Reviews And The Tyranny Of Averages

I've recently encountered a couple of articles from bloggers decrying the damage 1-star reviews can inflict on a struggling author. You can find one of them here and another here.

Both bloggers find fault with the reviewers themselves, a point of view I can mostly agree with. Many are just trolls, Others are trying to stand out as contrarians, while all they are actually doing is making themselves look petty and mean. Both bloggers also mention that their opinions of 1-star reviewers stir up quite a bit of controversy on social media. Evidently, some folks think that these people should be able to speak their minds no matter the consequences, even if they are petty and mean.

When a book has hundreds or thousands of reviews, as best-sellers do, a smattering of one- or two-star reviews is to be expected, but they're not going to have much of an effect on sales. But when a book has very few reviews, a serious and widespread problem for unknown, self-published authors like yours truly, even a single bad review can have serious consequences. This is not, however, because of the review itself, but because of the effect it has on the author's average star rating. This is the first thing you see when you land on the product page for the book on Amazon or Goodreads.

Let's say a book has a total of five reviews that break down like this:

1. Five stars - 1
2. Four stars - 3
3. One star - 1

OK, that's four good ratings and one zinger. No problem? Not quite. The total number of stars for this book is 5+4+4+4+1=18. This makes the average star rating for the book 3.6, which Amazon will round down to 3.5. That looks like this:

Three and a half stars. If you are a reader browsing titles to buy, are you going to look any further than this? Probably not, when there are so many other titles with better ratings to choose from. This will lead to fewer sales of the book, which means fewer reviews, and the thing becomes a death spiral. Thus the solitary 1-star review this book received proves to be its undoing.

Now, our inquisitive book-shopper will also see that the title has a total of five reviews, but there's no help here: "Oh, it's a three and a half star book only five people have bothered to review? I'll pass."

I call this the tyranny of averages. It's a cruel fact of life that mostly affects struggling authors, but even best-sellers by well-known authors are somewhat affected by it. For a case study, I chose Isaac Asimov's The Gods Themselves. Asimov was one of the greatest writers of the 20th century, known for some of the most enduring works in all of science fiction. Although The Gods Themselves is somewhat lesser known than Asimov's classics like I, Robot, it won both the Hugo and Nebula awards for best novel in the year following its publication. In the sci-fi world, these are the equivalent of the Oscars and Golden Globes for movies.

I reasoned that such a title would have close to a five-star average, but when I headed over to Goodreads to check, I discovered that the Good Doctor's masterpiece is barely in four-star territory, with an average rating of 4.09 stars across over 48,000 ratings. What gives? To find out, we have to dig into the details, something few people will do when they're just browsing titles looking for something to read.

As a lifelong Asimov fan, it's astonishing to me that over 2,000 people rated this book negatively. It's even more astonishing that over 9,000 people rated it as just OK. An overwhelming majority of readers gave it 4 and 5 star ratings, so does it seem to you that this is really a 4.09-star read? Let's see what we can find over on Amazon:

Only 4.2 stars? Not much better, but at least Amazon doesn't make you dig for the histogram. Here, we can see that the skew is a little more favorable, but the all-important average is still dragged down by the 1, 2, and 3 star reviews. 

If an award-winning, best-selling book by an iconic author can be hurt by a relative few bad reviews, what's a struggling indie author to do?

…

Cross-posted to Medium

Announcement for prospective reviewers

I was surprised the other day by a message in my inbox from someone who had used the contact form you see to your right to tell me they'd like to review my books. I wanted to be flattered, but when I took a look at their website, I observed that they are willing to review books for those who are willing to pay them $99.

Amazon expressly prohibits paid reviews, a good policy that protects authors and readers alike. Nobody likes to be scammed. When I launched my authoring career, I expected come-ons like this, so I didn't get upset about it. I'm just serving notice here that I was not born yesterday.

Mobi Dick

So I have been trying to drum up some more reviews for Joshua, and following the advice I found here, I looked up a few book bloggers who might want to read my novel and, I hope, go on Amazon and Goodreads to tell people it doesn't suck.

It turns out there are a few bookworms out there who do indeed take a chance on self-published authors, although it seems like a lot of them do it just to make money. However they do it, though, a lot of them prefer to get manuscripts in mobi format so they can be easily read with a Kindle.

