I automated daily TikTok posting, and I still don't know if it's worth it

The groundwork for GradePod started back in 2018. From the very beginning I knew my audience would be teenagers, worldwide. For those of you who live with teenagers, we all know they spend a lot of time online, and it was my challenge to attract their attention, help them, and convince them to convince their parents to pay £490 to join my online course.

I know, I know. It’s a stupidly long sales funnel.

Anyway. You’d think I would have embraced YouTube and TikTok from the very beginning, but I didn’t, and here’s why. I just couldn’t be bothered.

I could not stay consistent on posting regularly and I wasn’t particularly interested in trying. I’m honestly not sure how valuable social media is in building a business anyway.

So I built long form, evergreen content that lasted. Hour long tutorials, 1,000 word blog posts, books. I figured if a teenager was motivated enough to engage with longer content, they’d be the kind of person who would want to join my course.

The gamble worked. GradePod hit #1 on every Google keyword related to IB Physics, I welcomed hundreds of students into the course, and it stayed that way for years.

Then two things shifted. I started to notice the decline of online courses in 2024 (that’s another entry) at the same time as my Google traffic started suffering. A lot of businesses are noticing that now, and it is down to Google answering the question itself rather than providing a list of businesses that match the keyword. It is a seismic change in how businesses get found. The other shift was attention. As short form swept everyone’s feed, my blogs, books and hour long videos gained less traction.

Here’s the problem. I still can’t be bothered to record a video every day. And I didn’t want one of those expensive, honestly not very good, SaaS products to make short clips for me.

So I worked it out myself.

I downloaded the .srt subtitle files for all my long form tutorials. Claude read the scripts and identified clip segments, 60 to 110 seconds, one learning objective each, and returned timestamps, hook text, caption and CTA for every clip. It then wrote three shell scripts: cut_topic_clips.sh, rename_[topic]_backgrounds.sh and composite_topic_clips.sh, all ffmpeg, all running locally on my MacBook. Those cut the right sections out of the video, renamed each clip, and dropped it onto a background I had built in Canva (1080x1920 template, video card at x=54, y=528, w=967, h=534). I exported the backgrounds, ran the scripts in order, and was left with 300+ short videos, each one teaching a single learning objective.

I’m still wrangling with posting them automatically. It’s simple enough on YouTube. TikTok makes it very difficult to attach an API, so I’m using Buffer for that. Honestly, right now, I upload batches of ten or fifteen at a time to both platforms and schedule one a day. I feel like the process could be automated better, but I can’t be bothered doing that either.

I’ll reiterate what I said at the start. I still don’t know if it’s worth it.

Yes, I have a daily posting schedule, and you can see that videos get around 300 to 750 views. Yes, I’m talking to the right audience.

But are the videos resonating with teenagers? Are they making their way to the website and buying?

I don’t know yet. I can’t see the relationship between daily posting and actual money in the bank yet. This is a work in progress and I’ll get back to you.

What I do know is that I’m posting daily on YouTube and TikTok, I’ve wanted to do that for a long time, and I couldn’t be bothered. AI made it possible. And there isn’t an obvious AI voice or AI generated avatar in sight. It’s all me.