My Dog's Servant said:
Anarchist...appreciate the thoughtful post. The one thing that struck me, though, is it would seem that many of those--quite sensible!--rules would seem to make it even harder for new or struggling authors to get traction. Given your knowledge of the change search rules for Google, what would you think writers would have to do to get the algos' attention?
With time, it will become increasingly difficult to manipulate Amazon's algo. Authors' efforts toward that end will have an ever-diminishing influence over it.
From Amazon's perspective, that's a good thing.
Put yourself in Amazon's shoes. Your algo must deliver search results that do the following:
1. Instill trust in users. In the same way people have been trained to expect useful results from Google for any given search query, customers must be trained to think of Amazon's search functionality as useful.
2. Optimize the user's experience. If a user visits a book's sales page and subsequently clicks his browser's back button, that signals his discontent. As Amazon, you want to minimize user discontent. So your algo would suppress listings for books that have a bounce rate that exceeds a given percentage.
3. Maximize long-term sales. This outcome stems from trust and user experience. As Amazon, you must maintain tight control over which books appear when a user conducts a search. Or you must minimize authors' influence. Keyword usage would be the first item to receive your attention because it's one of the easiest to game.
My Dog's Servant said:
Also....how did you computer savvy types know what Google was focusing on at any given stage so you could figure out how to adjust (or game)?
We (search marketers) had mountains of data to examine. Additionally, Google wasn't doing rolling updates to its algo back then. It would do a major update once every couple years, and later once every few months. That made it much easier to test.
Imagine having 100,000 books in your backlist. Now imagine having software that pulled each book's ranking for a series of relevant queries. Now imagine having software that did the same for every other author's books. Now imagine that your software tracked changes in real-time and highlighted areas to adjust based on those changes.
As I mentioned, it was an arms race. I had so many software programs on my machines that I couldn't remember which programmers I had hired in which countries to design them.
My Dog's Servant said:
Amazon to date hasn't been very forthcoming with specifics.
It's in Amazon's best interest to keep its ranking factors under wraps. The moment it discloses how to rank, its algo will be gamed.
You can already see it happening with keywords.
On a related note, I suspect a similar effect will happen with purchased ads. If you're Amazon, you do NOT want people juicing their ranks, visibility and sales with external tools - think BookBub. You do NOT want authors to manipulate your ranking algo (not just search algo). Again, you want to maintain control.
I see a day when traffic from known promo entities will count less toward rank, visibility and sales. Meanwhile, Amazon will get its act together with AMS. In other words, it will start reducing the influence of sites like ENT, Robin Reads and BookBub, thereby making its own advertising platform more attractive.
Again, just speculation. But if I ran Amazon, that's how I'd do it.