Fascinating statistical analysis. It makes me wonder if same/different themes would emerge on LinkedIn/Medium, ie how Substack specific is this or does the general public all gravitate to the same topics at the same times.
Thanks for doing the setup, analysis, and distillation!
Those are great questions! Not sure about Medium, but LinkedIn would sue me into oblivion if I tried something similar with their posts/articles! Their TOS is super restrictive.
I was riveted by this data. Not shocked that politics drives the most engagement. Social networks figured this out a decade ago. I think that's one reason many folks are so angry these days, lol.
Everyone is politically divided yet constantly exposed to political content.
In any case, excellent article. Thanks for sharing it freely.
Thank you! 🙏 I’m going to collect feedback on the analytics feature before I decide what to do with it. In the meantime, I’m going to build out the curated collections a bit more and fulfill a few user requests (support for Spanish-language digests, digests with UTM links for tracking click-throughs, better mobile design).
A deep dive into data analytics so interesting. My polymathic autodidact learning is generously rewarded by engaging in Substack. The conversations one hungers to have are here. Your content appraisals quite worthy reading.
Amazing work, Karen! Glad to see StackDigest shaping towards such a robust tool with all these key data points!
I'm trying to build a data analytics dashboard tool as well, albiet for individual publications who want to analyse their posts in a nutshell. Would love your feedback on it once I have a basic POC ready for beta test!
Excellent job and very useful insights! You are building a great application for the Substack community.
As a newbie is Substack (7 articles so far), writing AI/ML technical content, I was also exploring Substack endpoints and built my own application for the newsletters I mainly like to follow, and I will write how I built it in a couple of weeks.
It just contains 30 newsletters on technical AI/ML content, but maybe you will find some inspiration in it to further develop your app, or add newsletters that you still do not have on the list.
One of the main problems I see in general is how to categorize/cluster the newsletters or articles themselves, as they are pretty diverse. AI/ML might be technical, product, news, politics, trends, ethics,... This was an issue for me, as I was interested mainly in technical content, and from the 500 newsletters you get from the endpoint, I ended up reducing it to 30 technical, though there are a few more.
All in all, thanks again for the great job, and keep pushing the app forward!
Yes, categorizing newsletters in meaningful ways can be challenging…and since AI and ML are non-deterministic, you may need to try a few times before you get a categorization framework you like.
Looking forward to checking out your app and your newsletter collection! 🙏
Substack grew from poached Twitter political writers so it's not surprising politics and the American reader and current news on the United States still dominates. Even how Substack is implementing Trending, displays how incredibly U.S. centric the app is. Which doesn't really reward contrarian independent journalism, but increasingly a kind of group think. For many of us the majority of our paid readers are from particular US states, which incentivizes us to write a certain way.
💯 The current discovery mechanisms are doing nothing to break people out of existing content bubbles. I’m thinking about using my data to map out the various Substack content ecosystems…but I worry that might end up becoming a roadmap to the echo chambers. 😆
My head = 🤯 - Wow. Thanks Karen - unbelievable insights! Business didn't really feaure as it's probably horizonal across leadership, AI & some of self improvement. The "who to....on Substack" is a common category when a new platform is in town - we can the initial period the "installation" era when people write and build products to help new users. So this one makes perfect sense & surprised it's not top. Anyone WOW ! 🧡
Given the randomness built in to clustering, it’s possible a Business topic might surface on a subsequent run…and agree the Substack-centered theme should almost always appear in the top 20 somewhere!
This is so good Karen. Have now signed up to beta (sorry for sleeping on it!) and can't wait to start using it to help shape my reading (and writing) habits.
An impressive achievement, Karen! A very well laid out format and commentary.
A few things struck me. First, I always wondered about the views on an article, and the comments statistically. Not that I have a ton of articles, but I just had one break 700 today and another 300 views and these are not recent pieces. I don’t post any notes, just comment on others work.
Secondly, a lot of folks are obsessed with subscriptions numbers. It would be an interesting to see if there is a correlation between subscribers/followers an article engagement. But I don’t know if you can pull that subscriber count data. And not that you need another challenge.
Lastly, my gut told me that real-time events are the best sellers, and the your data supports that. A couple of folks jumped on the Coldplay affair, the day after, and comments and likes were through the roof. The article quality, not outstanding to me, but people like to get their 2 cents in on the hype, I guess.
You’re very welcome! I’m looking into how I might enhance newsletter records with subscription numbers…I’m thinking results are probably very different for newsletters with huge (100K+) followings versus smaller ones.
Fascinating statistical analysis. It makes me wonder if same/different themes would emerge on LinkedIn/Medium, ie how Substack specific is this or does the general public all gravitate to the same topics at the same times.
Thanks for doing the setup, analysis, and distillation!
Those are great questions! Not sure about Medium, but LinkedIn would sue me into oblivion if I tried something similar with their posts/articles! Their TOS is super restrictive.
This is absolutely brilliant!! Thanks for doing this work and writing it up.
You’re very welcome! It was a fun project!
