Face or Faceless?
Might be one of your burning questions.
One isn’t better than the other. Both are powerful in their own way.
If you decide to go with the latter, this email digs into the faceless options you have in front of you.
The “easy” route might be to use one of those dozen SaaS tools that generate “faceless” content… but you’ll be stuck with the AI-style formats they force on you.
Also, they suck.
You can do better than that, easily.
Here are some of the best hacks we’ve done, seen, or come across + tips:
Slideshows:
Let’s start with slideshows. The best use cases are:
Two ways to find images: Pinterest or GPT-4o.
Actually, a fantastic hack to avoid worrying about copyright is to download Pinterest images and run them through GPT-4o with a small prompt tweak.
Something like:
“Make this person more tanned.”
It should change the person while keeping the same image style and making it look realistic.
As with every piece of content, the hook is the most important part.
Make them in bulk. Repeat hooks.
To source hooks, check out nearly 100,000 of them extracted in our Social Growth Engineers resource section (or just borrow from your competitors).
This is perfect for both hook/format testing and scaling a working strategy.
Let’s assume I’m making a carousel for a hair analysis AI app:
There you go. Infinite variations.
Now, using an image API, you can automate this with just a few lines of code.
Once you’ve created the next variation, go back to Figma and line everything up.
Add the same hook again, export, send…
That’s slideshows. Pretty easy.
A quick repurposing hack for IG is to convert them into videos with each frame at 1 or 2 seconds.
Clipping:
Clipping is the process of chopping and remixing existing content. Through this process, you can make a nearly infinite number of videos.
The best route is usually to either hire an army of clippers/posters or bring on a full-time editor.
You’ll provide them with a bank of video assets and guidelines on how to create “viral” content.
In Cheaterbuster’s case, the app used to provide a large asset bank, video guidelines on how to create the content + learnings on what performs best.
Operations are not negligible here. You’ll likely need to create your own clipping Discord server.
Find clippers (on other Discords, IG, Reddit, etc…) and onboard them with a fixed CPM.
They’ll create the TikTok and IG accounts on their side and manage them.
This model works great because if they don’t get views, you don’t end up paying them. And you define the CPM.
Clipping can be competitive and hard to scale in-house.
But if you can build up your own network, you’ll have a strong ongoing cost effective distribution channel.
Repurposing:
Repurposing content from other platforms – or your own content into a different format or localized version – can be a huge hack to boost output volume.
Look at this app. They split the screen half-and-half with a repurposed meme at the top and a hook in between:
Some could argue it’s also “clipping.” And it is, or at least a mix of both.
Rayz is another good example here. They repurpose an existing video into a new one with a CTA at the end:
Programmatic:
This is often a mix between clipping and repurposing, or now, generating partial frames with AI and stitching them together programmatically.
I’m using MoviePy myself to generate these. I create transitions with just a few lines of code:
For the app I’m growing now, I’m using a combo of Pinterest + GPT image API + MoviePy to create animations.
Programmatically creating bulk video content, through animated images, AI, or AI avatars, is a powerful way to scale aggressively.
That said, this technique should only be used after you’ve found content–market fit.
–
I hope this gave you some fruitful ideas.
There’s so much you can do with creative AI image/video generation.
The scalability it offers with programmatic video gen is going to unlock tons of new faceless content opportunities.
Time to create bulk faceless content like you’re social growth royalty.
Cheers,
Guillaume @ The Social Growth Engineers
Leave a Reply