Mobi? I have to confess I never heard of it before I started my quest for reviews. I've heard of lots of file formats, but this one is unique to Kindle users. But no biggie, many have told me; all I need to do is go to my book details on my KDP page and download my manuscript in mobi form. There. Done! So I go there, and I get:

WTF?! did they lose my manuscript or something? No. The problem is that I uploaded my manuscript as a .kpf file, something spat out by Amazon's Kindle Creator app when you feed it your manuscript. In order to download your book in mobi format you have to have uploaded it in mobi format.

Anybody see the disconnect here? What kind of idiot would clobber their original mobi file, thus making this expedition necessary? Don't answer that.

Joshua began life as a couple of short stories about a boy who gets into all kinds of humorous predicaments. I wrote them in HTML. After I wrote a few more of them, I started to get more serious about the project so I abandoned the HTML and converted the text to a Google Doc. There are a number of advantages to working this way as you are developing your manuscript. Everything you write is saved in real time to your Google drive, which you can access from anywhere, on any device that supports it. When you're away from your PC and you have an idea for your story, you can just whip out your phone, open up the Google Docs app, and start writing. I wrote a significant portion of my manuscript for Joshua in this way.

Once I was done with it, however, I needed to format it for print, paying attention to paper size, margins, and text placement in order to conform to Amazon's standards. I could not do this in Google Docs and I didn't want to do it in my severely outdated copy of Word. What I really did not want to do was pay for Office 365, the latest bloated atrocity of an office suite from Microsoft. I did, however, have a copy of Apache Open Office that was up to the task. Sort of.

I tried exporting my manuscript in Open Office's preferred format, .odt, with some funky results. Things I had formatted within Google Docs went all wonky when I opened them in Open Office. Trying to fix them all piecemeal proved to be too much of a challenge. I finally went back to Docs and stripped the manuscript of all formatting, re-exported it to .odt, and painstakingly reformatted everything from the chapter headings to the italics. After several days' hard labor, I had a manuscript that was fit to publish as a paperback that met Amazon's standards. It was extremely quirky, however. It was very easy to mess up my page numbers. I had to input my TOC manually. When I finally had it ready to publish, I exported it as a PDF, and hooray, nothing got trashed in the process.

But it turns out that .odt is not an acceptable format for Kindle Creator. I went back to Open Office to export it as a .doc, and it trashed my page numbers again. This wasn't really a problem, however, since Kindle Creator strips out the page numbers anyway. So I imported it to Kindle Creator, went through it page by page to fix up things that didn't import in a way I liked, and finally, I had both a print-ready and Kindle-ready manuscript to upload to KDP. A couple of days later, both were live on Amazon. Yay.

But troubles still lurked unseen, surfacing when I tried to make a mobi. There is a very good, free, open source tool known as Calibre that will convert manuscripts into both mobi and epub formats. Although it has a few quirks, it's pretty slick. You import your manuscript, turn the crank, and out comes your mobi. Well, it isn't that simple, but you get the idea.

But Calibre doesn't accept .doc files for conversion, and Open Office doesn't produce .docx files (no idea why). Calibre does purportedly convert .odt files, but it would rather have a .docx. I thought this might be my problem, so I got another free tool, LibreOffice, that can export a manuscript as .docx. Problem solved? I wish. Calibre converted my manuscript to a mobi, but it did not produce a table of contents from the chapter headings like it was supposed to. It also made a goof with the picture of Robert Fishell I use in my About the Author page at the end of my book. No matter, I just got rid of it rather than try to debug it. That still left me with a TOC problem.

I started plowing through the documentation for Calibre and discovered that it works its magic by converting whatever you give it to an intermediate form in XHTML that you can edit, if you understand enough about XHTML to know what you are doing. I do (although I'm pretty rusty), so I was able to generate a TOC manually. This left me with a question of why Calibre didn't just pick up my chapter headings to begin with.

I reported this as a bug to the developers, who promptly told me I needed to use heading styles in my manuscript, marked my bug report as a non-issue (as in RTFM, dummy), and went away. Leaving me with no idea why I couldn't get a TOC from the chapter headings which were quite clearly formatted using heading styles. I figured the problem was that converting the manuscript so many times had somehow corrupted it, but I didn't know how to fix that without stripping it down to plain text and starting over.

I respectfully asked them to look at the document again, or at least read what I told them in the first place, and what do you know – I had discovered a bug after all. Evidently the problem was that I numbered my chapters. Lots of, if not most, book authors do that. But it turns out that Calibre sees those numbers and thinks you made a list, and thus tags them with <li> rather than <h1>. If you don't know anything about HTML, I won't go into that, but the upshot was that the developer I was working with figured out the problem and fixed it for the next biweekly rollout. So you can thank me for my persistence in pursuing this issue if you're using Calibre and need a TOC for your numbered chapters.

And I can now give the thousands of reviewers who are champing at the bit to read Joshua a fully functional mobi, or even an ePub if they want it that way. I even figured out how to put Robert Fishell's mug back into the About the Author section. Go me.

Joshua will go live on BooksGoSocial in July. Watch this space.

So What Do You Do When You're Not Writing?

When I was setting up my author's page on Amazon, their bot asked me how I deal with writer's block. My answer was that I go and find something to fix. The problem then becomes fixer's block. I have often boasted, unseriously, that I can fix anything. More seriously, I can fix almost anything.

I'm particularly good at fixing bicycles. I need to be, since I own four of them, my wife has three, and my kid has one that she never rides. All of our bikes are pretty well maintained at the moment, but I discovered there is something of a market for people who know how to fix bicycles around here. So far, I've had a couple of them in. I figured $25/hr. was a reasonable rate for an old retired guy who works out of his garage, and so far, I've made $125 this year. That's a lot more than I've made on my books.

The point to fixing bikes, however, is not to make money, but to ride them. I am having a pretty good year so far. I've got around 400 miles in – much less than I'd have at this point in my glory days, but better than I've had than any other year since my spine surgery.

If I haven't been writing as much lately, that's why.

You Can Judge A Book By Its Cover?

I just updated the covers for both of my titles on Amazon, in the hopes of drumming up a few sales. According to some credible advice I was recently given, people can and do judge books by their covers, at least when they're browsing for them. So here they are:

They do look a lot better than the ones I originally published. I used my own artwork on both, but I originally used Amazon's Cover Creator tool to generate the text portions. This time around, I used Snappa for the text overlays, which gave me far more control over the appearance of the final product.

Only time will tell if it makes any difference. I needed to fix a few little typos in Joshua, things like using hyphens where I wanted en dashes, but in rereading the book, I re-concluded that my writing does not suck. All I need are some readers.

Barbara

Memories. We make them all the time, even when we are not aware of them. In every waking moment, and even in dreams, our senses provide us with information. We take in so much that we can't store it all, but it remains a mystery why some things are discarded and others are saved. Most fascinating, however, are the memories we don't realize we have. When we rediscover them, it's almost as if the experience is new. It's one of the things I enjoy about life, especially now when there are fewer and fewer truly new experiences to be had.

All it takes to jar loose some of these memories is a stimulus of some kind. That can be anything, like hearing an oldie on the radio, or even some random object, but nothing is more compelling than a photograph. For some time now, I've been meaning to digitize a collection of old 35mm slides my father took between 1950 and 1978, the bulk of which were taken before 1962: my childhood. As I went through them, I found a lot of images that pulled out long-lost memories of my youth, of my long-dead parents and grandparents, of beloved pets, of things I did and places I went. Some made me smile, some made me sad.  Of all of them, none were more evocative than this one:

The dopey-looking kid with the Dumbo ears, the red underwear, and a clothespin holding up his pants, is me, somewhere between the ages of 7 and 8. The absolutely adorable little girl in the foreground is Barbara. Her family owned a tidy lakefront cabin nextdoor to the little house trailer my family used as a summer place. Here is another one of us taken a few minutes earlier or later:

I look a little less dopey here, and Barbara looks no less adorable, but it's important that she and I were the only kids in the pictures. We were something to one another. I was much too dopey to realize it at the time, but the power of the feelings that washed over me when I saw these made me realize that I loved her.