Karen,
I was riveted by this data. Not shocked that politics drives the most engagement. Social networks figured this out a decade ago. I think that's one reason many folks are so angry these days, lol.
Everyone is politically divided yet constantly exposed to political content.
In any case, excellent article. Thanks for sharing it freely.
:)
Thank you so much! 🤗
Love these analysis Karen! Would like to try out these features, and I now wonder what the next step you are going to push for!
Thank you! 🙏 I’m going to collect feedback on the analytics feature before I decide what to do with it. In the meantime, I’m going to build out the curated collections a bit more and fulfill a few user requests (support for Spanish-language digests, digests with UTM links for tracking click-throughs, better mobile design).
You are onto something big, I can tell 🥳
🤞
Well done and very insightful. Thank you.
You’re very welcome!
A deep dive into data analytics so interesting. My polymathic autodidact learning is generously rewarded by engaging in Substack. The conversations one hungers to have are here. Your content appraisals quite worthy reading.
Thank you! 🙏
Amazing work, Karen! Glad to see StackDigest shaping towards such a robust tool with all these key data points!
I'm trying to build a data analytics dashboard tool as well, albiet for individual publications who want to analyse their posts in a nutshell. Would love your feedback on it once I have a basic POC ready for beta test!
Awesome! Would be happy to test it! 🎉
Wow. Had to make pop corn to read this one! Great info
So glad you enjoyed it! 😊
Hi Karen!
Excellent job and very useful insights! You are building a great application for the Substack community.
As a newbie is Substack (7 articles so far), writing AI/ML technical content, I was also exploring Substack endpoints and built my own application for the newsletters I mainly like to follow, and I will write how I built it in a couple of weeks.
It just contains 30 newsletters on technical AI/ML content, but maybe you will find some inspiration in it to further develop your app, or add newsletters that you still do not have on the list.
https://substack.martindatasol.com/
One of the main problems I see in general is how to categorize/cluster the newsletters or articles themselves, as they are pretty diverse. AI/ML might be technical, product, news, politics, trends, ethics,... This was an issue for me, as I was interested mainly in technical content, and from the 500 newsletters you get from the endpoint, I ended up reducing it to 30 technical, though there are a few more.
All in all, thanks again for the great job, and keep pushing the app forward!
Yes, categorizing newsletters in meaningful ways can be challenging…and since AI and ML are non-deterministic, you may need to try a few times before you get a categorization framework you like.
Looking forward to checking out your app and your newsletter collection! 🙏
Cool! I hope you like it. Feel free to ask if you have any questions!
That s really cool
Thank you! 🤗
Substack grew from poached Twitter political writers so it's not surprising politics and the American reader and current news on the United States still dominates. Even how Substack is implementing Trending, displays how incredibly U.S. centric the app is. Which doesn't really reward contrarian independent journalism, but increasingly a kind of group think. For many of us the majority of our paid readers are from particular US states, which incentivizes us to write a certain way.
💯 The current discovery mechanisms are doing nothing to break people out of existing content bubbles. I’m thinking about using my data to map out the various Substack content ecosystems…but I worry that might end up becoming a roadmap to the echo chambers. 😆
Loved this piece of analysis. Thank you for sharing
You’re very welcome! 🤗
My head = 🤯 - Wow. Thanks Karen - unbelievable insights! Business didn't really feaure as it's probably horizonal across leadership, AI & some of self improvement. The "who to....on Substack" is a common category when a new platform is in town - we can the initial period the "installation" era when people write and build products to help new users. So this one makes perfect sense & surprised it's not top. Anyone WOW ! 🧡
Given the randomness built in to clustering, it’s possible a Business topic might surface on a subsequent run…and agree the Substack-centered theme should almost always appear in the top 20 somewhere!
🙏💜
This is so good Karen. Have now signed up to beta (sorry for sleeping on it!) and can't wait to start using it to help shape my reading (and writing) habits.
Thank you so much…and would love any feedback you have on the tool! It’s still a work in progress! 🙏
An impressive achievement, Karen! A very well laid out format and commentary.
A few things struck me. First, I always wondered about the views on an article, and the comments statistically. Not that I have a ton of articles, but I just had one break 700 today and another 300 views and these are not recent pieces. I don’t post any notes, just comment on others work.
Secondly, a lot of folks are obsessed with subscriptions numbers. It would be an interesting to see if there is a correlation between subscribers/followers an article engagement. But I don’t know if you can pull that subscriber count data. And not that you need another challenge.
Lastly, my gut told me that real-time events are the best sellers, and the your data supports that. A couple of folks jumped on the Coldplay affair, the day after, and comments and likes were through the roof. The article quality, not outstanding to me, but people like to get their 2 cents in on the hype, I guess.
Thank you for sharing!
You’re very welcome! I’m looking into how I might enhance newsletter records with subscription numbers…I’m thinking results are probably very different for newsletters with huge (100K+) followings versus smaller ones.
Yeah, most likely.
Wow! Super helpful data points. A lot of work went into this. Thanks for sharing.
You’re very welcome! 🤗