10/15/2026
What happened when I used AI, psychology, and seven unconventional advertising concepts to sell my thriller series?
Advertising books is one of those things that sounds simple until you actually try doing it.
Write a good book. Create an interesting advertisement. Put it in front of people who like reading. Collect royalties.
If only.
I've been advertising my Ryan Weller Maritime Thriller Series on Facebook for years. Some ads work. Some don't. Occasionally, an ad that looks like a winner turns out to be a complete waste of money. Others that don't seem particularly remarkable keep bringing in readers month after month.
After sixteen books in the series, I've accumulated a fair amount of advertising experience, although I wouldn't claim to have figured out Meta's algorithm. I'm not convinced anyone has.
Recently, I decided to try something different.
Rather than continue developing variations of traditional thriller advertisements, I wanted to explore the psychology behind what makes someone stop scrolling through Facebook.
And I had an enthusiastic research assistant willing to help.
ChatGPT.
The Theory: Interrupt the Scroll
The premise was straightforward.
People scroll through Facebook almost automatically. They're looking at vacation pictures, political arguments, funny videos, family updates, and whatever else the algorithm decides they need to see.
An advertisement has a fraction of a second to interrupt that behavior.
Traditional book ads usually rely on familiar elements: dramatic artwork, recognizable genre imagery, a compelling headline, and perhaps a book cover.
But what if we approached the problem differently?
What if we designed advertisements around the psychological mechanisms that interrupt habitual scrolling?
ChatGPT and I developed a matrix of advertising concepts, each intended to trigger a different response.
Instead of asking what a thriller advertisement should look like, we asked what would make someone stop and pay attention.
It sounded promising.
Seven Ways to Interrupt Someone's Day
We developed several distinctly different approaches.
The Contrarian Statement
Challenge something the reader believes. Make them stop long enough to consider whether they agree.
One concept centered on the idea that the ocean doesn't forgive mistakes. Another argued that thriller heroes who never screw up are boring.
The goal was to provoke an immediate reaction.
The Micro-Story
Tell a tiny story that raises a question without answering it.
For Dark Water, we used a situation involving Ryan Weller, a cartel, and a photograph distributed to gunmen in Tampico.
The advertisement ended with Ryan showing up anyway.
Would curiosity make someone click?
The Confession
Make the advertisement resemble an ordinary personal Facebook post.
I wrote about creating Ryan Weller, a former Navy EOD technician turned commercial diver, and why I prefer a capable but imperfect protagonist.
Rather than selling the book directly, the idea was to create a connection between author and reader.
The Specific Reader Callout
Address a particular type of thriller reader.
Someone who would rather be on a boat than in a boardroom. Someone who enjoys diving, shipwrecks, pirates, cartels, and characters who know what they're doing.
This approach was designed to make the right reader recognize themselves immediately.
The Fake Ordinary Content
Create something that didn't initially look like an advertisement.
We built a Google-style search screen showing someone looking for maritime thrillers, realistic diving, and books similar to Clive Cussler's adventures.
The theory was that familiar, ordinary content might slip past the brain's automatic resistance to advertising.
Unexpected Humor
We used a meme about someone intending to read one chapter of a Ryan Weller thriller and still being awake at four in the morning.
Funny. Relatable. Familiar to anyone who has stayed up too late with a good book.
The Argument
Start a debate.
Our advertisement declared that a hero who never screws up is boring, then asked readers whether they agreed or disagreed.
Comments, we reasoned, might increase engagement and help the advertisement circulate.
Each concept had its own creative treatment and accompanying copy.
We weren't simply changing headlines or swapping photographs. We were testing different reasons people might stop scrolling.
And then we put the ads to work.
The Good News: People Responded
The advertisements attracted attention.
People liked them. People commented. Some of the concepts generated exactly the sort of engagement we'd hoped to create.
From a psychological standpoint, the experiment appeared to be doing something right.
We had found ways to interrupt the feed.
Unfortunately, I wasn't running a psychology experiment for academic credit.
I was spending money to sell books.
And that's where things got interesting.
The Bad News: Attention Doesn't Pay Royalties
I track my advertising performance against sales and Kindle Unlimited page reads.
For anyone unfamiliar with Kindle Unlimited, Amazon pays participating authors based on the number of pages subscribers read. The payout varies, but it amounts to a fraction of a cent per page.
That means a few hundred page reads don't generate much revenue.
On October 14, my tracking sheet showed the following cumulative results for the newer advertisements:
| Advertisement | KU Page Reads |
|---|---|
| Photograph | 1,178 |
| Mistakes | 103 |
| Reader Callout | 12 |
| Confession | 12 |
| Debate | 0 |
| 4 AM Humor | 0 |
| Native Feed | 0 |
These weren't perfectly controlled tests. The ads started on different dates, had different amounts of delivery, and weren't all given equal time or money. The figures also don't capture every possible future read or sale.
But the early results were hardly encouraging.
Even the best of the newer advertisements, Photograph, had produced only 1,178 recorded page reads.
At an illustrative Kindle Unlimited rate of $0.0045 per page, that works out to approximately $5.30.
That's not a particularly exciting return when you're paying Meta to deliver the advertisement.
Meanwhile, several of my older, more traditional thriller advertisements were still running.
