Academy of Wayward Authors: Would I Trust Meta or Amazon to Run My Ads? The Limitations of AI-Optimized Technology

Advertising with Meta and Amazon

I get the emails just like you.

Emails declaring the machine god will do a better job than me regarding audience acquisition.

Could it be that easy?

After spending this much time at the Academy, I hope you question what you read. Here and elsewhere.

Despite your doubts, the writer seems confident in what they say. Additionally, author clients attest to the astounding results, claiming that nothing worked until now.

It is the latest craze, amplified by the buzz around AI. However, with so much noise and uncertainty, desperate people may be tempted to buy into something that seems like magic beans.

pexels pixabay 533446

I’m going to call bullshit.

It doesn’t matter which platform we’re talking about. They simply optimize what you identify as important. These platforms act as amplifiers that will do your bidding if they see a profit, but the payoff is for them, not you.

Let me explain this further.

First, this all falls apart if someone reads your book and finds it uninteresting. Therefore, most advertising programs collapse.

Next, your advertising account has discrete information, such as your pixel and various ads. For Amazon, there is your author name, ASIN, and various ads.

Behind the scenes, data is being attached to each of these items, and when you look at this as an algorithm, we see that these are optimization loops. The system looks to get you more of something—impressions, clicks, and conversions.

Machine learning only looks to optimize. It will be better and faster in future iterations, but it’s working from the constraint of what it has learned. Garbage in, garbage out. Let me give you an example.

One of the prevailing methods of direct sales marketing suggests you run ads for a free book. Then, after the sign-up that gets you this prospect, you immediately offer a one-time discount on a book bundle.

The money from the bundle sale should then be invested back into ads to scale up.

Sounds good, right?

Since this is a direct model, you can use sales-based conversion that allows Meta to collect data and optimize the sales conversion. This is a powerful tool, and when done right, it can optimize well. Here’s the problem…

What are you optimizing for?

If you just let Meta find people on its own, it will optimize on people who will buy quickly when offered a discount. You further optimize by spending more on those ads that attract those conversions.

They post: I made $50,000 this month selling directly. I’ve never had such success! They forgot to mention that they spent $30,000 on ads.

Did they attract their ideal client?


pexels porapak apichodilok 346752

They are training the machine to find one-click bargain buyers and are building a list of such people. Can we nurture them to buy books directly at full price?

What if you start with your ideal client? Figure out who they are, what they like, and where they hang out.

“But Joe, I don’t know where to find them.”

Exactly. If you don’t, why would Amazon or Meta be able to do it? Maybe they don’t exist on those platforms.

This is where you can flip the script with a nurturing and retention system. If you are trying to build a direct store of readers who buy books at full price upon release, you can design your system to reward and highlight that behavior.

You should also end up with other segments of customers, like those who don’t pay full price. You can sell to them later when you decide to discount.

Now that you have customers who behave in a certain way, you can activate them and build look-alike audiences that you know are like the customers you seek.

Another approach is to use an existing segment as a seed audience. If you already have a list and can segment a small group that represents what you are seeking, however small, you can build on this base. This is best done through repeated launches, where you can drive that audience’s behavior into the purchasing systems of platforms without the clutter and confusion of other traffic.

Are you using the sale of a full-priced book as that key performance indicator or a sale that is easier to convert? Whatever you tell these systems to optimize, they will.

Here’s the big takeaway: There will always be hacks for quick sales. I don’t want hacks; I want to keep tipping cumulative advantage towards me. I haven’t found anything with the reliability or repeatability of identifying the ideal audience and deliberately nurturing them to know, like, and trust a good author.

Once I have that, I can use the market to amplify the results.

P.S. The answer to the opening question is that I don’t trust them, but I do let them optimize my objectives. When you know what you want, they can be great at finding more readers like the ones I seek.

Read: What Does It Mean to Belong? Why Community Is the Backbone of Your Business