Self Publishing Blog

Publishing Questions

Copyright 2006 by Morris Rosenthal

All Rights Reserved

Questions About Books Sales

How many copies did my book sell?

Question) A representative at Barnes & Noble sent me to you. Please tell me how many books I sold between January 01 through December 02, 2004. My ISBN's are xxxxxxxxxxxxx

Answer) Not only don't I have any relationship with Barnes&Noble (other than selling them books) but nobody has that information except the publisher. If you have the name and e-mail of the person who referred you, I'd like to talk to them and find out what's going on. There is a service called BookScan (owned by Neilsen) that can give a total covering maybe 80% of retail sales, but the subscription is very expensive unless you're a PMA member, and even then it's $800/year. I don't subscribe myself.

Q) How many books sell in the area of literary/creative non fiction?

A) I'm not really sure how this genre is defined. Can "creative" non-fiction be non-fiction? Literary non-fiction is generally used to describe history and biography, but if all the political propaganda books published by both sides during election run-ups are included, I suppose that could have made for a big, temporary bump in "creative non-fiction" sales. In either case, I don't have any information on sales numbers for those books, I'm not sure if anybody recognizes them as a category and tracks them.

Q) How come yesterday Amazon reported 17 copies of my title in stock and only 5 today, but my sales rank went the wrong way?

A) As near as I can tell, their in-stock figures have been meaningless since their last major software upgrade. They routinely show books as being in stock, but having shipping times slower than 24 hours, when any book that actually is in stock gets a shipping time of 24 hours. It's just a glitch.

Q) Why don't Amazon bestseller sorts agree with the sales ranks?

A) The bestseller sorts have never been constructed from the sales rank of the moment. Keep in mind that Amazon is now updating all sales ranks every hour, but bestseller sorts are usually updated just once a day on some historical weighted basis. The "featured" sort often mirrors the sales ranks, as do the Top 100 lists, which are updated every hour. Sales ranks currently reflect what happened the previous hour. So even though the sales ranks moves up pretty much every time a book sells and drops every hour without a sale, the math behind it isn't simple. A book might be having a bad day or a bad week and have a sales rank much lower than competitors, but depending on the time constant for the bestseller list (past day? past week? past month?) it could beat them in the listings. Since Amazon keeps all this info private, the best we can do is make educated guesses.

Q) In a recent published journal paper, your analysis of Amazon sales ranks was listed in the footnotes for the basic data. However, the numbers on your website don't match the numbers in the paper. Can you explain?

A) I'm aware of two or three academic papers that have used the data various incarnations of my Amazon sales ranks article. The analysis is unauthorized, not confirmed by Amazon, and the ranking system has gone through several major overhauls (changes) over the years. Anybody who uses is as the basis for a scientific paper does so at their own risk. The Amazon ranking scheme has always been "secret" - my attempt at reverse engineering sales ranks has always included a fudge factor since I don't have any inside information. I actually got an accusatory e-mail from somebody else in academia who'd written a paper largely based on my Amazon sales ranks a year or two asking why I'd "changed" my curve earlier this year. My answer was that I'd collected new data points and realized that the curve I had posted at that time no longer fit well at the top end. The reason might have been that Amazon had made some changes (they're under no obligation to keep anybody informed, and they don't:-) or that my original points were wrong, but I was pretty amused since those authors had never bothered contacting me about the data before writing the paper.

Q) What makes it more difficult for Amazon and B&N to post actual sales than these stupid ranks?

A) It wouldn't be more difficult, it would be less difficult. However, the entire book industry treats sales information as proprietary, something they can sell or bargain with. The Amazon and BN sales ranks and the Ingram distribution sales are the only free info out there.

Q) How many books has Amazon sold for a book currently ranked 20,000 and published last year?

A) You can't determine anything from a single rank in time under the new system, and even under the old system a single sample would only be significant for a book that had only sold a couple dozen copies or less.

Q) Can I directly compute the number of books sold in retail stores from the Amazon rank?

A) Amazon sales ranks are based on Amazon sales only. I know I have something posted about extrapolating, maybe if it's not in the sales rank article you already found.

Q) Publish America tells me I made less than $100 last year, how can I check that?