Here's what their cumulative page-read totals looked like on the same date:
| Advertisement | KU Page Reads |
|---|---|
| Reason to Hunt | 17,055 |
| Boat Diver Try + Trouble | 6,412 |
| Boat Diver Meet No Trouble | 5,053 |
| Sunset Try + Trouble | 4,644 |
| Story to Tell | 3,932 |
Those ads had been running longer, so the totals aren't a fair head-to-head measure of efficiency. Comparing profitability requires accounting for advertising spend, sales, and the time readers take to work through a book.
Nevertheless, the contrast was difficult to ignore.
My conventional advertisements were still generating reading activity.
The new ones were mostly generating attention.
And attention doesn't pay the advertising bill.
So What Went Wrong?
I went back to ChatGPT with the results.
Its first attempt at interpreting my tracking sheet was spectacularly wrong.
It confused the columns, misunderstood which advertisements were generating page reads, and confidently explained conclusions that weren't supported by the numbers.
After I corrected it, we returned to the actual question.
Why weren't people responding to the psychological disruption?
Eventually, we arrived at a distinction that should have been obvious from the beginning.
Getting someone to stop scrolling is not the same as making someone want to read a book.
Consider the debate advertisement.
"A HERO WHO NEVER SCREWS UP IS BORING."
That's a statement people can agree with, disagree with, or argue about.
But does it make someone want to spend the next ten hours reading Dark Water?
Not necessarily.
The humor advertisement presented a similar problem.
Someone can laugh at a meme about staying awake until four in the morning reading a thriller without having the slightest interest in buying that particular thriller.
The confession might create a connection with me as an author, but a personal connection doesn't automatically translate into interest in Ryan Weller's adventures.
We had designed advertisements to provoke reactions.
We hadn't consistently given people a compelling reason to read the books.
What Traditional Thriller Advertising Gets Right
Look at the imagery commonly associated with action-adventure fiction.
A shipwreck lying in deep water.
A diver descending into darkness.
A helicopter hovering over a vessel.
A sailboat alone on an empty ocean.
For someone who enjoys maritime adventures, those images communicate something immediately.
Danger. Mystery. Exploration. Suspense.
More importantly, they suggest the experience waiting inside the book.
A reader who enjoys Clive Cussler-style adventures doesn't necessarily need an advertisement to challenge their beliefs about fictional heroes.
They want an adventure.
They want to discover something, solve a mystery, escape danger, or follow a character into a situation they'd never willingly enter themselves.
And they want to know that the author can deliver.
That's what my traditional advertisements have generally attempted to communicate.
The psychological-disruption experiment often moved away from those expectations.
In trying to create something that didn't look like a book advertisement, we sometimes removed the very elements that make a book advertisement appealing.
Did AI Lead Me Down the Wrong Path?
In a sense, yes.
ChatGPT was very good at generating theories, suggesting psychological triggers, and developing creative concepts.
It could explain why a particular advertisement might interrupt someone's scrolling behavior.
What it couldn't establish was whether that interruption would translate into profitable book sales.
And it was entirely too enthusiastic about the concepts.
That's a weakness worth recognizing when using AI as a marketing partner.
An idea can sound brilliant in conversation. The reasoning can appear perfectly logical. The creative can look professional.
None of that means the advertisement will work.
AI can help develop and test hypotheses, but it doesn't replace actual market results.
I also bear responsibility for the experiment. I liked the premise, approved the concepts, and spent the money.
That's how testing works.
Sometimes you discover what succeeds.
Sometimes you pay to discover what doesn't.
What I'd Do Differently
I'm not convinced psychological disruption is useless.
Quite the opposite.
I still believe understanding why people stop scrolling is important.
But the interruption needs to be connected to the reason someone would want the product.
For my books, that means combining attention-grabbing creative with recognizable maritime adventure.
An unusual underwater photograph might interrupt the feed.
A dramatic situation involving a shipwreck might create curiosity.
A compelling line of copy might make someone wonder what happens next.
But those elements also need to communicate what readers will experience when they open Dark Water.
The goal isn't to collect reactions from people who enjoy a clever advertisement.
The goal is to find people who want to spend time with Ryan Weller.
There's another practical lesson here: advertising experiments need clearly defined success criteria.
Likes are nice.
Comments are encouraging.
Click-through rates can be useful.
But none of those numbers tells me whether the advertising is profitable.
For that, I need to know what I spent and what the resulting readers earned, including sales and Kindle Unlimited royalties.
And because I'm advertising a sixteen-book series, the value of a reader may extend well beyond the first book.
That's where readthrough becomes important.
A reader who starts Dark Water and continues through the series is worth considerably more than someone who reads a few pages and moves on.
Finding those readers is the entire point.
Back to the Drawing Board
For now, I'm putting more emphasis on the traditional adventure-driven advertisements that have been producing results.
That doesn't mean returning to the same handful of images forever.
There's plenty of room for creativity in maritime thrillers. Shipwrecks, salvage operations, diving, pirates, smugglers, dangerous waters, and complicated characters offer an almost unlimited supply of possibilities.
The difference is that the next round of creative will begin with the experience I'm trying to sell.
Not with a psychological trick.
And I'll continue testing, because advertising never stays solved for very long.
An advertisement that works today can stop working tomorrow. Audiences change. Algorithms change. Readers get tired of seeing the same thing.
The only way to know what works is to put it in front of people and measure the results.
This experiment cost me money, but it also taught me something useful.
You can win someone's attention without winning their interest.
And in book advertising, interest is what eventually turns into readers.
The ocean doesn't forgive mistakes.
Apparently, neither does Facebook advertising.