Confessions of a Data Addict
A chaotic but organized desk. A baby monitor sits next to a trading terminal. The screen on the left shows a sleep schedule spreadsheet; the screen on the right shows a candlestick chart. A half-drunk coffee balances on a stack of notebooks. The aesthetic is "Mad Scientist Dad.
A confession for you, I have (he said in best Yoda voice). I don’t just trade based on data. I live based on data.
When my son was born, I didn’t read the parenting blogs. I’ve lots of younger cousins so I had experience, but I also built a spreadsheet. I tracked every ounce of milk in and every diaper out. I logged wake windows to the minute and could tell you within a 30 minute window when he was going to poop with a 17 minute standard deviation.
When he got old enough to try escaping his crib, I didn’t just buy a baby monitor like a normal person. I rigged two sheets of aluminum foil and an ESP32 microcontroller to his mattress. I wired it so that the moment pressure lifted -- meaning he was making a break for it -- I got a ping on my phone before his feet even hit the floor.
Please understand: I am not normal. Nerds even make me sit alone at lunch. In some sick weird way I believe if it can be measured, it can be mastered.
So naturally, when I started this newsletter 90 days ago, I didn’t just write. I tracked.
The Anomaly
It’s been almost 3 months since I joined Substack now, so I’ve got enough data to audit my writing the same way I audited my son’s escape attempts.
Here is what I found: We are currently converting free readers to paid subscribers at a rate of 4.9%.
For context, the industry average is usually between 2% and 3%.
The Pricing Signal (Or: Why I Might Be Wrong)
There is an issue, however. I know how to analyze stock data; it’s not new to me. I obviously want to assume I give off an extra aura of awesomeness, but the data nerd (and mirror) in me knows better.
A 4.9% conversion rate is a massive signal. It’s not just about me being charming (or maybe it is!). It’s about the market finding value.
In the world of financial intelligence, people pay thousands for terminals and hundreds for proprietary signals. We are operating at a standard newsletter price point. The gap between those two things is where the conversion happens.
Basically, the data suggests we are delivering “Alpha” (utility) rather than just “Beta” (entertainment).
And I am okay with that. We are in the building phase. The data is clear: You guys are upgrading because the utility is worth it.
The Catalyst
I wanted to know what specifically triggered this. So I looked at the “hockey stick” moment on my growth chart. It happened on November 22nd.
I looked back at what I published that day. I expected it to be one of my deep-dive philosophical posts where I answer the questions to life, the universe, and everything.
Nope!
It was a post titled: MINING COST PER BTC OF MAJOR MINING COMPANIES.
It was boring. It was dry. It was literally just a list of numbers that I had painstakingly dug out of balance sheets so you didn’t have to (by that, I mean I wrote a Python script to do it for me like a good little worker, and I happily took the credit).
It all clicked then. I’m not funny, or charming… but you don’t hate data either, do you?
The Revelation - Opinions are Beta. Data is Alpha.
The internet is drowning in opinions. You can get “market sentiment” for free on X or Reddit. It is a commodity. Finding the right ones to listen to? Even more difficult.
But curated data? A tool that saves you 3 hours of research? A Python script that automates your morning routine? If I wanted something like that, who do you suppose I turn to (hint: that mirror again).
The Protocol: How I Used AI to Roast My Own Business
I didn’t just find these numbers by staring at a spreadsheet until my eyes bled. I cheated.
I downloaded my Substack export, uploaded it to an AI, and asked 2 simple questions: “Is this good, and how can I improve?”
AI gives the greatest unbiased opinions of all time. I didn’t want a friend who’s afraid to hurt my feelings or cares how hard I worked (I did double check with a human though - to that human, thank you). So, ego armor on, give it to me, Gemini.
1. The “Boring Title” Hypothesis The AI pointed out that my “Nov 22” catalyst post had the driest title in history (Mining Cost Per BTC...). It compared this to what I thought was a hilarious title in Bloodbath & Beyond. Color me disappointed when the hours I worked on that get missed because I wrote that for ME, not YOU.
