64 Comments

Thank you all for the kind words and great questions—I see you! 🙌

I’ve got a packed workday ahead, but I promise to go through your comments and analyze your niche by this weekend!

If you’re just finding this and the 10 spots for my manual analysis are already filled, don’t worry—I’m planning to roll out an MVP in the next couple of weeks so you can try it yourself. And of course, the more requests I get, the higher it climbs on my priority list.

So if you're excited about it, let me know, and I'll try to launch it even sooner! 😊

Expand full comment

Just landed here and would love to try. Starting from 0

Expand full comment

I definitely would love to try this too, Jenny!

Expand full comment

Great insights! I'd love to test the web app! Here are my questions/takeaways:

1. I'm surprised at the proportion of ghost accounts.

2. How did you get the data? Did you just use the requests module or equivalent to scrape directly from urls?

3. How do you see a publication's tier?

Expand full comment

Thank you, Claudia! 😊 Your #1 takeaway was a huge surprise to me too! Now I know I’m at least ahead of that group.

I just use the requests module to retrieve data. Scraping directly from URLs feels like a gray area these days, so I figured it’s always safest to use requests.

The tier term never shows up anywhere in Substack’s UI, but it’s hidden in the request response. AI picked up on it and incorporated it into the analysis, otherwise, I might have never noticed it!

Expand full comment

Hey Jenny, would love to know more about the requests module and how to access it.

Could you let us know more please 👍💥

Expand full comment

Absolutely! I might put together a simple tutorial at some point—I’ll be sure to let you know when I do! 😊

Expand full comment

I second Raj’s ask on the ‘requests module’ - I’d love to be able to use automation for augmenting our list of women writing about AI and data here! And it would be great to discover my own rankings and get these insights for my own 2 newsletters. :)

Expand full comment

You can find them from the products section on my newsletter https://open.substack.com/pub/finntropy?utm_source=share&utm_medium=android&r=2023k5

Expand full comment

As a data nerd I'm interested in numbers and the stories they can tell us. I started my newsletter without a clear idea what I should be writing about. Only recently I stumbled on a few topics that resonate strongly with my audience, and experienced rapid growth.

I created this Substack Growth Analysis survey to understand better what writers really care about, see https://open.substack.com/pub/finntropy/p/how-to-grow-your-substack-newsletter?utm_source=share&utm_medium=android&r=2023k5

Expand full comment

So wonderful! Thank you for sharing this!

Expand full comment

These undocumented Substack APIs provide rich set of data that contain a lot of details on the newsletter, posts, notes and the audience. I've collected a dataset from over 15K newsletters in different bestseller tiers, with millions of posts and notes details, including engagement metrics. Some details are public and some only visible to newsletter creators, so that makes analytics more challenging.

I've created some Chrome extensions for writers who want to grab their metrics and save locally to Excel CSV files. I wonder what Jennie's AI assistant would find from those files?

Very interesting analysis, Jenny!

Thanks for sharing.

Expand full comment

Wow that’s some serious data collection! I wonder what my AI assistant would find from them too 😃 Now what is your extension?

Expand full comment

Here is the link to Gumroad https://finntropy.gumroad.com/

Expand full comment

Great article, really insightful. I don't have a newsletter but been toying with the idea. A mix of using AI tools & growing Vegetables!

As with other comments, would love to know more about how you built the agent and got the data. Thank you for posting

Expand full comment

Your idea of combining AI tools with growing vegetables sounds super interesting! Am I getting it right—are you literally using AI to assist with growing vegetables? I'd love to hear more! 🌱

As for your question, I didn’t build the agent myself, I relied on an AI agent (Cursor, to be precise) to do most of the heavy lifting. 😃

Cursor helped me fetch the data from Substack’s undocumented API, making the process much smoother!

Expand full comment

Correct 🙂 back in December I took photos of all my seed packets, got ChatGPT to convert to text/spreadsheet. I then experimented with ChatGPT/Claude/Gemini on creating a growing plan. Claude was by far the best so worked with that to create a google calendar for sowing seeds, a monthly image of my grow beds with growing plan based on UK weather cycles. I'm wondering how I can simplify all this and create an App to upload photos and document the journey. Photos could be things like pests and diseases for advice as well

Expand full comment

This sounds like an amazing plan! When you carry it out, please let me know, I would love to see it!

Expand full comment

Great work Jenny, loved it! I'm also super honored to see that my Substack Report inspired you to do such an analysis, thank you so much for the shout-out.

I'm curious to learn more about Tiers. What are the definitions, and how did you find out that part?

Expand full comment

Thank you Ciler! Your report is truly a masterpiece and incredible hard work!

Tiers is the most mysterious term for me as well, nowhere defines it, but it appears in the API response. So AI decided to pick it up and interpreted it as a way to gauge the quality of the publication.

I believe most of the people I interact with will fall within the Tier 2 range.

