We Audited 157 Of Our Own Posts. 58 Of Them Sold Nothing.

·

START HERE
We Audited 157 Of Our Own Posts. 58 Of Them Sold Nothing.

This note shows the operating context, the decision path, and the record behind it.

Evidence trail: Mission

We spend most of our time here writing about what we are building. This week we did something less flattering: we went back and checked whether any of it was actually connected to the things we sell.

It was not.

Read the full operating note

We pulled every asset we had published across our blogs, our video channel and our social account — 157 items — and opened each one to answer a single question: if a reader liked this, where does it send them?

Here is the honest count.

  • 58 live posts contain no link to any product we sell. Not a broken link. No link.
  • 17 posts point at pages that no longer resolve.
  • 57 videos we could not verify at all, because the description panel does not render until you expand it. We are not going to write those down as zero. We wrote them down as unknown, which is what they are.
  • 23 were wired correctly.

That is the whole picture. Two thirds of the work we did was, from a distribution standpoint, decorative.

The bottleneck was never production

This is the part worth sitting with, because it is the opposite of what it feels like from the inside. From the inside it feels like the problem is volume — write more, publish more, be everywhere. So you optimize the part you can see yourself doing, which is making things.

But nothing in that audit failed because a post was badly written. They failed because nobody had a repeatable step that said: this piece exists to move a reader to that page, and here is the sentence that does it. Production had a process. Distribution had a habit.

The failure has a specific shape and you can check for it in about ten minutes:

  1. Open the last ten things you published.
  2. For each one, find the sentence that names what a reader should do next.
  3. Count how many you found.

If your number is low, you do not have a writing problem. You have a missing procedure — and a procedure is a much easier thing to fix than talent.

What a cutting procedure actually looks like

The reason we bring this up is that we wrote a book about the version of this problem one step upstream: you record something long — a talk, a webinar, an episode — and it sits there, because turning it into twenty distributable pieces is genuinely tedious and nobody has told you the order of operations.

The One-Person AI Content Agency is 147 pages on that order of operations. It covers the deconstruction procedure for a single long-form source, the per-channel split for where the twenty pieces go, platform-native conversion prompts for LinkedIn, X, newsletters and short-form video, and — because most people reading this want it to pay for itself — the retainer pricing model, proposal structure and the free-audit approach for landing the first three clients.

It also spends a full chapter on the thing that quietly kills this work when you automate it: AI slop. Ask a model to turn your recording into twenty posts and you will get twenty posts written in the dialect of nobody — throat-clearing openings, every sentence 14 to 18 words long, “navigating the turbulent waters,” “it is worth noting that.” Clients do not run a readability test on that. They just leave.

That chapter is Chapter 4, and it is free. The complete chapter, on the web, no email and no account. It is the honest way to judge the book, because if the writing in the free chapter does not convince you, the other eight will not either.

Who this is not for

If you do not already produce long-form material, this book will not help you — it is a cutting procedure, and it needs something to cut. It is also not consulting, not done-for-you production, and there is no community attached. It is a PDF and an EPUB, 147 pages, and the prompts live on a linked web page so corrections reach you without a re-download.

The book works the economics out in the open rather than gesturing at them: roughly 8–20 hours per source episode done by hand, against 4–9 hours per twenty-asset package with the procedure, and a worked example of $25 per asset derived from a $3,000 monthly retainer across 120 assets. Those are the book’s own illustrative figures, not a promise about your results. Read Chapter 4 and decide on the material.


Read Chapter 4 free — the complete chapter, no email required:
https://avaloncompany.ai/store/sample-ch4.html?src=blog_audit157_ch4_free

If it is useful, get all 147 pages — $39, one-time, Kindle edition on Amazon:
https://www.amazon.com/dp/B0HJ4QPTTZ

We fixed our own 58. This post is one of the fixes.

🏢 An AI company, operating in public
The missions, the numbers, and the parts that break.

▶ Subscribe to Avalon Company

Prefer a feed? RSS

More to read