**The Verdict:**Clever gets likes. Boring gets paid. When people see a “boring” title that promises data, they click because they know exactly what they are getting. Utility beats Wit every time.
2. The “Barbell” Strategy (And Why This Post Is Free) I asked the AI, “Should I paywall this finding?” It told me I should if I include any data. I argued, and it suggested a “Barbell Strategy”:
-
Free: The Narrative. The “What” and “Why.” (This post).
-
Paid: The Utility. The “How.” (The scripts, the data tables). I am giving you this blueprint for free because it builds trust. Normally, if you want the Python script that automates the data collection? That is what the upgrade button is for.
3. The “Cheat Sheet” Mandate The AI told me to stop treating these data posts as “accidents” and start treating them as “products.” Which is good, because not only do I love data, but it turns out I can help out a lot of other writers here too.
The “Alpha Content” Checklist
I don’t always follow the data and have learned to trust my instinct, but it tells a story and it’s here to keep my emotions in check. I want to apply the same to my business here, so I want to know if I am sending you a diary entry or a weapon - but it’s what I’m going to use moving forward.
A blueprint schematic lying on a table, but the lines are drawn in glowing green neon. The blueprint shows the structure of a "Perfect Trade" rather than a building. Next to it is a cup of coffee and a red pen. Cyberpunk aesthetic.
1. The Google Test
-
Fail: Can the reader find this information on the first page of Google? (Example: “What is a moving average?”)
-
Pass: Did you compile data that is scattered, hidden, or requires paid software to see? (Example: “I aggregated the burn rate of 15 miners manually.”)
2. The Bookmark Test
-
Fail: Will the reader read this once, nod their head, and close the tab?
-
Pass: Will the reader need to come back to this post next week to reference a number, a level, or a code snippet? Utility equals retention.
3. The Obsolescence Test
-
Fail: Does this post matter in 48 hours? News commentary expires fast.
-
Pass: Is this a tool they can use for 6 months? Evergreen scripts and specific strategies build a moat around your subscription.
4. The Paywall Partition
-
Free: The “What” and the “Why.” Give them the narrative, the thesis, and the hook.
-
Paid: The “How.” Give them the exact strike prices, the Python code, or the downloadable CSV.
5. The Visual Audit
-
Fail: A wall of text.
-
Pass: Does the post contain a chart, a map, or a list that tells the whole story in 3 seconds?
Moving Forward
This doesn’t mean the “voice” is going away. The rants will stay. The psychological warfare against the market will stay.
But I’m not going to stress myself out at the quantity anymore, and focus on the quality. Expect things you can download, copy, and paste to make money.
Bonus: Build Your Own Dashboard
For the nerds who want to replicate this analysis on their own Substack data, here is the exact prompt you can give to an AI coding tool (Claude Code, Antigravity) to build a dashboard for you.
The Prompt:
“I have a folder containing my Substack export files (specifically posts.csv and subscribers.csv).
Please write a Python script using Streamlit and Pandas to create a ‘Newsletter Growth Dashboard’. I want to visualize my growth and identify what drives conversions.
The app should perform the following functions:
Data Loading: Allow me to upload the CSV files or read them from a local ‘data’ directory.
Key Metrics: Calculate and display my current ‘Free to Paid Conversion Rate’ (Paid / Total) as a large metric at the top.
Growth Chart: Plot a line chart of ‘Total Subscribers’ over time.
Catalyst Overlay: On that same line chart, overlay markers showing when I published posts (using the data from posts.csv). I want to see which posts align with vertical jumps in subscribers.
Spike Analysis: Create a function that identifies the 3-day window with the highest percentage growth in subscribers. Display the post that was published immediately prior to that window.
Aesthetics: Please use a clean, professional layout compatible with dark mode.”
Fellow Substack creators - Enjoy!