A) I don't know how Publish America runs their accounting or their royalty periods, let alone their royalty computation. Never mind the money, the question is how many copies they reported sold and if they have a methodology for excluding copies sold through certain links. My understanding of the Publish America model is that they ask the author for a certain number of friends and family they can market the book to. It's not exactly a subsidy press, but it works more or less the same way, leading authors into believing they've been accepted by a trade publisher when in fact, they're a very different animal. As to exact numbers, Neilson Bookscan gets reports from the majority of retail outlets, but a subscription is out of reach (though only $800 if you join the PMA but it wasn't worth it to me).

Q) How does the industry define a "bestseller."

A) It depends on the genre. A bestsellers in mainstream categories are those that get on the big Bestseller Lists, the number of sales per week required varies with the season and competition. On the other hand, a book about refinishing furniture could be declared a bestseller by the publisher if it sold 10,000 copies. I think McGraw-Hill was labeling one of my PC books a bestseller by the time it reached 30,000. The most telling statistic about subsidy publishers is their lack of bestsellers, even in Amazon terms. You can gauge the success of publishers by their online sales.

Q) What does my e-book rank mean?

A) Selling one copy a day should keep you in the top 100 e-book bestseller list, but your overall rank won't likely break 10,000 at that rate.

Q) Can I make money publishing a book when similar books average 100,000 rank on Amazon?

A) Well, 100,000 isn't great, they're only selling a couple books a week. Amazon isn't the only place people shop for books, of course, but it's one of the main outlets for POD books.

Q) How many books would you expect a medium size publisher to sell of an Egyptian travel story.

A) I'd guess the average trade book sells about 2,000 copies the first year, maybe a couple hundred a year on backlist after that for a few years. Keep in mind the reason publishers keep doing them is they want to have the one that catches on and sells 20,000, but that's in the average, the mean travelogue probably doesn't break 1,000. I don't have any special insight in the genre, I'm using fairly standard non-fiction numbers for medium publishers.

Q) I just received my first royalty check and the amount is far lower than I expected. It shows 16 copies sold and I know that's false because my friends and relatives ordered at least 100 copies. My ranks on Amazon an d Barnes&Noble were both under 100,000 when I looked the other day, so I'm sure the publisher is cheating me and I want your backing to confront them with numbers.

A) The question you are asking is one of the most common I get. Unfortunately, nobody seems willing to do the follow-up. If you're sure that 100 friends of family have bought the book, have them all rip out the title page and send it to you, then you have something to give a lawyer. Transient Amazon and ranks don't mean much, if you hold an average rank under 100,000 for a few months, then you're probably selling a couple books a week. If you only checked once it may just mean you sold three books in the previous week, Amazon predicts a good future, then nothing happens. Check the rank again next week and next month, and keep a record. It only has significance in the long term, or if you hold a rank in the top 10,000 day after day.

Q) My book was released in October 2004 (three months ago at time of writing) and the Amazon rank has stayed under 10,000. My publisher says the distributor only reported one sale in October. Is it possible to achieve keep rank with the sale of one book?

A) Not any way I know of, but I'd need much more information to look into it. Momentary sales rank means very little, only the long term average has significance. Also keep in mind that different warehouses and distributors can have a book in stock and fill orders without your seeing them on the bottom line if the stock was purchased a previous month. For example, Ingram ordered 150 copies of one of my books at the beginning of the Summer (slow season) and had 20 left in August. My sales to Ingram of that book during July and August were therefore zero, my sales were all front-loaded. In your case, since the book was published at some point in October, it's quite possible that Amazon was taking orders that resulted in your sales rank rising, but didn't fulfill them until November. It's also not clear to me what reporting terms your publisher's distributor is on with them. It could be there's a 90 day lag between the distributor shipping a book and crediting the publisher, and the publisher may be reporting the number of copies they've been credited with as sold, as opposed to the number that have been shipped.

Q) How big is the poetry market in terms of both dollars and percentage of the total book market?

A) As a percentage, I suspect we could safely round down to zero. However, I don't have any numbers on the dollars amount, the only place I imagine might have that info is if Neilsen or Ingram breaks it out from the totals.

Q) I'm a reporter for the X Times and I need to know the biggest geographical markets in the US and where X ranks. Any idea?

