A problematic feature of the world in 2018 is that the social networks we have built seem to spread misinformation faster and wider than its more accurate counterpart, and this can lead authors to make decisions counter to their interests. One of the enduring myths surrounds “The Amazon Algorithm.”
That phrase alone is a red flag to me because anyone who has a good idea about how Amazon works knows there is more than one algorithm — for example there are separate algorithms which determine how books are ranked in the bests seller charts, in the popularity list, plus which books appear in search for a given term and what order they appear in. And that’s just the tip of a very big iceberg.
That’s not the only thing about the phrase “The Amazon Algorithm” which sets me on edge — speaking about it in such a way betrays a fundamental misunderstanding of what algorithms are, and what they can do.
Humans tend to anthropomorphize, well, everything. But characterizing algorithms in such a way is just plain inaccurate. “The Amazon Algorithm” isn’t some immensely powerful AI. People sometimes speak of it as if it was a 21st century Greek God, whose capricious nature is only matched by its wrath.
In fact, algorithms are pretty dumb.
I don’t mean that they aren’t complex, or that they can’t be designed in intricate and beautiful ways — of course they can. But an algorithm doesn’t possess any sentience or independence or creativity or latitude. It’s simply a set of rules for the system to follow. There is nothing arbitrary about it whatsoever.
It’s often a simple script which says something like (translating to plain English) “only include ebooks that are 30 days old or less.” Some are a little more complex, of course, but you get the idea.
Like the algorithms which determine what updates make it into your News Feed on Facebook, the website to appear first on Google, the video which will autoplay next on YouTube, or the artists who will be recommended to you on Spotify, the exact make-up of the algorithms aren’t disclosed by Amazon. Unsurprisingly. We’ve seen what scammers can do with just a little knowledge of how things work.
But that doesn’t mean we can’t try and reverse engineer what these algorithms probably look like. There are over seven million books to observe in the Kindle Store after all, which is quite the data pond. As a result, over time, we have built up a pretty reliable picture of how Amazon works, particular the bit which sells ebooks. (In fact, a large group of data crunchers had it pretty well figured out by 2012/13.)
At this point, there is a lot of good information out there on the topic, but there is a lot of bullshit spoken by people claiming to know a lot more than they do.
Patient Zero
One particular guy — who I won’t name — springs to mind as I believe he’s the source for “The Amazon Algorithm” bullshit. He presents himself as million-selling author, and an expert, when he’s neither. I dug into his background and found a cute little cheat had propelled him to what looked like chart success, when really there was a skeavy little formula behind this appearance of a fanbase.
Back when this was possible, he would drop the price of his books to $0.01 on Google Play, and then self-report the lower price to Amazon, whose bots would dutifully match that price, despite this being against the TOS. It was like having a free book in the paid charts, and at that price it attracted a lot of downloads, of course. The book would then rise up the popularity list also and start getting recommended to Amazon customers, at which point he’d raise the price to $2.99 and drop the price of his next book.
This heavily touted “success” of his — where he’s waving around sales numbers rather than quoting income — was than parlayed into a thousand-dollar mastermind course a couple of years ago, which he proceeded to sell to hundreds of fiction writers, without disclosing this wheeze, or that all his books were non-fiction: public domain prayers he had simply repackaged.
He’s not even a writer! Now he’s an expert talking about “The Amazon Algorithm.” What a world.
Affiliates and Kickbacks
One thing that annoyed me about this particular guy’s course is that it was heavily promoted by a bunch of people who should have known better. They were all affiliates, presumably earning a kickback every time someone from their audience signed up for the course. And most of them weren’t disclosing the fact either.
This is a largely unspoken area of murkiness in self-publishing. While affiliate relationships are not automatically shady, I personally don’t like this web of kickbacks for two reasons.
First, that financial incentive can be considerable. I’ve been offered seriously lucrative affiliate arrangements — hundreds of dollars per sign-up — to promote various courses and products and services to my audience. (I’ve never agreed.) While that arrangement might be standard in all sorts of industries, I think rewards like this act as a massive disincentive to kick the tires properly.
I don’t believe any of the people endorsing or promoting the course above knew about its shady provenance, but, if we are all going to be very honest with ourselves for a moment, perhaps the prospect of thousands of dollars in kickbacks dulls the spidey sense a little?
Second, these affiliate relationships are rarely disclosed. That’s not just ethically questionable it’s also illegal. The affiliate relationship is often masked as an offer too. “Use my special code SHADYBLOGGER25 when checking out to get a fab discount!”
I’m not saying that anyone offering or benefiting from affiliate arrangements is suspect, clearly that’s not the case. But I’m damn sure that this course would have gotten way less publicity if it had to rely on organic endorsements or word of mouth. Perhaps some of the endorsers would have done a better job at checking the background too.
I think we all owe it to our audiences to be honest and upfront when we financially benefit from something we are endorsing, whatever the legalities of the matter are, and if there is a reason you are reticent to make that disclosure, perhaps you should examine why. That little voice is a reliable bloodhound.
Anyway, there are a lot of bullshit artists out there — there will always be chancers and schemers where the internet and money intersect. You have to get good at spotting them.
Back to the algorithms (note the plural). Keep in mind that the Kindle Store is not ruled by Algogod, whose nature you can never understand. It’s true that certain aspects of Amazon’s immense and complex recommendation engine are a little more opaque than others, but we do have a pretty good understanding of how the whole thing collectively works.
Which means we have a lot of knowledge that we can apply to our marketing campaigns and our metadata to help us build audience and sell more books… without the need for thousand-dollar courses by internet marketers masquerading as writers.
Mark Twain famously said, “A lie can travel halfway around the world before the truth can get its boots on.”
Or did he? In an early example of fake news, turns out he never said it at all.
