A trillion star review — via NASA’s Spitzer Space Telescope

A more thoughtful Amazon

Big data to quell the tiny trolls

Craig Mod
Craig Mod
Sep 10, 2014 · 5 min read

by Craig Mod


You, the author. Years of work poured into your book. The pub date set. The day arrives. You shake with excitement. Your breath is held. And, yet, within minutes — despite there being no possible way for the work to have been read — the book has a one-star rating on Amazon.


Tough luck. The world’s unfair, right? It follows, too, that so is the internet. Comments aren’t fair. Reviews aren’t fair. Ratings aren’t fair. There’s vitriol and indignation splattered upon everything. It’s rare the community that manages to rise above snark and takedowns.

But, does it always have to work this way?


The landing page for most books is Amazon. You know this but, still, try it: Search for any book. Its Amazon page will usually be, if not the top result, most certainly in the top five. And for good reason — Amazon, despite publisher kerfuffles, is still the easiest way to get information about a book or the book itself. Physical books arrive next-day at a fraction of the cost of a brick-and-mortar shop. Digital books within seconds.

And so our online literary bed has been made; Amazon provides shelter for our words. This is mainly a good thing, I believe, because it consolidates and facilitates effortless commerce. Readers benefit from a consistent interface with stored shipping and payment credentials. Publishers benefit (usually) from the ease of sale.

Authors, however, get the short end of the stick.


Amazon, the de facto home for a book is a home over which authors have little control. And so it becomes a home subject to the whim of vandals. I have seen the newly released books of dear friends unjustly one-stared moments after going live. The trolls lurking in the shadows, ready to pounce. Those single-star attackers — of which there are many and whose work is voluminous — giddy and triumphant.


Of course, one-star reviews are sometimes justified and just. Not every book is a good book. Some are dangerous. Some should be marked as such. Debates should be thoughtful and thorough and multitudinous.

What makes a review “unjust”? Unjust reviews are usually marked by ad hominem attacks. Often, the reviewer supports none of his complaints with examples or exposition. It’s the equivalent of a paragraph length, “Meh.”


The reason we are upset by Amazon trolls is that their actions truly damage — both psychologically and economically. And yet these damaging acts themselves are devoid of rigor: Seven seconds to attack that which took 700 days to produce.

With little effort the attackers attach their bile to something greater than that of which they are capable. This dissonance — venom effortlessly and thoughtlessly spat upon the diligence or precision of long hard work — is from where the tension is born. It’s what enrages us. Especially those of us who create. And so, the factor dividing the “good” reviews from the “bad” is very often, simply, the presence of rigor. Of being thoughtful.

Thoughtfulness need not only be supportive. A thoughtful and rigorous criticism — even a takedown, perhaps especially a takedown — is invigorating, exciting, and, welcome. So what we are looking for is a way to sift away the thoughtless (which may, indeed, also include five-star reviews from an author’s friends or family), not to ignore, categorically, the contrarian.


Thankfully, Amazon has an incredible amount of data. And with great data comes great responsibility. Or at least the ability to make things right.

I am going to propose something. This something is very obvious. Has Amazon thought of it? Most definitely. Will Amazon listen to my proposal? Unlikely. Amazon tends to not listen. For years we have clamoured for a Kindle API. Has Amazon responded? Yes. By ignoring that clamouring and winning the market. So it goes. Still, allow me to shout into the Bezos void.

Amazon can both minimize the impact of thoughtless reviewing and simultaneously make the entire Amazon browsing experience better for all. You see, each reviewer has a rating. A “helpful” rating. And if someone is sufficiently helpless or useless — as many attackers tend to be — the “helpful” rating of their account can drop to 20% or lower.

I propose Amazon begin to use this helpful data in calculating ratings. Giving more weight to the reviews of those most helpful. Allowing the opinion — and stars — of one helpful reviewer, for example, to cancel out five thoughtless trolls.

I would happily choose — as would many of us— an Amazon in which I only saw reviews and star ratings as calculated by the most thoughtful. An Amazon of rigor. An Amazon that deprives the thoughtlessly angry or vindictive of attention. For without attention or impact, attacks cease to make sense.


But how does one fairly calculate these weights? Determine the “helpfulness” of someone? Currently, other users can click the “helpful” button on reviews. Though, considering the vast amount of data at hand, there is certainly an opportunity for a deeper, more passively calculated algorithmic nuance. Amazon could use a combination of the number of reviews written by some reviewer, their total “helpful”” votes, the frequency of their reviews, semantic language analysis of their text for obvious anger / attacks, whether or not they verifiably purchased the item in question, etc. The variables and weighting between an Amazon user, their actions, and their reviews is almost endless. But Amazon has near endless resources. And this feels like an area worth exploring.


Big data is boring data if it’s not used to inform or make our lives or experiences better. High-pass filters to skim the dregs from the online world seem like a no brainer. Exposing these filters and allowing us to dial them up or down is a must.

As for Amazon — like the world — it can be unfair. But through data smarts we can change that. Or at least make it a little more thoughtful.


For now, the newly published will continue to hold their breath, and those in waiting will keep their heads down, as always, working on what’s next, a wrinkle of rigor set deep in their brow.


The Message

A Pandaemonium Revolver Collection. Season 2 stars @anildash @alanalevinson @ftrain @hipstercrite @itsthebrandi @jamielaurenkeiles @vijithassar @yungrama @zeynep. Season 1 available on DVD shortly.

Thanks to Robin Sloan, Virginia Heffernan, Rex Sorgatz, and Evan Hansen

    Craig Mod

    Written by

    Craig Mod

    Probably walking on a mountain … http://craigmod.com

    The Message

    A Pandaemonium Revolver Collection. Season 2 stars @anildash @alanalevinson @ftrain @hipstercrite @itsthebrandi @jamielaurenkeiles @vijithassar @yungrama @zeynep. Season 1 available on DVD shortly.

    Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
    Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
    Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade