A few days ago, I was staring at an Amazon advertising campaign that had managed to collect more than 1,200 impressions without producing a single click. That campaign is gone. I rebuilt the advertising from scratch, changed the message, changed the targeting, and decided to stop paying attention to what Amazon thought I should bid unless the numbers actually made sense.

Now I’m starting to see what the new approach does, and for once, the news is not terrible.

My new BookBub advertisement has now been shown 150 times in Canada and has received three clicks. That gives it a 2 percent click-through rate, and those three clicks have cost me a total of $1.86, or about 62 cents each. Three clicks obviously do not make a bestseller, but compared with some of my earlier advertising experiments, three people voluntarily clicking on the advertisement feels remarkably close to progress.

The new ad says, “Twelve Houses Rule. One Was Erased.” Underneath that is “Military Sci-Fi • Ancient Warships.” It is a much simpler advertisement than what I started with, but it also gets much closer to the real hook of Sovereign Knight.

Humanity has been ruled by the Twelve Houses for centuries. Everyone knows there are twelve, and everyone accepts that there have always been twelve. Unfortunately for them, ancient machines have very long memories. They remember a First House, one that humanity somehow forgot, or was made to forget.

Canada appears to be responding to that idea. The United Kingdom has been quieter so far. The same advertisement has been shown there 48 times without a click, but that is nowhere near enough data for me to pronounce the British population collectively opposed to forgotten interstellar civilizations. For now, I’m leaving it alone.

Amazon has been more interesting. When I first launched the replacement campaign, almost nobody saw it. After the first couple of days, I had accumulated roughly ten impressions. Apparently Amazon and I disagreed about how much the privilege of showing an advertisement for an $8.99 ebook should cost.

Amazon had some suggestions. A few of those suggestions approached the price of a small lunch, so I declined.

Instead, I increased the bids only on a few searches that actually describe the readers I am looking for: military science fiction, military sci-fi, and military space opera. That finally did something. The Amazon campaign has now reached 24 impressions, up from only ten yesterday.

There are still no clicks and no money spent, but the important difference is that Amazon is finally showing the advertisement to people searching for the kinds of books I actually write. For now, I’m leaving both campaigns alone. That may be one of the harder lessons I have learned about advertising. When a campaign is clearly failing, change it. When you simply do not have enough data yet, leave it alone long enough to find out what is actually happening.

Meanwhile, Sovereign Knight itself continues moving forward. I am at roughly 14 books sold, three of those came through Barnes & Noble, and I now have my first five-star review.

Yes, I know the reviewer.

Yes, I am still counting it.

She read the book and liked it. I am not giving the star back.

So the independent-author scoreboard is looking a little better than it did a few weeks ago: about 14 books sold, three Barnes & Noble sales, one five-star review, three Canadian advertising clicks, and Amazon has finally decided that perhaps somebody should actually see the new advertisement.

It is not exactly galactic domination.

But I’ll take it.