Automating Native Ads with AI 🤖

Can AI supercharge your native ad game? 🤔

Nadim Kuttab has spent the last few years putting tools and processes in place to answer that question with a Yes. And he was gracious enough to put his entire toolbox on display at a recent Affiliate World event – so we’re sharing an inside look with you here.

The most challenging part about native ads is creativity, and AI can play a huge role in helping marketers and affiliates be more creative. Nadim and his team use a five-step AI-enhanced process to create powerful native campaigns at scale, and with the right methodology and tools, you can too.

1️⃣ Defining Your Audience

Defining or segmenting your audience starts with understanding who you’re trying to sell to. Why do people buy this product? What problem does it solve? Who does it solve that problem for?

When you’ve completed your own “brain dump,” you can ask ChatGPT for additional suggestions based on what you already know about your audience. 👇

It’s important to start with what you yourself know, however, so that you aren’t limited by the suggestions that your AI tool provides.

2️⃣ Creating Your Headlines

Hooks or headlines further segment the users that you want to find. If you want to reach women over 50, for example, you need to make sure that your output appeals to women over 50.

Nadim recommends creating hooks and headlines manually during a brainstorming period before iterating your ideas through ChatGPT prompts:

By using ChatGPT to develop specific variations of your headlines instead of doing that manually, this whole step will take minutes instead of hours.

3️⃣ Writing the Content

The most time-consuming part of native ads is creating content, and Nadim’s tests found that AI-written pages lost to human-written pages 100% of the time. That’s why human-adjusted content is crucial for successful native ads.

For the advertorial, Nadim asks ChatGPT to write 400-500 words that start with an introduction to a problem, present why the problem is relevant to the consumer, and then present a solution with a sense of urgency towards the end. This prompt helps ChatGPT understand the goal of the output. 👇

The structure of the content is always the same, and once you have the advertorial text, you can edit it in a text document or send it to your copywriter for review. Again, you know your product and your audience better than AI, and tailoring the text according to that knowledge is where your competitive advantage comes in.

4️⃣ Finding the Right Thumbnails

There are a lot of different places to get creative assets, including Taboola stock images that you can edit and re-upload or AI-generated images courtesy of DALL-E 2. Nadim also recommends Canva Pro for editing images because it’s easy and simple to drag and drop, adjust, and export images.

A good rule of thumb is to find 10 interesting images, either stock or AI-generated, to support the ad content. Sometimes AI images that look “uncanny” can be useful as scroll stoppers that get users to click.

5️⃣ Choosing the Right Mix + Upload

When you’re uploading native assets, you need to know your funnel backward and forward. Literally.

Nadim recommends starting from the bottom of your funnel and working your way up to the ads at the top. To do that, you upload one offer page to the platform, then 2-4 content pieces, and then 10 ads. Limiting each campaign to one offer page makes it easy to determine where the funnel is performing well and which variations can be cut.

With these assets in place, you can start testing variables, find your top performers quickly, and optimize for conversions based on results.