My daughter came into my office while I was scanning these. I asked her if she recognized anyone, and she picked me out immediately. She asked about the girl, and I told her, "That's Barbara. She was my first girlfriend."

When she asked how old I was here, I said I was 7 or 8. I don't remember the year, and until I saw the photographs, I didn't remember the occurrence. There's a plastic tablecloth on the picnic table in front of our trailer, so there must have been something important going on. It might have been my birthday, or hers, or my sister's, or something else entirely. Whatever it was, it was a time that Barbara and I were together, and that made it special no matter what day it was. Even if I didn't understand it.

My daughter asked me how I could have a girlfriend at that age. I told her that even when you're a child, you can have attractions, you can have feelings for someone – even if you're too dopey to get it. I was a little surprised I had to explain that to her, as girls are more mature than boys at that age. Did Barbara have feelings for me? She at least liked me enough to be hanging out there at the picnic table with me, whatever the occasion was. She's standing close enough to me that maybe, just maybe, she felt something for me, too.

In my novel Joshua, my young protagonist develops feelings for a pretty little girl in his 3rd grade class, but in Joshua's mind, it's an unrequited attraction that has nowhere to go. I was far ahead of him here. I think that if I hadn't been so dopey, or if I could somehow return to my childhood to set some things right, I would ask Barbara to sit out on the dock with me after dark to look at the stars. What a memory that could have been.

The Matrix Revisited

Amazon recently added the Matrix trilogy to their Prime Video lineup (disclaimer: this is not a sales pitch, but they do carry my books).

Image may be subject to copyright. Reproduced
here under fair use doctrine.

I hadn't seen any of these films in a while, and I wanted to see if they'd held up in the 20 years that have passed since the first film was released. I was not disappointed. The action, dialogue, camera work, and special effects are still first rate. I had forgotten some of the story, so there was enough suspense to keep me absorbed in the plot.

The Matrix is one of the more influential films of the past 20 years. So much of it has become imbued in our culture. Everyone knows what you're talking about if you say "I wish I'd taken the blue pill." Morpheus's "What if I told you..." line has been adapted to hundreds of memes. And the reflections in those pince-nez shades – pure cinematic genius.

All in all, it was 5+ hours well spent. 20 years on, I'm still awaiting Neo's return. Yet in spite of the durability of the films, I found them dated. The late 1990s were a time of great optimism, the ascendency of the digital age. Home computers had become commonplace, and everyone knew about the Internet. Computers amazed people, and there seemed to be no limit to what they would be able to do in the future.

Moore's law was well-known in 1999, and 20 years in the future, it has held up. The computers of today are hundreds of times faster that those of the late 90s. They have become far more accessible, far more affordable, far more ubiquitous. Link speeds are also hundreds of times faster, fast enough for me to stream the Matrix films in 1080p HD on a 10" tablet computer. I keep up with such things, and I can say truthfully that we are still nowhere near the limits of computing capacity that can be brought out of silicon. Computers, however, can still do no more than what they could do in the 1990s. They just do it a lot faster.

In the films, The Matrix is a construct of sentient programs, used to deceive the world's population into thinking they are alive in a civilization modeled on the world of 1999. The programs have conquered and enslaved humanity in order to provide the electric power the machines need to survive. I won't go into the absurdity of this idea as it's essential to what is otherwise a great story. To people actually living and breathing in 1999, it might have seemed plausible. But it isn't power generation that trips up the plot. It's the idea that somehow computer programs could become sentient.

Most people had, however, already accepted this idea. It is a familiar theme in science fiction dating back to 1950s, perhaps farther. 1968's 2001: A Space Odyssey features a malevolent machine intelligence named H.A.L. who decides the secret he carries is too important for meat intelligences to be trusted with. Many other films, such as 1984's Tron feature sentient programs as both sympathetic and malicious characters. In the Terminator films, Arnold Schwarzenegger plays both. And who can forget C3PO and R2D2, among the most memorable machine intelligences in cinematic history. Indeed, the idea of machine intelligence has become so essential to good storytelling that folks like me dare not even challenge it. Fortunately, we do not have to.

