Why No Monthly Ad Updates?

Someone asked why I haven’t done a monthly ad results post in a while, and the reason is that starting in October of 2025 I mostly stopped using Facebook ads.

Why? Facebook ads became drastically less effective as Meta added more and more badly designed AI targeting features to the ad platform.

Now, why is this a bad thing? Am I refusing to use Facebook ads because of my frequently stated dislike of generative AI? No. The reason is that the AI targeting doesn’t work nearly as well as the old system and ends up being a waste of money because the AI targeting is too broad and random.

This is bad because the key to successful online advertising is to narrow your audience as much as possible.

What do I mean by this?

Perhaps an example will help.

Let’s say, for example, you sell motorcycle parts online. Obviously, when purchasing motorcycle parts, you need to match up with the exact model and year, otherwise you’re going to run into trouble. Say you have a lot of parts for the Harley Davidson V-Rod Muscle motorcycle model from 2009, and you want to sell them online. The best way to do that would be to target your ads as specifically as possible.

In the US, at least, motorcycle owners tend to be older and wealthier, and are more likely than not male. So you would want to target your ads to men 60 and older who have expressed interest in Harley-Davidson motorcycles, and potentially in that specific model from 2009. Facebook used to let you target audiences to that level of granularity.

However, in the last few years Meta has jumped on the generative AI bandwagon along with nearly every other large technology corporation, It’s followed the same trajectory as every other company that has used AI. The CEO gets a brainwave that AI is The Inevitable Future, and doesn’t want to be left behind. Employees are told to implement as many AI features as possible and to use AI for everything. That inevitably becomes part of the Key Performance Indicators on their performance reviews, so the employees dutifully stuff badly-designed AI features into everything in hopes of keeping their jobs the next time the stock price drops slightly. (If you’ve wondered why AI crap seems to be stuffed into everything these days, that is why.)

Gradually Facebook has eliminated the ability for granular targeting like I’ve discussed above in favor of their Advantage AI system, which does the targeting for you.

There are two major problems with this. First, Advantage is a black box. You don’t really see how it works, and it isn’t well-documented anywhere online. Second, the targeting is way too broad. Advantage is really good at getting a lot of clicks on your ad, but this is the important part: the clicks on the ads don’t usually convert to sales. Which is better, an ad that gets 10 clicks but results in 5 sales, or an ad that gets 50 clicks and converts to one sale?  From Facebook’s and Advantage’s point of view, the second is better, but it’s not better for the advertiser.

Like, to return to the motorcycle parts example, say you’re trying to sell Harley Davidson V-Rod Muscle 2009 parts, and you try to run an ad with Advantage doing the targeting. You can’t specifically target male motorcycle/Harley owners over 60 years of age. No, you have to mention “motorcycle” and “Harley parts” in the ad text, and hope the system steers it correctly. What will happen is that the Advantage system will show the ad to a bunch of people who have no interest in Harley parts. Honda motorcycle owners will see it, as will Yamaha owners, and probably people who own mopeds and e-bikes as well.

In certain situations this could work. Let’s say you have a giant warehouse of motorcycle parts which includes Honda and Yamaha parts, so there could be some overlap. If you have a big enough inventory of motorcycle parts, a broad targeting reach might work.

The trouble is that books are very, very specific products, and targeting outside of genre lines never works. Like, if you’re writing dark fantasy romance and the system shows your ad to someone who only reads World War II historical novels, it’s not going to work at best and at worst the reader will become annoyed. Even for someone like me who has like 160 fantasy novels, it wouldn’t work. I could try advertising to “fantasy” novels, but CHILD OF THE GHOSTS, FROSTBORN: THE GRAY KNIGHT, and CLOAK OF DRAGONS are all very different books in very different sub-genres of fantasy. So I have in essence a giant warehouse of books, but books are specific enough that trying to use Facebook Advantage to advertise them doesn’t work.

So now I don’t really use Facebook ads. I did try again in June to boost HALF-ELVEN THIEF OMNIBUS ONE before its Bookbub deal, and the results from the Facebook ads were in the bottom half of mediocre at best, and outright useless at worst. (The Bookbub Featured Deal itself, by contrast, was excellent.) When I have a new book out and I do the “I published a book!” post on Facebook, I’ll use the boost function, which shows it to more people based on my audience, which does increase engagement and isn’t too expensive.

But for the most part I don’t really use Facebook ads any longer.

That said, I do a LOT of Amazon and Bookbub ads, way more than I did before Facebook ads became unusable. I’ll typically run about twenty to twenty-four short Bookbub campaigns a month advertising my permafree books on the non-Amazon stores. While I am not sure of the exact number, I think have something like 550 Amazon ads campaigns running across about thirty-five different titles on Amazon. (Fear not, this is less expensive than it sounds.) I haven’t posted about it because it’s a ton of data, and I haven’t figured out a way to do so in a concise way that isn’t overwhelming or a giant wall of math and text.

That said, there should be a way to do it, at least for the Amazon ads. Generally Amazon ads are way more effective with Kindle Unlimited books and nonfiction, and Bookbub ads are very effective for the non-Amazon stores, especially for permafree books like FROSTBORN: THE GRAY KNIGHT and CHILD OF THE GHOSTS.

So that is why I haven’t done an ad review roundup post in a while. Once I work out a good way to do it with Amazon/Bookbub ad data, I will probably start posting them again.

-JM