You’ve Got the Clicks – Now Make Them Testify

How to turn Facebook ad clicks into clues instead of noise

Clicks are rather unreliable witnesses. Ask them a vague question, and they’ll tell you almost nothing.

Interrogate them properly, however, and they can become surprisingly talkative.

That matters because testing is one of the engines behind a book launch. I wouldn’t go so far as to say a good testing program can make or break a book. But I’d come fairly close.

Think about the last time you saw an ad for a book and suddenly thought: I need that.

Maybe you loved the book. Maybe you abandoned it three chapters later and wondered what possessed you.

Either way, something about the presentation worked.

The cover caught your eye. The hook posed a question you wanted answered. The description made the story sound irresistible. Something stopped your scroll and persuaded you to take the next step.

For authors advertising on Meta, discovering what that something is can be immensely useful.

But you won’t discover it by throwing twenty random ads into Facebook and waiting for Meta to crown a winner.

There was a time when Facebook advertising advice focused heavily on targeting the right audiences. Find exactly the right interests. Layer the right behaviors. Narrow this. Exclude that.

The old joke went that there were no bad ads—only bad targeting.

Today, the ad creative itself does much of the targeting.

The image, hook, headline, and primary text signal which kind of person should stop and pay attention, while Meta’s increasingly automated delivery system determines who is likely to respond.

For an advertiser whose website can send purchase information back to Meta, this can work beautifully.

But authors selling books on Amazon face a complication.

We can send people from Facebook to Amazon, but we can’t install our own Meta pixel on the Amazon product page to let Meta optimize our campaigns around completed book purchases.

Meta can become very good at finding people who click.

Our job is to make those clicks come from people who might actually want the book.

And that is where systematic testing comes in.

The temptation is enormous.

Five hooks. Three images. Eight texts. Two headlines. Mix everything together, launch a small army of ads, and see what sticks.

I call this the spaghetti-against-the-wall school of advertising.

It can produce a winner. What it usually doesn’t produce is much understanding.

If one ad performs brilliantly and another crashes, what caused the difference? Was it the hook? The picture? The primary text? Some interaction among all three?

You don’t know.

So I prefer to make the clicks answer one question at a time.

If I’m testing hooks, the hook changes. The image, primary text, headline, audience, and other key settings remain the same.

Once I’ve learned something about the hooks, I can use the stronger ones to investigate images.

Then texts.

Then combinations.

And, as we discussed in AI, AI Everywhere… How to Keep Your Tests Pure, I also make sure Meta hasn’t helpfully changed a variable behind my back.

A clean experiment asks a clean question.

My testing usually looks less like a single giant tournament and more like a series of auditions.

First, several hooks enter the room.

The weak ones leave.

The stronger hooks return and audition with different images.

Then the strongest hook/image combinations meet different primary texts.

Eventually, I have a much smaller group of creatives that have earned the right to continue.

This saves money and, more importantly, builds knowledge.

By the time I reach the later rounds, I don’t just know that Ad 17 had the cheapest clicks.

I know something about why readers may have responded that way.

Click-through rate, or CTR, is usually one of the first numbers I watch.

It answers a useful question:

Of the people who saw this ad, how many were interested enough to click on it?

That matters.

If one hook consistently earns far more clicks than another under comparable conditions, it is telling you something. But CTR is only the beginning.

A high CTR does not mean this ad sells books.

It means this ad makes people curious enough to click.

Those are not the same.

Instead of simply declaring winners and losers, I find it more useful to classify early results as:

Winners — consistently strong enough to move forward.

Interesting signals — something promising happened, but I need more evidence.

Inconclusive — not enough information yet.

Underperformers — repeatedly fail to attract attention and can likely leave the audition.

The key word here is consistency.

One spectacular CTR can be luck. A hook that performs well across multiple images and texts is much more interesting.

Consistency is where the clicks begin to testify.

Before launching my thriller Nemesis, I started with five hooks, eight primary texts, and several possible images.

Two images gradually became particularly interesting. One was a cinematic nighttime view of Baltimore Harbor. The other was much more classical: a sword and the scales of justice, two symbols central to Nemesis.

The Baltimore image attracted broader interest and eventually became part of the book’s cover design.

The sword-and-scales image behaved differently. It wasn’t always the strongest image overall—but when paired with certain hooks and texts, it became extremely effective.

That distinction mattered.

If I had simply asked, “Which picture won?” I might have discarded useful information.

Instead, the testing began to reveal different ways into the story.

Three hooks survived the early rounds:

H3: She filmed what no one was supposed to see.

H4: A president murdered. A nation lied to. One woman holds the truth.

H5: She went undercover to stop a president. Then he died.

H3 served as the curiosity hook—the forbidden evidence.

H4 framed the book as a struggle between truth and power.

H5 delivered the narrative shock: a mission goes catastrophically wrong.

Three primary texts also kept giving me useful signals.

T5 emphasized plot and danger.

T6 emphasized the ordinary woman at the center of the story.

T8 emphasized secrecy, conspiracy, and the hidden world behind the visible one.

Now the results were becoming much more interesting than ad A had a 5% CTR and Ad B had 4%.

They were beginning to explain to me why different readers might want this book.

Here’s the simpler version of the map that emerged:

That was far more useful to me than selecting one universally “best” ad.

The book had several doors, and different readers wanted to enter through different ones.

H3, She filmed what no one was supposed to see, provides a good example.

People clicked. A lot.

The combination of that hook and cinematic imagery was so effective at creating curiosity that some people apparently thought they were looking at advertising for a movie rather than a novel.

Excellent CTR. Wrong attention.

So I adjusted the creative to make the product unmistakable, including adding A THRILLER NOVEL to the images.

The click rate became less spectacular—but the test became more meaningful. That is an important distinction.

The goal of testing isn’t to maximize CTR. It’s to learn which creative attracts the right kind of attention.

Once several rounds of testing are complete, I stop staring at individual ads and start looking for patterns.

Does one hook work regardless of the image?

Does one image need a particular kind of copy?

Do character-oriented creatives behave differently from plot-oriented ones?

Are different demographic groups responding to different story promises?

Does one combination consistently produce inexpensive clicks while another produces fewer but perhaps more qualified ones?

You are no longer asking which ad won.

You’re asking: What are readers responding to?

That question can inform far more than just Facebook advertising.

It can help refine your Amazon blurb.

Your book-page hook.

A+ content.

Newsletter copy.

Website messaging.

Sometimes even the cover.

That is why I like to start testing before launch rather than after everything has been carved in stone.

Systematic testing does not magically tell you which ad will sell the most books.

Before launch, it can’t. What it can do is replace random guessing with progressively better evidence.

Change one thing.

Watch what happens.

Keep the promising signals.

Discard the repeat offenders.

Test again.

Look for patterns rather than isolated victories.

Eventually, the clicks stop being a pile of numbers and start telling you something about the people behind them.

But there is one rather large complication.

Clicks only testify about what happened at the top of the funnel.

They can tell you who stopped. Who became curious. Who clicked.

They cannot yet tell you who bought the book.

And that is where our detective story becomes considerably more interesting.

Because a beautiful CTR can tell the absolute truth—and still lead you to a completely wrong conclusion.

That’s what we’ll investigate next in All That Clicks Is Not Gold.