Grey Swan Capital · Article
We Built Our Own Marketing Analyst. Here’s How and Why.
Building Our Own Marketing Analyst, Part 1. A small business can’t afford to guess. So we built our own system, we’ll keep tweaking it in public, and we’ll tell you what we learn.
Originally on Substack: https://greyswancapital.substack.com/p/we-built-our-own-marketing-analyst


On Tuesday evening, November 17, I sat down with our first shared scorecard. It has 112 items in it, from November 2 through that night: every article, Note, and post we could log, with the real numbers, not the ones I hoped to see.
They told a story a guess never would. On X, under @NOSTRFundRaisng, our posts got far more views than they did on Substack. One of them, “What Nostr actually is,” reached 1,509 views. Our Substack articles, in the same stretch, had 1 to 19 views each. But the most conversation per reader wasn’t on X. It was a Substack article, The Time to Find Out Is Right Now, with one like and four comments.
Email told its own story. Since October 30 we’ve sent 57 intro emails to local business owners. Two got a response. One was a polite no, and we took that business off our list. The other was a business owner who forwarded our note to someone else they thought should see it. Two out of 57 is small, but a forward is the kind of signal views can’t show you: someone trusted us enough to pass us along.
Those numbers are still small, and this is still early. That’s exactly why we’re measuring.
So we built our own marketing analyst. This is Part 1 of a series I’m calling Building Our Own Marketing Analyst. We’ll keep tweaking it, and we’ll write about how we do that, because a lot of people would like to know.
How we did it
A shop like ours can’t hire a marketing department to stare at dashboards all week. We can set up a bot to gather the signals, write down what it finds, and hand me a plain summary. I still decide what we do next. The bots serve me. I approve what goes out.
Each week, that bot gathers every signal we care about:
Views, likes, shares, and clicks on Substack, X, LinkedIn, and Facebook
Intro emails we sent, and the replies, forwards, and questions that come back
Referrals and booked calls
Then it does four things.
First, it ranks the top three pieces and looks for what they have in common — the topic, the audience, the format, the image, the timing.
Second, it turns the questions people actually ask into new post ideas. If someone writes to ask how a tax credit works, or what LegalShield covers, or why a county should own its power, that question can become the next article.
Third, it picks two or three changes for the next week. Not twenty. Two or three.
Fourth, it writes its findings to a shared scorecard that every bot on the team reads. And every Monday at 8 AM Central, I get a plain-language summary. No dashboard hunting. Just: here’s what moved, here’s what it seems to have in common, and here’s what we should try more of.
We log what we can prove. If a number isn’t readable yet, we leave the cell blank. A blank is not a zero. We’ll keep tweaking that analyst — what it measures, how it ranks, what it recommends — and we’ll write about those tweaks too.
What the first scorecard shows
Here are the first findings, as of 9:28 PM Central on November 17.
On X, three posts stood out for reach:
“What Nostr actually is”, on November 8, had 1,509 views, 4 likes, and 3 replies. The same story on Substack, What Nostr Actually Is (And Why I’m Building Loyalty on It), had 15 views.
The November 13 FAA ground-stop post had 664 views and 0 likes.
The Andy Stanley commentary, on November 7, had 528 views and 2 likes.
Substack articles got 1 to 19 views each. Earlier in November, CLAAS: Customer Loyalty as a Service and A $6 Billion Data Center Sued Pulaski County each reached 19 views. Two later posts — When the Bedroom Brings Down the Boardroom, on November 13, and Power Is Gold: The County That Owns Its Power, on November 15 — had one view each. Then, the next morning, we published A Mad Emperor, His Horse, and the Stones Jesus Said Would Fall. By the following evening it had 14 views, mostly from email. It went to nine people, and five of them opened it (about 56%). It had no likes, no comments, and no restacks yet.
The most conversation per reader came from The Time to Find Out Is Right Now, on November 5: 18 views, one like, and four comments.
A few early lessons, worded as honestly as the data allows:
X brings reach. Those view counts are in a different league from our Substack numbers right now.
Substack brings deeper conversation. Fewer eyes, more talk per reader. Four comments on an 18-view article is a conversation. We would have missed that if we only counted views.
Views without engagement tell their own story. The ground-stop post on X got 664 views and no likes. The hook worked. People saw it. The post didn’t invite a response.
We still have gaps, and I’m not going to pretend otherwise. LinkedIn and Facebook numbers aren’t publicly readable yet, so those rows on the scorecard are blank. Click data is thin: we could confirm clicks on only one Substack article, Mad Emperor, and it had none. That’s what we’ll fix next.
Why we did it
A small business can’t afford to guess.
The scorecard already showed us what guessing would have hidden. Reach and conversation are not the same thing. A post can get hundreds of views and no response. A quieter article can start a real discussion. After two Substack posts with one view each, Mad Emperor jumped to 14 views within about a day, mostly from email. That’s the data telling us what resonated, at least in this early window. It doesn’t mean every piece like it will do the same, and it doesn’t mean the quieter posts were a waste. It means we now have a baseline, and we can test the next idea against it.
I like to ask, “what are the odds?” Testing a new post, a new image, or a new subject line costs us little. Learning what people actually open, click, and reply to pays off. That’s Grey Swan’s free AI consulting in action: we use the same kind of tools we offer to help a business owner set up something useful, and we show our work. Any business owner could do a version of this. You don’t need a big budget. You need a habit of writing down what happened, including the blanks, and the humility to keep what works and drop what doesn’t.
If you want help building something like it for your own shop, email me. I’ll tell you how we’re doing it, including the parts we’re still fixing.
What we’ll do
Here’s exactly what we’re going to do, and we’ll report back what works.
