Amazon Sales Rank vs. Author Earnings
In responding to a question about an ebook-only publisher, PG suggested checking out the Amazon Best Sellers Rank as one way of determining how well the publisher’s books were doing and assessing the value-add the publisher provided over self-publishing.
PG has seen much less discussion about what Nook sales rank means, but that might be another place to check.
So, here’s question #1: Is this a fair way of evaluating how well an ebook-only or mostly-ebook publisher performs for authors?
Question #2: While we’re discussing Amazon Best Sellers Rank, has anyone seen any credible discussions concerning how sales rank translates into dollars for indie authors? If you’re at 20,000 with a $2.99 ebook, are you making $500 per month from that book?
Or is Amazon’s algorithm too volatile to make it a useful gauge of the dollars coming in the door?
Question #3: Everybody knows that lots of ebooks have experienced sales spikes immediately after Christmas as new ereaders and tablets are packed with ebooks. What about other sales patterns? For example, do sales go up on the weekend when many people have more leisure time to read?
PG hasn’t seen any analyses of sales rank that address these kinds of questions.
He doesn’t want to trigger obsessive-compulsive behavior centered on your KDP dashboard, but would be interested in whatever tribal knowledge his readers might have about sales rank.
, Amazon will pay Passive Guy a small affiliate commission on any of your purchases. 
This is the first full month that my zombie thriller “Orpheus” is at $2.99 on Amazon. For the first week of June, my rank probably averaged between 25,000-30,000. I’m on pace for a $200 month. That may or may not help you.
For whatever it’s worth, since I bumped it up from $.99, sales are just about the same, but profits are way up.
Might want to try $3.99 and see what happens.
It’s the first book in what is most likely a trilogy, so I’ll price #2 and #3 at $3.99 and hopefully start rocking and rolling.
I notice when I post a new book in a series that I get an up tick for all the other titles in the series.
Good luck trying to figure out Amazon’s ranking system. I’ve seen the ranking of one of my trad-published books improve by hundreds of thousands with what I assume to be one sale. Apparently, Amazon factors in the long history of a book’s sales performance when determining what a single sale does to its ranking. After the “excitement” of that one sale, my title settles back into the 400-thousands.
Paper books are counted on a different system though. I’m not sure if you’re talkping ebooks only, but I just want to put that out there. The paper book listings can get beyond 1,000,000.
In the 400k zone, any sale will give you a six figure lift in the rankings. As you approach the 40k mark, you start needing more sales. Between 20k and 10k you need (last month at least) twenty or thirty sales a day to hold your momentum.
I think the real metric to watch is the popularity list, rather than the best sellers list. That seems to be what customers browse on amazon.
Amazon sales rank doesn’t mean anything as far as I can see. It is recalculated regularly by a secret algorithm. I’ve never even seen proof that the same sales rank sells the same number of books across different genres.
I certainly agree that Amazon sales rank is a complete and total mystery to me.
I’ve been tracking daily sales data on Amazon for my titles for the past two years. I have zero correlation to day in terms of sales–they don’t sell more or less on the weekends in any statistically significant sense. I do see a slight bump in sales at the start of each month and the end of each month but, again, not really significant.
Biggest driver of sales, as nearly as I can tell, is releasing a new book in a series and/or having a sample chapter of another book as a bonus, complete with a link to buying that book in the back of the book they just bought/downloaded/borrowed. Second driver is linked sales from review sites.
I see that same pattern at the ending of one month, and the start of the next.
I release about 10-30 new works a month. The 3rd week of the month is ALWAYS bad for me. (Salaries? Power, phone, gas bills? Not sure.)
I would bet that sales spike at the beginning and end of the month is people starting a new month in their budget (or ending one and finding that they still have some entertainment dollars left).
I don’t know how accurate this post is, but these are the only estimates I’ve seen given on rank vs. sales: http://www.theresaragan.com/p/sale-ranking-chart.html
From my own experience, this is fairly accurate. I had a book in the Kindle top 20′s and found those numbers to be true as well the numbers up to 8500. Of course, sales vary daily with my heaviest sales on the weekends and my slowest sales on Tuesday.
This ranking is the closest to how my books track.
That said, on any given day it can fluctuate widely. The other day I sold 35 copies of book #1 and the rank was in the 6000s, but based on the chart(and other days)I should have been ranked much higher, so I just assumed everyone else was really having an incredible sales day!
One area that seems to be off to me — at least for book #2. On days when I’ve sold 10 copies, I’ve never made it to a rank of 8,500. It’s been more like 10,000.I suspect the historical sales trend and when you press the publish button for a particular book has a big effect on why writers encounter such big variances in their data.
Oh, and on sales spikes. Weekends are best for me and I’ve always suspected that it’s linked to genre.
Phoenix Sullivan and Debora Geary both watch sales ranks like hawks and could probably give you some solid figures. The mega-thread on author’s experiences with Select over at Kindleboards’ Writers Cafe has a lot of data salted throughh it as well.
Of course, ranks are relative and so the numbers to hit a certain rank vary over time. Plus Amazon changes their algorithms now and then too, which can upend everything.
IMHO, Amazon’s sales rank is nothing mysterious: it’s a rolling average which probably weights more recent sales, but doesn’t necessarily: it fluctuates wildly because hourly rankings DO that. (B&N seems to work similarly, but since they record it daily, it is much much less volatile.)
Also IMHO: rankings are extremely over-rated, especially by indies. It is a single measure, and only a snapshot measure at that. But it’s an easy and public measure, so people go for it.
I think the reason people love to look at rankings and quote them and wave them is because we writers are a really insecure bunch. We’re constantly looking for any hard number to prove or refute our worst fears, to defend ourselves against our (imagined) detractors.
We all have different goals, and because ranking is relative (you do better when others do badly, for instance) it isn’t a good measure for ANY of them.
These graphs were in the rough ballpark when I compared them against several authors’ self-reported ranks and sales.
http://www.fonerbooks.com/surfing.htm
http://www.fonerbooks.com/kindle.htm
I was surprised, given how old the data is.
Has anyone compiled a more up-to-date graph?
Paul Draker
[email protected]
@pauldraker
http://www.facebook.com/paul.draker
[email protected]
The rankings are all relative, and relative to a field that constantly changes; a book could fall in rank but rise in sales level. But they are a way to compare without everyone posting their actual sales numbers in public, and for non-indie authors, it’s often the only indication they have about their book’s sales.
Sites like NovelRank will track the Amazon ranks for you, but don’t believe the estimated sales numbers NovelRank shows; they are meaningless.
That said, very broadly speaking, a book in the top 100 is probably selling hundreds of copies a day; top 50 might be thousand(s) of copies a day. From ranks 500-2000 will be double digit daily sales; 2000-5000 should be a handful of sales (5-20). Under 5000 means it’s not selling large numbers, maybe 5 one day and 1 the next. The rankings are a bit sticky, especially at higher ranks; at lower ranks there tend to be spikes. But this is all VERY broad and varies enormously. It’s like being the tallest person in the room, then going to another room where you are the shortest, even though your height didn’t change at all.
Q1: Yes.
Q2: 300,000-400,000 average = 1 sale a month. 30,000 steady = 10-20 a day. It needs to be steady to be countable. Uk rankings can be very low with fewer sales. No analysis available because people don’t have enough books when they care to know (newbie level) and later don’t want to disclose information due to mob mentality and attacks. Few people will disclose actual numbers. Some will Just lie. (Plus you get a feel for it after awhile and don’t need to ask.)
Q3: Weekends suck. Tuesday to Thursday is best. Exception is Sunday when Amazon flushes all sales to your account for the calcalation. Reporting slows on Saturday day and at end of month, also the 15th. Holidays are painful. People don’t read when there are big events on.
I should note that I write erotica, scifi, thrillers, fantasy and some romance and horror. As I’m at 59% erotica right now, you could say that people don’t read me in the weekends because they are too busy ’living’ their erotic fantasies. Ahem…
You know maybe that explains holidays too. Oh my!
Hmm, romance writer here. Mondays and Tuesdays are generally very slow. Weds-Fri morning picking up steam, and Friday afternoons through the weekend are generally good.
People must want romance in the weekends to warm up and erotica to tie them through the week until they can get that romance. It also could be a male-female thing. Men need a fix me up during the week and women need preparation for the weekend staring mid-week. Everyone is exhausted from all this on Monday.
lol, could be!
As noted by some, the data is extremely hard to separate out in any meaningful way. First, you need to know the algorithm, which we don’t — we do know it isn’t linear. Which means trying to game it or draw conclusions from it is pretty much worthless — someone who’s new and selling like hotcakes right now might rank above someone who has sold consistently but unspectacularly for several months, yet the unspectacular might have higher sales (and more income).
Equally, genre matters a LOT. Romance has higher numbers than some, but price average is different than mainstream.