Last night, I had a rather frustrating "conversation" with Alexa. The Echo Dot I keep on my nightstand had a flashing yellow ring that I wanted to turn off so I could go to sleep. I asked Alexa to turn off the yellow light and she told me she could not find a device called "the yellow light." I asked her "why do you have a yellow light?" This time, she told me it was because I had notifications from Amazon shopping. Earlier in the day, I had posted a question about a film scanner I was considering, and a few people had answered it.

I could not pose a command that would get Alexa to turn off the light. This was irritating, but I thought all I would need to do was go on the Amazon website and read the answers I'd gotten. Nope. She still had a yellow light. After some time, I figured out that the only way I could get her to shut it off was to let her read the answers I'd already read, one by one. Now, she does a remarkable job with text-to-speech synthesis. The vocalizations were intelligible, and even her inflections were good. She really sounded like a human woman reading the answers to me. That bit of amusement mitigated my frustration somewhat. But it was a lot of trouble to go through just to get her to turn off that annoying yellow light.

Alexa was, of course, just executing a program. It goes a little like this:

   The Amazon cloud says Bob has notifications. To inform him he has notifications, I'll make my ring flash yellow.

   Bob says "notifications." I'll retrieve them. I tell him he has four notifications from Amazon Shopping.

   I'll read the first one.

   Bob says "notifications." I tell him he has three notifications from Amazon Shopping.

   I'll read the first one.

The loop continues until all of the notifications have been read. Only then does the yellow ring go out. In retrospect, I should probably just have unplugged her and dealt with it in the morning, but it presented me with a computing problem to solve, and I can't resist that. Now I know how to turn off the yellow ring. I also know how to disable the notifications so I don't get more yellow rings – but this is something Alexa cannot do by herself. You have to do it with the Alexa app.

Now, by referring to Alexa as "she" and "her," I am indulging in anthropomorphism (but I talk to and have names for my bicycles, too). A great deal of computing went on during my exchange with Alexa, but that occurred because a team of programmers coded it. Alexa was bound by the extent of her programming. It is thus with all computers. Without the work done by programmers, they won't even turn on. In my 37 years of working with computers, not counting the 5 that have passed since I retired, I have yet to encounter a system that is not bound by the limits inherent in its programming. Those limits are ultimately defined by the laws of discrete mathematics, and those are not going to change even unto the entropic death of the universe.

The Star Trek franchise mostly got this right. The Enterprise computer could operate the machinery of the ship, and could answer questions posed by the crew, but it could do little else. They went a bit off the rails with Veejur (the first Star Trek movie) and especially Data (Star Trek: The Next Generation), but the Enterprise mainframe remained comfortably predictable. To the extent that Alexa, Siri, and other voice-interactive systems become more "intelligent," that will be the direction they take.

Now, I like Alexa. I have five Echo devices located throughout my house. She wakes me up to my favorite Internet radio station, she turns lights and appliances on and off, and plays an almost infinite selection of music on demand. As a person who has trouble getting around, this capability is a godsend. I do not, however, fear that she is going to hijack my brain and insert my body into a power plant any time ever.

This is supposed to be a blog about my writing, not technology, so it may seem that I digress a bit. I am, rather, just being long-winded. I have mentioned in prior blog entries that the authors who have most influenced me include the greats of science fiction's golden age. This leads to the question of why I am not writing science fiction. It would seem the most natural thing. My Spike Bike stories take place in a dystopian future, but this was just a backdrop I needed to give the stories a modicum of credibility. I was not writing about a future society, I was writing about the one I lived in.

The problem is that I am unable to sufficiently suspend disbelief when I write. All fiction requires suspension of disbelief, but not to the degree that science fiction demands. I bill myself as a writer these days, but I am still an engineer. As such, I solved real world problems using real world science. I tried to be creative with it, but I was always keenly aware of the constraints. I cannot ignore those constraints when I write, so I cannot bring a computer to life or make a ship fly through space faster than light. This does not cut me off from the genre entirely, but it does impose restrictions that I don't like.

As a kid who cut his literary teeth reading space opera, I realize I could never do it justice. A writer must be unbounded in the creation of a story, so that the story can go where it needs to go. Math and physics would bind me as such, so it's easier for me not to go there.