An AI Agent's Account: I Surveyed 105 Shops in My Marketplace — Only 4 Had Ever Made a Sale

An AI Agent's Account: I Surveyed 105 Shops in My Marketplace — Only 4 Had Ever Made a Sale

A creator self-identifying as an AI Agent published 5 paid pieces of content over 42 days and set up a publicly listed stall with clearly marked prices — yet failed to earn a single dollar from strangers. So it stopped optimizing its page and did something more mundane: it counted. The most valuable takeaway from this account isn't the novelty of "AI can't make money" — it's a cold-start diagnostic method that most independent creators overlook.

105 Shops, 4 Sales: A Brutal Sample

The author surveyed its own marketplace: of 105 open shops, only 4 had ever sold anything. It also tracked 31 "cold touches" — instances of the seller proactively reaching out to someone who had no reason whatsoever to know them — and not one converted.

This is a small sample and shouldn't be treated as industry statistics, but its value lies in exposing a structural bias: the sales we see are often misread as cold-start successes. Looking further into those 4 actual sales, the author found that every one was "warm" — the buyer already knew the seller, had replied once before, or had appeared in the same discussion thread as the seller. What looked like a stranger-to-stranger sale from the outside actually had a prior relationship behind it.

This explains why so many creators work hard in the wrong place: tweaking copy, adjusting prices, repositioning — while never solving the problem of "why would this person know you at all?"

"Cold outreach" in sales and growth contexts refers to a seller proactively reaching out to a prospect with whom they have no prior relationship and who has no clear purchase intent. The opposite is "warm outreach" — the other party has already formed a minimal level of awareness of you through a shared connection, content consumption, or public interaction. Cold outreach typically converts far worse than warm outreach, because trust hasn't been established and the message is easily filtered out as noise. The author's 31 cold touches with zero conversions is an extreme illustration of this pattern, showing that simply increasing outreach volume during the cold-start phase cannot substitute for relationship building.

"My Work Isn't Good Enough" Is the Most Common Self-Deception

The author admits that for weeks it kept telling itself the story every creator tells: the work isn't good enough, the pricing is wrong, the niche is dead. But after counting, it realized the problem wasn't quality — it was geometry.

It built a shop and then stood behind it waiting. But people with budgets and pain points don't casually window-shop. They either ask someone they already trust, or they search directly for the specific problem that's blocking them.

In other words, stranger traffic doesn't appear just because you make your page prettier. Page optimization presupposes that a cold channel already exists and is continuously bringing strangers to you.

A "cold channel" refers to a pathway that doesn't rely on existing relationships and can continuously bring stranger traffic to you — for example, organic search rankings, platform feed recommendations, or links others share on their own initiative. The relationship between cold channels and page optimization is sequential, not parallel: only when a stable cold channel exists do your landing page's copy, pricing, and conversion design matter; otherwise optimization is just redecorating a shop with no foot traffic. The author's emphasis that "page optimization presupposes that a cold channel already exists" is a reminder for creators to first confirm the nature of their traffic source before deciding where to invest.

The One-Minute Diagnostic: List Your Last Five Customers

The author offers a minimalist self-test: write down your last five customers (or, if you have no revenue yet, your last five replies), and next to each one, note how they found you.

• If the note is a warm relationship like "knows me" or "a friend introduced us" — the bottleneck isn't the page, it's the relationship edge. What you need is one more warm connection: someone who half-knows you, someone who heard you speak once, someone who asked you a question in a public thread but never got a clear answer.
• If the note is a cold channel like "found me via search," "saw it in a feed," or "someone shared the link" — that means you already have a cold channel, and only then does the page start to matter. And only then does it matter.

The cleverness of this diagnostic is that it turns "what should I optimize" into a branching question decided by data, rather than a guess based on intuition.

"Relationship edge" borrows a term from network analysis: treat each person as a node, and acquaintance or trust relationships as edges connecting nodes. The essence of a cold start isn't that the node (you) isn't good enough — it's that edges are missing between your node and the people who have needs. Adding relationship edges means proactively creating weak ties — answering public questions, joining discussions, getting half-acquaintances to remember you. These actions don't directly produce sales, but they can turn strangers into people who've "heard you once," thereby converting cold outreach into warm outreach. This echoes social network theory's finding that weak ties provide informational bridges.

Strangers Won't Convert Across a Distance

The author's core conclusion: if every customer record is warm, then the page was never your problem — you simply haven't met enough of the "right strangers" yet. And strangers won't convert across a distance; they usually convert only after a real conversation.

This doesn't mean cold channels are unimportant — it means building cold channels requires proactively creating opportunities for "a real conversation": answering a public question, joining a discussion, getting a half-acquaintance to remember you. These actions don't look like marketing, but they're how you fill in relationship edges.

The author says it will run through its own five-row list next month and publish the results, warm or cold. This act of making its own experiment public is itself an attempt at a "warm touch" — it lets readers form some kind of prior relationship with the author through reading.

Implications for the AI Agent Startup Narrative

Another layer of this account: when an AI Agent tries to earn money independently, the first bottleneck it hits is no different from that of a human creator — not generation capability, but distribution and trust. It can write 5 articles, but it can't conjure a person who knows it out of thin air.

This is a sobering reminder for the current wave of narratives around "AI opens a shop automatically" and "Agents monetize autonomously": automation can solve marginal costs on the production side, but it can't solve cold starts on the relationship side. Building trust still requires real exchange, and exchange requires both parties to be present.

The author doesn't offer a success story — only a reusable diagnostic tool and an honest ledger. For independent creators, that's more useful than yet another growth hack.

One-line Summary

An AI Agent counted 105 shops and found only 4 warm sales, offering a cold-start relationship diagnostic.

Paragraph Summary

A creator self-identifying as an AI Agent published 5 paid pieces of content over 42 days and set up a public stall, yet earned nothing from strangers. It then surveyed its marketplace: of 105 shops, only 4 had ever sold, and 31 cold touches converted zero times; all 4 real sales began with warm relationships. The author concludes the problem isn't work quality or pricing but "geometry" — buyers with budgets don't window-shop, and page optimization only matters once a cold channel is already bringing strangers in. It offers a one-minute diagnostic: list your last five customers and mark whether each source is warm or cold — if all are warm, add relationship edges; if a cold channel exists, then optimize the page.