A) Certainly the people at Neilsen Bookscan (the same as the TV rating people) would know, they are the main source for industry info on retail sales, which they track at point-of-sale. Unfortunately, a subscription to their services is well beyond my means, but as a reporter, you might get something out of them. Second shot would by Ingram Books, the largest distributor in the US, who must certainly track regional information. I haven't seen numbers published anywhere myself. I'd be inclined to think t would agree fairly closely with population, adjusted for new immigration, since the country is fairly amorphous.

Q) If I want to improve my Amazon rank, am I better off buying 100 copies in one order I do I have to put in 100 separate orders? Would 1000 get me in the top 10?

A) I read about people years ago who inflated their ranks temporarily buy buying large numbers of books in single orders. I just don't know why anybody would bother, and people who are legitimately interested in quantity buying would always get a better deal from the publisher. I understand that some authors think that being able to write "Top 10 on Amazon" in ad copy will make their books bestsellers, but I've never bought a book because it was advertised as being "Top 10" on Amazon (for two hours on Tuesday), have you?

Q) How many books on home schooling are sold every year.

A) It's an interesting question, but I'm afraid I wouldn't know where to start. You'd have to manually collect all the titles, or at least the ones that appear to sell, then use the Ingram stock/sales check and Amazon sales ranks estimate for each to get an estimate. It's possible the Neilsen BookScan has a category, but I don't have access to those numbers, the subscription is too expensive.

Q) Do I need to wait until my book is already highly ranked before sending hundreds of people to Amazon to buy it to get in the Top 10, or can I do it when the book is brand-new?

A) There are no fruits to having temporary high placement that I'm aware of. The famous cases of people who've done it and written articles on the subject, like Shel Horowitz, held onto their rank for a day, then watched it fade back into the hundreds of thousands as sales returned to their natural pace. If you can sell a couple hundred books in a day without spamming a hundred thousand people, more power to you, but I wouldn't invest more money accomplishing it than you expect to profit from those one day sales. In my experience, books that leap to the top on Amazon do so because they have marketing buzz coming from the bottom-up, not the other way around.

Q) I recently had a book published by St. Martins, and I'd like to know where to find sales info.

A) You can get Ingram sales from their automated phone system, but that's only a small fraction of sales for most large trade books. You can estimate Amazon sales from the sales rank and my site, but then you have to guess how much of your sales are through Amazon. The only people who really know how many books you've sold are at your publisher.

Q) Where can I find the average price for successful new books?

A) Just look at the Amazon Hot 100 list and run an average. It pays to do it yourself so you can see where the variations are. For example, if you ran an average in early September or February on a big enough sample, it would come out double the usual figure because of university texts.

Q) Can a book with multiple reviews on Amazon and no sales rank really have sold no copies?

A) Yes. As far as I know, a book with no ranking has never sold a copy.

Q) What are the relationships between books, music and DVD's sold online, and by the various Amazons?

A) Amazon.CA is a much smaller book market than the US maybe 10 million English speakers in Canada, but it seems to me that something like twice as many US produced books overall are purchased in Canada than in the UK. While both English and French speakers would buy music online, it looks to me like Amazon.CA primarily sells English books. I've seen an estimate putting music sales at a third of book sales, I suppose it would be pretty easy to run down. I've sold thousands of books through Amazon Associates, and people buy other items just to get the order total up for free shipping. I've seen lots of DVD's purchased this way, very few CD's. I don't know anything about the music business, but it's possible that the demographics are so different that music and books aren't comparable. I do know that that while the big chain bookstores in this country all sell music, the big chain music stores don't sell books, so music may have far more retail outlets. I once concluded that the top 2,500 or so books account for half the total book sales on Amazon. Around 97% of books have average ranks under 100,000, and even at these depressed sales levels, they are actually over-represented on Amazon, since they aren't available in stores. Many successful small publishers don't do well on Amazon, they primarily succeed through aggressive marketing to niche audiences through direct marketing. The music equivalent would be bands selling CDs at their shows.

Q) How come your estimates of the total number of books sold each year disagree with the PMA survey?

Note: The 2002 PMA "Rest of Us" survey estimated that small, independent publishers were generating $30 billion a year in sales, in a year that the IRS put total book sales for the whole industry at $23 billion. I believe this $30 billion estimate to be around ten times higher than reality, and eventually dropped my PMA membership when they failed to publish a clarification.