On weekend issues, the Kindle boards are full of anecdotal stuff that is completely meaningless. “I sold a lot more on Friday, which is when I had my sale day” — well duh, really? Or it was a Wednesday before a long weekend, blah blah blah. Plus genre stuff again. Some people argue for anything except Tuesdays when new movies and video games are released but nada. Seasonal slumps seem to hit in late winter, early spring, once the Xmas rush is over, but nobody seems to be able to say if it is really a reduction or just a different part of their own sales curve with distortions.
The only data out there I *do* really like are the ones from Bufo Calvin who has decent and consistent methodology on his site, running similar algorithms each month.
Good luck,
PolyWogg
PG, you had to go and trigger my KDP-OCD…
@Ace–Funny. I have my fewest sales Tuesday-Thursday. Best days are Friday, Saturday, Sunday. Consistently, across all platforms.
Biggest day for me lately? Mother’s Day. Especially on the Nook. It was great. Other holidays are also big sale days.
I think it’s all about who’s buying your books. I write romance, and the days when I (mom with school-aged kids) have time to sit down with a pleasure read are the same days I sell a lot of books.
Not as many as I’d like, but I’m trying to be patient.
As for Amazon rankings… ugh, when you sell about 1/day or less on a title, it’s volatile, but seems to hover around 50K-100K. Two a day is more like 40K. But it’s always changing. I’ve had better months than I’ve had lately.
Sigh. Working on getting another book out this month…the only promo I can stomach.
My biggest sales days, aside from my promo days, have been Christmas day and a few days after, and Mother’s Day.
I am one of those persons who has obsessively recorded my sales every night since Dec 2009, and then does monthly compilations and quarterly compilations.
What this has revealed is that for my books the pattern of huge spike in sales in January-March followed by a steady decline held true apart from the KDP Select promotions. In April of 2011 and April of 2012 my sales were 28% of what they were for the peak month in the post holiday spike. The difference is that I sold twice as many books in the post holiday period (but I also had 2 versus 1 books available), so even with the same drop I am still selling about 2 times what I was selling last year at this time.
On the other hand, last year the category that gets me the most sales (historical mystery) only had 80 books in it and my book had been #1 for 6 months when the post holiday 2011 sales came, in 2012 the category had exploded (a glitch was fixed), and it has over 2000 books–many of the new arrivals onto the list traditionally published bestsellers with long selling histories. Without KDP Select promotions I would not have been able to stay visible on this list-and therefore not been able to take advantage of the post holiday spike.
In short, my experience is that my sales are influenced by a seasonal pattern, but in order to benefit from seasonal spike, and not be irreparably damaged by the yearly decline, I need to be careful about paying attention to where my books are on the category lists.
I do find that day to day I can usually predict how my sales were for that day by what changes there have been in my rankings. It is not a perfect alignment-for example, since Maids of Misfortune, my first book has now been out over 2 years and sold vastly more books than the sequel, daily sales put it higher in the overall and category rankings. But I can usually predict the direction (more sales or less sales) by the direction of the rankings.
I did find the link mentioned previously to be pretty much my experience in terms of rankings relationship to sales.
http://www.theresaragan.com/p/sale-ranking-chart.html
I haven’t been able to discern a consistent weekly pattern–my instinct is that I sell more on weekends than the first of the week, but between holidays and promotions, which artificially affect sales, just when I think I see a pattern it disappears. If there is anyone out there who would like to crunch my numbers I would be glad to donate my data. I just don’t have the time or inclination.
I have recently increased my prices from $2.99 to $3.99 and will be interested over the next 2-4 months to see if this seems to have had a positive or negative effect.
As to PG’s very first question, I belong to a Historical Fiction Authors Cooperative, http://hfebooks.com/ and from looking at books of authors we are interested in recruiting as well as keeping track of the sales of our 30 members, I can tell you that there are differences in how authors do when they are published by small presses–much of it related to pricing.
PG, thanks for the discussion. Always stimulating.
M. Louisa Locke
Thanks for the detail, Louisa.
Louisa,
Have you ever compared your data to the estimates from NovelRank.com ?
I would be very interested in the results. You can download a .csv file with this hour by hour estimates of the sales of Maids of Misfortune going all way back to 5-July-2010.
His estimate is that you’ve sold 6381 copies of the Kindle version on Amazon.com, 435 copies on Amazon.uk, and 7 copies on other Amazon sites since that date. Is that close to your figures?
Dear William,
I went back and counting just Maids of Misfortune,
starting with August 2010-through April 2012 (I don’t have final figures for May yet),
US Kindle 28,355 copies
UK and other non US Kindle 2,099 (vast majority are UK)
So if the data you came up with from novel rank is correct it is a really off-base calculation. I had always heard novel rank underestimated, which is why I hadn’t tried it.
I wonder if other people have found this much disparity???
Mary Louisa
My experience with the accuracy of NovelRank’s sales estimates is not good.
Agree on Theresa Ragan’s numbers jiving with my own sales (which have ranged from 800 in the Kindle store to… the outer darkness).
And agreed on Novel Rank. If you sell more than 20-30 books a month, the ranking is way way off.
I think part of Amazon’s ranking includes “dynamics” meaning how quickly a book sells in a certain period of time, and “relativity” how many people buy similar genre of books during that period of time. My book “Arboregal, the Lorn Tree” was #1 on Kindle on May 12 & 13 in the category of Children’s Books -> Science Fiction & Fantasy -> Science Fiction. Number 1 is good, but the number of books that put me in number 1 position was 116. Not a number to pop the champagne yet. It seems that my book sold briefly on May 12, and very few other books sold in that category.
Let’s face it; if you sell one book more than the other fellow, you’re number 1. In the end the number of books sold count, not the ranking, but ranking does affect sells.
I agree that sales velocity plays a big role, DG.
Overall kindle rank to sales relationship changes with overall sales volume. Right now a ranking of 2k translates to less sales than it did in January, because of Christmas insanity.
Rough guide?
8-12k… 10-15 / day? Daily fluctuations still have a big effect on rank this high up tho
5-8k steady = I dunno, 20/day, plus or minus a few books? Not many books are steady / sticky at these ranks though; stuff seems to stabilize at 3-4k or better?
Better than 3,500, sticky, you’re probably selling at least 1000/month. Rank of 800-1k is probably 90-100 copies per day? Pus or minus some?
After that it probably gets kinda wild.
In general, I’d say this is an excellent way to evaluate an ebook only publisher. Amazon is the big ebook pond; if they’re not performing there, they’re leaving a lot of money on the table.
I’ve got a paranormal romance, and it sells better on weekends, pretty consistently. Also I’d say 90% of sales happen between 9PM and 7AM. How’s that for trivia?
The vamps are all out during your best sales period, Sarah.
You’d think there’d be someone out there reverse engineering the numbers, so to speak, using them (rather, snapshots) to get an inkling of the algorithm. Someone, obviously, with more numbers knowledge than I have, and more time on his/her hands.
Weekends are better for me. Here’s a little data for the month of May for a couple of my books:
Cold my Heart 3.3 a day–typical rank: 35,000
Footsteps in Time 35.5 a day — typical rank: 2500
Winds of Time 21.3 a day — typical rank: 5800
The Good Knight 13 a day — typical rank: 10,000
The Last Pendragon 5.6 a day — typical rank: 22,000
Reasonable answers to PG’s questions could come from carefully analyzing the NovelRank output over time. Does everybody know about http://www.novelrank.com ? The guy who runs that site tries to estimate sales of books based on the hourly updates to sales ranks. His system isn’t perfect. It tends to underestimate the sales of the top selling books.
Yeah except he also says that a higher volume of sales will be quite inaccurate — which many people have found to be true. It’s because it counts a bump up as a single sale, whereas that bump could represent anywhere from one to 100 sales within that hour before the rankings change…
PG. No worries about triggering OC behavior. I’m already there without help.
Question #1: Is this a fair way of evaluating how well an ebook-only or mostly-ebook publisher performs for authors?
Not a clue, but with no other readily available tool, it’s probably the about the only game in town.
Question #2: I think the inherent problem is the relative nature of the sales rank. For example, my sales rank dipped a bit in the post Christmas period, but I sold a lot more books and made a lot more money. I was doing well, and quite happy with sales, but other authors were doing even better. I don’t think you can use a relative rank to quantify sales. A sales rank of 1,000 will net you different sales in different months. It’s not a question of volatility in the algorithm, as much as the nature of the beast. Also, I’ve never heard mention of a price component in the sales rank algorithm, and it seems to me that you’d need that to gauge revenue. For example, ten weeks ago I raised the price of my only book from $2.99 to $4.95. My sales rank is down, but revenue is much better.
Question #3: What about other sales patterns? For example, do sales go up on the weekend when many people have more leisure time to read?
I’ve watched this closely for about 6 months and there’s no pattern that I can discern.
Following are last 6 weeks SR vs. sales for Deadly Straits after a price increase from $2.99 to $4.95 on 1 April. Sales volume dropped substantially in the first two weeks of April, then started to recover. I finished the month even with March (revenue wise) because of the increased per book revenue. By the end of the April, sales had more or less stabilized at the following levels and have remained there for the last 6 weeks:
B&N – SR almost rock steady @ 2,000 – Average 16 Sales/Day
Amazon US – SR Range 2,000 to 1,500 – Average 49 Sales/Day
Amazon Uk – SR Range 4,000 to 1,300 – Average 8 Sales/Day
What I’d really be interested to see is other authors’ experience with regard to price increases vs. revenue. I think it will always be anecdotal unless/until someone starts soliciting detailed sales data and crunching the numbers, but anecdotal is better than nothing. It’s really pretty nerve racking to pull the trigger on a price increase and watch your SR head south, but I don’t think free and $.99 are going to cut it much longer.
Hope this helps. Anyone else who feels like sharing, please jump in.
I watch the six weekly payment chart like a hawk. It’s a good indicator of income when changing prices of just generally. You can also compare 1 week to another easily, which helps.
Ace. Thanks for the tip. I never paid much attention to the 6 week chart. I’ll keep an eye on it.
Another thing to keep an eye on — especially if you are in the “book every couple days” territory, or lower — is that sometimes sales rank and book-sales-reported-on-dashboard lag. Sometimes I see the sales rank shoot to the 50K or 80K range, but no sale for a few hours, and sometimes I see the sale, but the sales rank does not budge for those few hours.
I suppose if everyone went to Amazon and bought my books tomorrow, I’d be able to give more data…
Oh, and then there’s the “not a US sale” issue, where the royalties drop to 35%, though the ordering is through the US store. Affects sales rank, sure, but doesn’t give nearly the same amount of royalty.
(And then there’s the short story someone returned after a few hours of purchase; the sales rank seems to retain the “it was bought!” even a day afterwards, despite the return!)
Don’t forget items brought at the end of the month that are returned. Negative sales suck.
I’ve managed to miss those, thankfully. I console myself that someone who got their money back is less likely to gripe than someone who didn’t, but didn’t like the book anyway. (Though for the short stories, I sometimes wonder if it was read-and-return. >_> Oh, well.)
The interesting thing is that returning it didn’t affect the sales rank boost that the purchase gave! (I wonder how long before that becomes an attempted tactic to artificially boost sales… While, “Buy it! If you don’t like it, return it! Amazon makes it easy!” is a fair thing to suggest, if someone manages to scrape together an extended family to all do this on the same day… (Or if someone just pays a service to do that with a lot of fake accounts!))
I wonder that too, with my shorts, Abeth. Amazon says they keep an eye on accounts that have too many returns, though~
Since my shorts are all 99c, at least I’m only out a bit more than a quarter if someone does that. *wry* (It’s a bit sadder with the novels.
But so far, that’s only happened about once, that I recall. 
)
And then there was the time one story got two sales and in short order one was returned — I wonder if somehow someone managed to purchase two copies! Definitely a reason to want a refund on one, if so.
I track my sales daily on Amazon, because I can and I like to see what my books are doing, and I’ve found that sales do spike on weekends most of the time. It’s not set in stone, but it does happen for the most part. I’ve often found my books sell better on Mondays or Wednesdays too though. It’s all a black art!
Sales ranks are very difficult to use to determine the number of books sold per month if you maintain that rank because it all depends on whether Amazon is having a good month of traffic or an average month. Obviously if Amazon is having amazing traffic, then sales tend to go up for each rank, but your book doesn’t necessarily change rank. Also, you could continue to sell the same number of books and go up in rank if those above you sell less and slip down the ranks. But in general, it’s a good measure of how well a publisher or book is doing. You just can’t use it to decide how much that book is making.
Plus, you need to take into account with a $2.99 book that there will be some sales from outside the US that only gain 35% royalty, not 70%.
I’ve found that it used to be easier to see patterns in sales, but since KDP Select and the flood of free books have led to lots of readers filling up their Kindles for free, it’s a little bit stormy on the sales front.
Good post. Agreed.
I have 18 months of daily sales and rank data and I’ve blogged them in a response to your questions, PG. You can see the full blog here: Amazon sales rank = how many books sold.
Thanks, Derek.
Every time I put a book on free promo on Amazon, it shoots up to #2 in Action/Adventure. If I move 20,000 copies, that’s where I end up. If I move 600 copies, that’s where I end up. I am curious as to the differences in the paid and free ranking system.
I wonder if price is taken into consideration in the “paid” ranking, though. It seems like it should since the person sitting at #6 with a price of $3.99 is probably making a lot more money than the person sitting at #5 with a $.99 price tag.
Splitter
Very interesting discussion, guys! I actually try not to obsess about rankings, instead worry about writing the books.
Though after the NYC publishing experience — and royalty statements I struggled to understand — having data/metrics available for review at any time is delicious!