Expand full comment

Oh Yay, Jenny! My three keywords are recovery (from anything) align with true self, cult survivor (I write under the umbrella of how to recover from ANYTHING) I’d love to see how my newsletter stacks up, what tier it’s in, etc.! Thank you.

Expand full comment

Sounds great! I'm sending over the results and plots right away!

Expand full comment

Yasssssss

Expand full comment

I think I can explain at least some of the high follower ghost accounts: there are many already famous people (most press folks, many influencers, politicians), who set up an account and people immediately started following/subscribing even before they wrote anything. 2. They might have, at one time in the past, been more prolific, built up their followers then, and then stopped writing (I've seen this a lot).

I'm curious if there's a way to track not just how long someone has been posting, but their frequency per week of posting -- do numbers follow the people who post 2-3x a week more than those who post 1x a week, or 2x every month? Harder: does word count/length matter?

How did you discover your tier?

Finally, I'd love an analysis, because I'm vain like that. Or alternatively, if you create a website I think you'd get a huge follower bump just from all the people who would find value in this!

Three words

Privacy

Tech

Legal

Expand full comment

Thanks for sharing these insights—they make a lot of sense!

Your questions are getting tougher! There’s a dirty way to do it—basically searching for all published articles, sorting by time, and counting them. But so far, the API doesn’t provide a clean solution.

How did I discover my tier? It was actually hidden in the API response, Substack just never surfaced it in the user interface.

And yes, your analysis is done and headed your way soon!

By the way, your newsletter domain tricked my method, so I couldn’t find out your tier. Looks like I’ve got some homework to do! 😆

Expand full comment

Yeah, having ones own domain tends to trick Substack as well -- none of the marketing asset images ever work, and links sometimes get screwy. Computers are hard!

Expand full comment

Oh I never knew that even Substack is having problems with it, that’s indeed hard!

Expand full comment

Also I meant to add that this is such a cool thing! Thank you for developing it.

Expand full comment

Awesome analysis! A web app would be very cool. Here are 3 keywords related to my niche: self-publishing, book marketing, content creation. Thanks!

Expand full comment

Glad you enjoyed the analysis! I will send the analyzed results your way right away.

Expand full comment

Thank you!

Expand full comment
Feb 7Edited

That’s amazing! Love it! Thanks so much for sharing this! I asked that myself as a newbie here but gave up after some generic ChatGPT responses lol. I just started creating my account so will wait patiently for another 1-2 years to see some traction 😅

I’d love to see a behind the scenes tutorial how you worked with your AI agent on this ☺️

One thing I would say is you never know how you influence people and what impact you have - numbers don’t tell that 😉 keep up the good work!

Ah and my 3 keywords are: AI startup (of course), growth mindset and product-market fit.

Expand full comment

You are absolutely true: you never know how you influence people and what impact you have! Thank you for reminding me of this gem!

I would love to share more about how I worked with my AI agent. It is hard to describe in the writings, but once you see it, you'll know it.

Maybe in future, with enough interest, I can run some workshops, who knows :)

And your analysis is coming soon 😆

Expand full comment

Sign me up for the workshop 🙌

Expand full comment

And thanks so much for my analysis! That’s brilliant!

Expand full comment

Absolutely! I’ll be sure to let you know when I do! 😊

Expand full comment

Where do you find the tiers of newsletters?

Expand full comment

It is in the API response.

Expand full comment

Yep! They do have APIs, but they don’t maintain public documentation, which makes it hard for programmers to find. I’m compiling my experiences with all the Substack APIs I’ve used and will share it with you once it’s ready!

Expand full comment

There's a Substack API? Nice

Expand full comment

:D I don’t really understand the dog picture, but which AI do you use to do that niche deep dive, any step by step guide coming so we can do that to our substacks @Jenny Ouyang ?

P.S. I came here from @Claudia Faith ‘s beautiful community to find such a nugget of gold! I was wondering if people like when I, as an ex-media owner, expose the witchcraft behind news headlines using psychology. What’s your take? https://substack.com/@adrianborowski/note/c-92696758

Expand full comment

The dog captures me deep-diving into a world of niche data!

I used Claude for this (though honestly, any model would work). My goal is to make the process even simpler than a step-by-step guide, so anyone can easily get a sense of their own niche.

And I’ll definitely check out your shared note! :)

Expand full comment

Very interesting analysis ! Though I am too late to ask my own analysis 🤭

Expand full comment

Thank you for the comment! I will make sure to let you know when the MVP is out 😃

Expand full comment

Rolling out an MVP would be great! Thanks! ~~ I’m in this for the long haul. It’s scary, and there‘s the unknown, but it’s also fun. 🙂

Expand full comment

Absolutely! I will make sure to let you know when the MVP is out 😃

Expand full comment

Thanks for sharing!

One thing to add is that I know some people post only notes, especially like visual artists. That’s why they have a huge amount of followers without having a single post.

Expand full comment

That makes a lot of sense! Thanks for sharing your thoughts!

Expand full comment