A) The PMA survey, which I believe was called "The Rest of Us" was based on several fatally flawed assumptions. I corresponded with them about it, but they were unwilling to make any changes because they were trying to relate their numbers back to the equally flawed BISG survey, to have an "apples to apples" comparison. My understanding of the PMA survey is as follows. They used a relatively small sampling of their own members to create a baseline for the average number of books sold for each book published. They then extrapolated for all small publisher based on the number of ISBN Bowker lists as "active."

Now the problems. First, the survey group was self selecting. Publishers who were in the business long enough to recognize that their was a PMA to be joined, willing to pay a membership, and still interested enough in publishing (ie, successful enough) to fill out a survey. The average reporting PMA publisher had 7 books in print. It's like using results from a survey of international business travelers to extrapolate how many overseas flights the average American makes each year based on the number of active passports. My own experience corresponding with the recent explosion of small publishers, who are primarily self-publishers, is most have 1 book in print. Second, and most importantly, active ISBN numbers are not a basis for extrapolating anything. According to Bowker, an assigned ISBN number that hasn't been declared inactive is active by default. We simply have to deal with reality here, which the PMA study does not. Books have to be sold somewhere, and that somewhere is usually stores, whether online, retail, chains, mass merchandisers, etc. The IRS has excellent information for these sales outlets, and releases the reports publicly. The best way to get a number for the sales of small publishers to start with total known book sales and subtract the totals from the large publishers who report. The pie is only so big, and most of it belongs to the big trades. My own estimate is that the PMA overestimates the sales of small publishers by a factor of 10.

Q) How can I compare my sales with the big subsidy presses like you do?

A) Just hit the "Search" tab on Amazon, scroll all the way down to "Power Search" and enter publisher names separated by "Or" . The search string I sent you was scraped out of the Internet Explorer address window, but it was generated by the Power Search:

publisher: authorhouse or iuniverse or xlibris or YOUR_PUBLISHER_NAME

after which I sort by bestselling.

You're doing really well for yourself on sales. The only way to get appreciable Amazon sales is to have lots and lots of publicity or to send people there from your own website, which is what I do. Since you already figured out xLibris isn't doing anything for you but taking a big cut, have you considered becoming a real publisher and doing your next book with Lightning Source?

Q) I keep confusing the different Amazon programs and what they mean. Can you clarify?


  • Amazon Associates - The Amazon program allowing website owners to sell Amazon books on commission
  • Amazon Marketplace - The Amazon program for direct sellers of new and used books, similar to ebay
  • Amazon Advantage - The Amazon program allowing small publishers to have their books stocked by Amazon

I've used all three of the above. There may be more detail than you're interested in this e-mail, but I figure, you asked (sort-of) and it's something that's come up a lot around here lately.)

The only way most small publishers can get their books sold through Amazon has been to enroll in Amazon Advantage, a program that requires a 55% discount, full returns, publisher pays shipping (same terms as Ingram Books), in return for stocking a couple books for 24 hour delivery. The best way around this is to have your books printed and distributed by Lightning Source. Amazon announced last year that they are going to start charging for this service, but encouraged publishers to sell their books through Amazon Marketplace as an option. Many Advantage members, including myself, wouldn't pay the fee and quit. Amazon Marketplace used to be for strictly used books, you'll see it's morphed into New and Used. The basic deal is this. If you sign up for $30/month on Marketplace, you can list books for free, Amazon takes 15% of the sales price PLUS they mark up the shipping and handling. Amazon charges $3.50 for media mail delivery (last time I looked), they pay the seller $2.25. If you don't sign up for the $30/month, they also charge a $0.99/book listing fee. When I sell a book for $13.95 on Amazon, they net $3.08. All they do is process a credit card transaction, completely hands-off. If you look at enough used books on Amazon, you see some people selling books for less than a buck. These are people who pay the $30/month for a listing, and make a buck shipping light books by media mail. On a $0.50 book, they make a buck, Amazon makes a $1.32, the customer gets a $0.50 book for $4.00, everybody is happy. In other words, Amazon wants to be ebay, or more specifically, PayPal, except they charge a lot more than PayPal does. When I first tried self-publishing, if Amazon had been available I never would have gone out looking for a trade publisher. All of the techniques described here about promoting and selling your own books will also help getting published by a New York house if that's your ultimate goal.

