made. I also just got interested more generally in what’s in Amazon’s web bookstore. I first looked at only English language Kindle titles (over 4 million). Then I asked how many titles are in the store irrespective of format (over 50 million); hardbacks (over 10 million); paperbacks (over 30 million). So with 6 million Kindle titles that comes to 46 million (presumably 4 million are in other formats that I didn’t search for).

    Curious, I searched for “used” Kindle titles — that returned over 1,000 titles (prompts the question just what the heck is a “used” kindle title? :mrgreen:). Just for giggles, I asked how many “collectibles” are in the Kindle store — that returned just two titles — both graphic sci-fi items. 😘

    Wanting to drill down a bit more, and see where specific counts might appear, I searched several genre categories (that are reputed to be the real growth areas for eBooks): romance (800,000), mystery/thriller (300,000), and sci-fi/fantasy (also 300,000).

    I then wanted to see whether any of the usual wildcard characters (“*”, “?”) had any effect. I searched for “X”, “X*”, and “X?” in the author field. All three returned exactly the same result: “more than 1,000) — seems that that search returned any title that had an “X” anywhere in the author’s name. I did manage to get one specific result when I asked for author “Ray Bradbury” any format — I got 107 results.

    So why is Amazon making this change? I can think of two possible reasons: one based on some competitive/business model analysis; the other based on some underlying technical change. I tend to discount the first of these because it would only make sense if the Amazon web bookstore was in some kind of competitive struggle — clearly, they are overwhelmingly dominant here, and I can see no advantage to providing fuzzy counts.

    The other makes more sense. All of Amazon’s web stores are built on an AWS foundation (in fact AWS grew out of the very infrastructure used to create the original web bookstore). As AWS (and the Amazon web store)gains in popularity, it may be that offering non-specific counts results in less load on the AWS servers (I find that hard to believe). The searches I reported here did seem to offer very snappy responses (but I don’t know if that’s any different from searches using the old methodology). My favored speculation has to do with the fact that all cloud vendors are looking to lever AI whenever possible. The current state of the AI art resides mostly on two underlying technologies: big data, and deep learning. “Deep Learning” usually involves some form of neural network. I can easily imagine how a new big data database engine, or (particularly) a neural network delivering exactly the kinds of results you are seeing. Both of these are more focused on discerning patterns rather than specific counts.

    What do you think?

    BTW I agree with Lady Galaxy: the post is too long, and (IMO) not very conducive to useful insights.

  5. Tom Semple Says:

    This behavior is for any search results for any product category on Amazon; it is not specific to books. I’m sure the web team didn’t think they were doing any disservice (much less becoming ‘less transparent’), rather that they were eliminating meaningless detail (which it is in most cases).

    Still you should be able to report when ‘over 6,000,000’ becomes ‘over 7,000,000’ and so forth. But after it hits 10,000,000 it might be awhile before the next ‘one significant digit’ number shows up (20,000,000).

    If you did counts for all sub-categories (as well as the category hierarchies) you might be able to derive something about growth rate as these increment by 100, 1000 or whatever. But it would be complicated – I think books can be listed in up to 3 categories and so that would need to be taken into account. And I’m not sure if publishers always select ‘leaf nodes’ (the most finely grained categories possible), or that the top 3 rankings Amazon reports are the same as the publisher-selected categories.

    You can also get a lower bound on category counts by tracking low-ranking books in each. For example Cosmology reports ‘more than 1000’ titles, but only returns 16×61 pages of search results (966 total). On the last page, there was a book that ranked #1980 in that category. So I’d infer there were at least that many titles. Somehow you’d need to develop a statistical model that could estimate total number of titles given a set of category counts.

    Likewise, there will be many books ranked dead last in sales rank (those without sales), and tracking these would at least that give a lower bound on total number of books in the store.

    So I think there is still something you could do to track this, but it will be different and the model would be subject to constant revision and refinement.

    • Bufo Calvin Says:

      Thanks for writing, Tom!

      Yes, you’ve identified some of the issues. In terms of the categories, books do get counted multiple times when you use sub-categories…and publishers sometimes choose contradictory categories to increase sales. I’ve mentioned before that I’ve seen the same book in both fiction and non-fiction categories.

      I could do some sort of analysis, but my continuity (which to me was what was most interesting…how Amazon is evolving) is shot. Given what appears to be a lack of enthusiasm for the posts generally, and the amount of work they entail, I’m going to hold off and re-think.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.


%d bloggers like this: