Last week I talked about how a negative review on Amazon can actually help sell books. This week I’m going to go more in-depth on Amazon’s rating system as it applies to book reviews. Full disclosure: I have no secret knowledge of Amazon’s algorithm. Everything I’m about to tell you is experience-based only.
I first realized something was askew with the Amazon algorithm when I had 16 five-star reviews and only one 2-star review, yet the percentages in the handy-dandy review chart that comes up for every product showed 22% of people rating it as 2 stars and, of course, 78% giving it a five-star rating.
I’m by no means a math genius, but even I can tell that 1/16th is not the same as 22%. I wanted an explanation.
Like any self-respecting writer, I penned a letter to Amazon’s support staff, asking them to adjust the numbers for me. I got a swift reply. Here’s what they said:
The overall star rating for a product is determined by a machine-learned model that considers factors such as the age of the review, helpful votes by customers, and whether the reviews are from verified purchasers. Similar machine-learned factors help determine a review’s ranking in the list of reviews.
The system continues to learn which reviews are most helpful to customers and improves the experience over time.
Nice. But customers don’t know the intricacies of Amazon’s algorithm, so if they glance all they see is that nearly 25% of readers gave my book a lackluster review. In all fairness, the next time I viewed my page, the 2-star review percentage had been knocked back to 12%. I’m pretty sure Amazon’s machine didn’t suddenly learn a new trick. I suspect there was a glitch and Amazon wasn’t copping to it.
Back to the algorithm, which considers three things:
Newer reviews are weighted more heavily than older ones.
To make the most of this aspect, encourage readers to leave feedback even beyond the launch of your book.
There are little thumbs up/down buttons next to all the reviews that says “Was this review helpful to you?” The more thumbs-up presses the positive reviews get, the more they’re weighted. Same goes for the negative ones, so it helps if your positive reviewers do more than simply say “Great read!” The more insightful a review is, the more likely it is to gain a thumbs up. I’d like to go on record as disagreeing with this aspect of the algorithm. Someone may have written a review that helped sell the book, as my negative reviewer did, but that thumbs up will only make that review count as, well, more negative.
To finesse this section, encourage readers to thumbs-up positive reviews that have helped them.
This means those ARC copies you give out for reviews won’t be weighted very highly by Amazon’s software. That’s a shame, since in my mind, they’re totally valid. It also means when someone purchases your book in Australia and you’re in the US, the review does not receive a verified purchase label. I have no idea why this is — I’ve queried Amazon and not heard back as of this writing. I don’t know if this happens between other countries but I know the Australia/US anomaly exists.
Verified purchases are the gold standard — encourage everyone to purchase your book, even if they’ve gotten a free advanced copy.
Besides being aware of the components of the algorithm to fine-tune your marketing, you also need to be on top of your book’s product page, managing it on at least a weekly basis. I’m fairly certain that 22% number was a simple error on Amazon’s part and not generated by their algorithm, otherwise there would have been no change in the percentage as a result of my letter. Let’s face it, Amazon’s a big company and lots of things can go wrong. It’s up to you as the product owner to take responsibility. So make regular visits to your book’s page and be involved. Amazon’s response to me was lightning-fast, and I don’t really care if they admit fault or not, I just want the percentage to be more reflective of reality.
And if you want to sell books, you will too.
See you on the next page!