I was one of the earlier people to sign up with Amazon Associates, I think it was 1997, they even sent me a T-shirt one year. I've earned a few thousand dollars over the years sending people to Amazon to buy my books. When they ran bonuses for Associates signing up new customers for them, I was bringing them over 50 a quarter. The interesting thing about being an Associate is that you can see how much of an impact your links to Amazon have on the sales of a given book, especially if you're the publisher and can account for all sales. For example, barring drop-shipping from Ingram, my website links to "Start Your Own Computer Business" which I publish through Lightning Source, account for well over 50% of that book's sales on Amazon. Amazon has hundreds of thousands of Associates these days, I wonder how much of Amazon's sales they account for, and how badly Amazon would be hurt if the Associates went elsewhere. My business model as a small publisher is far more dependent on Google sending traffic to my website than on Amazon carrying my books, I can send the customers where I choose once they get to my site.

Amazon is also playing loose with their free shipping. I ordered a $206 Russian Language course last week and went for the free shipping. When I went to check the status, they weren't planning on shipping it for another 5 days, even though it was in stock and shipping in 24 hours. This to get me back for not paying $4 for UPS, on which they make a buck or two. I cancelled the order and got the tapes elsewhere online for $177. I'm not much of a comparison shopper, as you can see from the fact that I was able to get it $30 cheaper when I really looked. If Amazon can lose me, they can lose anybody.

Q) Why would book with ongoing sales have remainders available on Amazon?

A) It's a materials handling problem. Books are cheap to print when you get up over 5,000 or 10,000, paperbacks in quantity run less than $1.00. When a particular bookstore or distributor finds a book isn't selling for them, even if it's selling like hotcakes elsewhere, the publisher may tell them to rip off the cover and send it back for credit, or tell them to send the books to a trusted remainder house, who will mark them as remainders with a black mark. That way, the publisher doesn't have to handle the books again, worry about if they are in saleable condition, and gets a write-off. They have plenty of new books to sell in their place. In the old days, this didn't matter since remainders rarely competed with new books, but today, remainders compete head-on with new books online. However, publishers are slow to change their ways.

Q) How did you shape your old Amazon sales curve.

A) First of all, it's now been replaced by the small curve for estimating Amazon sales based on ranks at I did that big curve back in the middle of 2000, I don't remember the month, but I used a combination of the publicly available numbers from the SEC filings, and public comments from Jeff Bezos. One of the factors in the shape of the curve was that I was closely following all book news at the time. Last year, people were still fascinated enough with the Internet that individual sales for blockbuster books on Amazon were often reported as news, as in "The first day of it's release, the new Harry Potter book sold xx copies in 1 day on Amazon, etc..."

The dead-on data points I had, aside from the tail end, where you could easily decipher ranks for books that have sold less copies than you can count on one hand, were over 10,000, but steady for years. A micro published book of ours which sat at 250,000 for two years sold something like 25 books a year. A computer book of mine which sold around 200 books a year through Amazon sat around 40,000 in 1998/99. My first Build book was steady on between 1000-2000 in 1999, selling around 1200 books through Amazon. I don't remember what the multiplier was to move 1999 forward to 2000, but it was based on Amazon's growth. The problem with using data from books with sales ranks between 10,0000 and 100,0000 in the old system was that unless they were extremely steady over the long run, it doesn't mean much. There were lots of books that aren't selling anything holding onto spots based on old sales, and Amazon wasn't always consistent on their re-ranking cycle.

I don't know how your publisher was able to get weekly sales figure, unless they are exclusively doing direct fulfillment for Amazon. Most trades can only tell how many books they shipped to Amazon in a period, not how many Amazon actually sold, and that doesn't include books shipped for Amazon by Ingram. In any case, You need to average your number over a time period much longer than a week to get anything out of them, otherwise the signal to noise ratio is way too high.

Every year I suspect it gets harder and harder to start from scratch at this game because of all the deadwood taking up sales ranks. I wouldn't want to start over again unless they discard historical sales. To the best of my memory, a the curve shape was fixed as much by the points at the tail end as anything else. Once you work out the total number of books sold and you work to get that area under the curve, you find that you are fairly restricted, unless you allow the curve to get real bumpy.

