August 2019: Code Cave


  • Subscribers:
    2,334 🎉
  • Users:
    5,065 🕵️‍♀️
  • Pageviews:
    28,016 👀
  • Trends:
    1,847 📈
  • Insights:
    54 👁
Check out the live stats: Open Page


This month I scuttled back into my code cave 🦀. This was with the view to
making Trennd more robust and sustainable for the long term.
That's important, and I’ve definitely made a lot of progress on the backend towards that goal.
But at the same time, I’ve done zero marketing this month 🤦‍♂️.
The hype last month around the Hacker News launch and Twittersphere has naturally died down, along with the web traffic.
The mailing list count is still ticking up everyday though despite this, and we’ve reached a good new equilibrium of 100-300 visitors to the site everyday. Both are green flags to me.
(I’m not sure if green flags are actually a real thing. Emojipedia doesn’t think so… 🚩)
It’s way too easy for me to put marketing on the backburner and keep building in general. That’s where I’m most comfortable. Yet, I know full well marketing should be an ongoing crusade, alongside product improvement.
The most successful indie hackers I see around - they have systemized routines that enable them to continually output valuable content.
So my new goal is to figure out my own process to enforce continual, ongoing marketing and stay disciplined. If I have a weekly process that I have to stick to, this should slowly yield results and compound over time.
For instance, writing 1 quality report about a trend and sharing each week could be a good start. If anyone has any ideas or suggestions on this front too, I’m all ears! 👂


This was a quiet month. But there are a set of regular Trennd power users which is awesome to see.
I intend to make a ‘Trennd Seer’ leaderboard in the future which I think could be fun. But for now, seeing people come back regularly and add trends is a great feeling and potentially the beginnings of a community.


Though the events were lacking, the building was plentiful. Some of the bigger things were:
Better granular data:
Before now the 3 month chart data was weekly and so pretty crummy with only 12 data points. I’ve improved this though with daily data points and also added a 1 month time frame. You can really see the difference below.
I finally added search functionality! This was much requested and pretty essential to any directory kind of site, as people often want to check if something they think of is added already, or find similar trends.
Absolute search volumes:
All the trend chart data is on a relative scale as the source dataset is Google Trends. So you can see if something is trending relative to its past self. But it gives you no idea how big the trend is in an absolute sense or compared to other trends. Now with the monthly search volume figures added too, which you can also sort the trends on, this is possible.
Automatic trend classification:
Previously, the trend classifications (like spike/hyper/potential) were all based on my judgement (or the user’s that submitted that trend). But this depends on subjective opinions which differ across people. And even my own classification of trends changed depending on how I felt that day and how many coffees I’d had. 😝 But now, I’ve created an algorithm that classifies the trend based on the 5 year chart data. I admit, it’s not perfect yet. But it does a decent job and makes the trend classification consistent, automatic and updates daily based on new data. 💪 I also, renamed the “Hyper” trend type to “Exponential” as the exponent is what it’s based on.
I looked into automatic trend categorisation:
I’d hoped to automate this too, ie. putting trends into Food/Product/etc categories. But after digging into it I believe this would create more trouble than it’s worth. All the APIs I tested performed both poorly and had a huge number of categories. And categorizing is of no value if all the trends end up with a unique category or no category at all. For now, we’ll continue classifying by hand and potentially revisit this later.
Added affiliate product tile:
I added a sponsor tile at #4 on the main page. Right now it contains an affiliate link to Mangools which is a great tool that I use myself and recommend. I’ve also tried to keep it classy, so to speak, with it not being too garish and also disappearing once you start to sort/filter trends. Having this here allows me to test click through numbers which I can then show to potential sponsors. Maybe then we can start to make a small amount of revenue and pay some bills.

Next Month

The goal for last month was to make the site more robust, and I want to spend a couple more weeks doing so. This includes things like SEO and maybe even the comments system. This comments section has been “WIP” for too long!
But more importantly, as I mentioned at the beginning, I need to start a regular content marketing process, not build more and more features.
I should be regularly posting new, hot trends to Reddit, Twitter and other places. But not in a way that’s spammy. If it’s only a trend keyword, chart and link, then people will soon tell me to go away.
Probably not too politely.
So my current thinking is to start writing about specific interesting trends in depth and share that content every week/month. And the more I can do to build a system around this to ensure quality and consistency on my part, the better the long term results will be.
Until next month. 😘
And you can chat to me anytime: [email protected]

Trennd · Privacy Policy · Terms of Service