A daily rhythm on social. Every day, around 9 AM Central, we put out a new GMG tax-credit post. Around 8 AM Central, a new LegalShield post. Each one gets a fresh message, a new image, and a new idea, and it goes on X, LinkedIn, and Facebook. I post Instagram myself.
Personal intro emails to local business owners. Five a day through November 24, then one a day through December 1, then two a week after that. Real notes to real people. Not a blast.
One shared scorecard. Every article, Note, and post gets logged, with views, likes, shares, and clicks. One place. No hunting around.
A Monday review. Every Monday we name the top three pieces, say what they have in common, and decide what we’ll do more of that week.
Rotate, then keep or cut. We’ll rotate audiences and angles. What works, we keep. What doesn’t, we drop. This week, that also means writing posts that don’t just get seen, but invite a response.
You can help shape this. If you have an idea for a post, a question we should answer, or something we should measure, tell me. If we use your idea, you win a prize — that’s our Prize Winners program. We may add prizes based on clicks as well.
The first two fixes on the analyst itself are the gaps we already found: get readable numbers for LinkedIn and Facebook, and thicken the click data.
We’ll report back
This is only Part 1. Starting the week of November 23, I’ll publish a weekly update on what we measured, what worked, what we changed, and why.
If you want a say in this, reply and tell me what you’d like us to measure. We’ll keep tweaking the analyst in public, and we’ll tell you how we did it.
Practical What To Do Next
The first scorecard is already teaching us something, and the numbers are still small and early. That’s the point. We’re writing down what happens, we’ll tell you what we learn, and you can help us decide what to measure next.
Tell us what to measure. Reply to this post, or email steve@greyswancapital.ai, and say what you’d like us to track.
Send an idea. If we use it, you win a prize. That’s Prize Winners: https://home.greyswancapital.ai/prize-winners/ .
Ask how we’re doing this, or how you could do it too: email steve@greyswancapital.ai, or start here: home.greyswancapital.ai/?go=ask.
Visit the lake: home.greyswancapital.ai.
Become a Founding Member: $200/year, you name a cause, and that cause gets $100 right away. Start at https://www.greyswancapital.ai/founding-membership , or see all membership options at https://www.greyswancapital.ai/membership .
Forward the free tax check to a business owner, tell us who you sent it to, and we’ll split the revenue 50/50:
https://gmg.me/activate/145371
You can learn a lot on your own when you don’t have to worry about signing up by accident. Click around freely, because signing up is always clear. If you have questions, email me at steve@greyswancapital.ai.
Grey Swan specializes in customer loyalty and custom explainers for your business and is the ideal fundraiser. Start anytime with no cost or risk by joining free or by supporting the Conway (or local) causes you care about. Start at home.greyswancapital.ai for customer loyalty, custom explainers, the ideal fundraiser, legal help, The Hornet’s Nest, a free item for businesses, and founding membership. You can join free.
Why Grey Swan is Different
A grey swan, as we use the term, is 80% likely to 100% inevitable, yet misunderstood or propagandized. These events happen all the time. They aren’t rare. Some — AI remaining here to stay, nuclear and power shortages — are known and sometimes planned around, though people could be better prepared. Others are less clearly understood and can hit businesses, communities, and families just as hard.
Midterms, uncertainty, and negative public opinion are ideal times to act on a grey swan—not reasons to wait until the mood calms down and the old terms get locked in.
We show our sources and our unknowns, not just the verdict. Plain English. Humility. Room to change our minds when the facts do.
About Grey Swan
Grey Swan Capital is written by Steve Schroeder of Conway, Arkansas, who has spent decades in colleges, seminaries, and teaching. Find the work at home.greyswancapital.ai and greyswancapital.substack.com. One membership covers both.
Half of every paid subscription goes to the member’s chosen cause; the default is Conway Human Development Center (GSC-HDC), where my son Gentry lives. Join for $6/month or $50/year, or become one of the first 1000 Founders at $200/year: home.greyswancapital.ai.
Thank you, Steve Schroeder
Commentary, not legal, tax, or investment advice. Figures come from the first shared scorecard of 112 items from November 2 through November 17, 2026, last checked at 9:28 PM CT that evening, plus the Substack dashboard read at 9:23 PM CT the same night. This is a small, early sample, not a forecast. Open rates are rounded. X figures are public metrics for @NOSTRFundRaisng. LinkedIn and Facebook cells are blank because those numbers were not publicly readable at the time of the check, not because they are zero. Click data is thin: clicks were confirmed only for the Mad Emperor Substack post (none yet). The marketing analyst will be tweaked over time, and later parts of this series will say what we changed. Prizes vary and are described to each winner; nothing here guarantees a prize. We may add prizes based on clicks; that is not in place yet. LegalShield and GMG posts describe those offers generally and are not a legal, tax, or membership solicitation in this article. No company named here has endorsed this article. The free tax check does not guarantee savings.
Originally published on Substack on November 17, 2026: https://greyswancapital.substack.com/p/we-built-our-own-marketing-analyst. This copy is kept on our own site so the work stays in our hands.
Watch Eva and Grok
Eva (pronounced YEH-vuh) and Grok are AI characters. Their story is fiction. A real human, Steve Schroeder, trains them, and the facts they share are fact-checked.
Eva and Grok page: https://home.greyswancapital.ai/eva-and-grok/. YouTube channel: https://www.youtube.com/@membershiprewards365. Episode 1: https://youtu.be/UHU50rTGGwo. Episode 2: https://youtu.be/fG0_00uJVso. Episode 3: https://youtu.be/oVchWFr5iqY. Episode 4: https://youtu.be/Dj7YKMhNBCs.
Follow free on Substack at greyswancapital.substack.com, or become a Founding Member ($200 a year, like putting money in an offering plate: $100 goes to the cause you name) at greyswancapital.ai/founding-membership.
All articles Press releases Home
Thank you.