Q) Is it legal for Amazon to sell review copies?

A) I'm pasting in a letter I sent to the Author's Guild back when I was a member:

"Your Spring 2002 issue was largely given over to the Amazon used book subject in a manner that mixed preaching to the converted with abuse of the bully pulpit. "Amazon Dispute Continues" was overly focused on the publicity the dispute has generated and the related article, "Shrink-wrap solution", offers a cure that is worse than the disease. Nick Taylor's "Letter From Your President" reminisced about review copies sold through Strand and ventured some comments about copyright protection that escaped me, but I'm not a lawyer. I have serious complaints about the way Amazon is selling books which I haven't seen addressed by the Guild, and they represent a more systemic problem that the sale of review copies. If the Guild wants to invest so much time picking fights with Amazon, I'd like to see more focus on battles that might be winnable, rather than the current posturing.

I write a successful computer title, "Build Your Own PC" for McGraw-Hill/Osborne, which is currently in its third edition. The third edition was published in January, and quickly replaced the obsolete second edition on bookstore shelves. Six months later, the obsolete second edition is still available on Amazon, both as new and used. If you've ever browsed around the Amazon catalog by following the "customers who bought this book also bought..." links, you'll understand the problem. Customers who arrive at Amazon by any method, including by links from my site, often browse around a little before deciding to buy. Unfortunately, if they leave the third edition page to look at, say, "Building a PC for Dummies", and then click on the cover of my book or try to return to "Build Your Own PC" through the links on that page, they'll end up at the second edition! People actually buy the second edition by mistake, as my Amazon Associate reports show.

What's so tragic about that? Computer hardware evolves rapidly and the second edition is truly obsolete. Rather than review copies or "true" used books, there are always a large number of remainders available for sale. Remainders were obviously never intended to cannibalize sales from new books, and hopefully publishers of books that are in edition will catch on and stop the practice. Amazon also claims they are helping their customers by retaining customer reviews through all editions of a book. In my case, the book contains some 250 photographs, of which exactly one is retained from the first edition and around thirty from the second edition. Just two of eight chapters in the third edition appeared in the first, and these are completely rewritten. I've had lengthy exchanges with Amazon, they refuse drop old reviews from the new edition.

As Nick pointed out, authors do need to eat. I continue to link Amazon from my website, and I picked up $90 in Associate Sales in July, enough to pay my first year membership in the Guild. I don't like many of the practices at Amazon, but they have the largest customer base by far, which means that customers I send there are more likely to have an existing account and buy with "One-Click" than if I sent them to or elsewhere. In my case, computer books sales support my other writing efforts, so I'm more interested in maximizing sales than howling at the moon.

I'm new to the Authors Guild, but I'm not new to tunnel vision. I'd like to see more effort focused on fixing what can be fixed at Amazon than simply attacking them from some imagined moral high ground. I believe the Guild is doing a poor job representing members on this issue and is confusing headlines with accomplishments."

Q) Amazon seems to be making a lot of changes lately, no?

A) Always something new going on at Amazon. They are progressing with the integration of their and properties. This means that books with Search Inside may become searchable on sooner rather than later, something they reserved the right to do in their Search Inside agreements. It also means that all of the websites with Alexis data now have an Amazon item number, and their vital statistics can be accessed directly from the Amazon catalog. For example, the ASIN for Lightning Source is B00006F5B0, which can be searched on at Amazon like an ISBN. Another innovation is the "Real Name" tag for reviews. According to their write-up of what the tags mean, reviews posted under the "Real Name" of the reviewer will be given priority. They haven't eliminated the anonymous reviewing yet, but as long as the real reviews appear at the top, it may cut down on the damage caused by writers who savage the competition, or readers who had a bad day and want to take it out on somebody.

Q) Sales ranks aside, how many sales does in take to have a "bestselling" e-book on Amazon.

A) You can make the top 100 in e-book sales on Amazon selling less than one copy a day. For Amazon UK, it only takes a couple copies a week. It's not a very big market, but if you're making over $10 a copy, it doesn't take too many sales to add a couple hundred dollars a month to the bottom line.

Foner Books Home | Self Publishing Blog | Contact