Couple years’ back, TikTok was all the rage in the UK. It was growing massively, and people were saying it would take over the Internet. Facebook, Instagram? Obsolete. Google, YouTube? Why bother spending 20 minutes looking for an answer when you can get it in 20 seconds?!
New York Times article from 2022
I wanted to capitalize on that momentum. So, I set out to quickly launch something.
The plan
I decided to build a service that surfaces emerging trends on TikTok. I figured it’s something that would be popular with a range of audiences - social media managers, influencers, news journalists.
Saw a few people asking about this kind of intel, too. And I spotted a large enterprise in this space. Evidence that people are paying for this type of product.
Building
TikTok had a web page where it showed the most popular hashtags over the past week. I set up a daily Python cron job to fetch that data, for the US and the UK, and perform some analysis to predict which of them will keep surging.
I decided to go with a less conventional framework for my web app: Astro in server-side rendering mode. I loved using the framework for my statically-built blog sites, and I wanted to see how its SSR capabilities fared. It worked like a charm.
Le beautiful app
Validation
I posted about it on my LinkedIn account. A friend asked me to extend data coverage to Pakistan, he was thinking of using it for his father’s (political influencer) social media accounts. I obliged, it was a tiny change; but it didn’t lead to anything. Another person wanted to join as a business cofounder - I told him to find a few customers to help validate this idea, he never did.
I reached out to social media agency owners. Response rate was poor, but I did manage to have brief exchanges with a couple of individuals. My tool didn’t seem to catch their interest.
One owner did ask me about another kind of tool - one that bulk messaged TikTok influencers about affiliate partnerships. I thought there’s potential, but it would’ve required a big pivot, and an even bigger investment to deliver a prototype. So I passed on it.
Relevant extracts from conversation
I reached out to news journalists, too. I figured that maybe this data would help them uncover stories that people want to read about. Conversations were started, but they ended shortly afterwards.
Lastly, I targeted influencers by posting on Reddit.
They were very well-received. But I stopped after a month, as I realised that this isn’t a good strategy to grow your TikTok profile. You can’t be chasing every trend, you need to have a consistent theme.
There was one final angle I thought about exploring - social media monitoring for companies worried about reputational damage. However, I realised how hard it would be to build something that integrates with TikTok, and I didn’t know anyone in that space.
And so, after a couple of months, I gave up.
Learnings
I think the biggest mistake was that I went into a domain that I was totally clueless about. I didn’t even have the app installed on my phone before!
But, I don’t regret it. I was pretty efficient with my time investment, I learned a lot about TikTok, and tried out Astro in SSR mode.