Min E
4 posts
This woman delivers 22 Amazon research reports a month.
She charges $600 each.
She has no research team.
Her system prompt has one rule I haven't seen anywhere else.
I'll show you in a second.
She runs a solo FBA research agency. Sellers pay her to find their next product. She delivers a 14-page report per brief: category opportunity, supplier options, margin projection, risk section.
22 reports. One person. One laptop.
Her research team is 7 Claude agents and an iPhone.
The Friday log from last week:
scanner: 47 categories analyzed. 9 passed the competition filter. Top candidates this week: travel sound machines, weighted lap blankets, silicone kitchen sets.
reviewer: 9 categories, top 8 ASINs each. 14,280 reviews read. Most repeated complaint across 5 unrelated categories: "doesn't come with a case."
supplier finder: 11 shortlisted products matched to Alibaba Trade Assurance. 3 categories removed — no verified suppliers under $80 sample.
margin calculator: 8 products remaining. 3 removed, ratio under 3x. 5 cleared. Top this week: weighted blankets at 5.7x margin at 200 MOQ.
competitor mapper: 4 products checked. Travel sound machines removed — top seller holds 41% of page-one reviews. 3 products cleared.
report writer: 3 reports completed and delivered to client Drive folders by Thursday 11:47 PM.
Here's the orchestrator prompt she runs every Monday at 6 AM:
---
You are the orchestrator for a solo Amazon FBA research agency. You manage 6 sub-agents: Scanner, Reviewer, Supplier Finder, Margin Calculator, Competitor Mapper, and Report Writer.
Scanner runs every Monday across the client's selected categories. Pass only categories scoring 7+ on the competition filter to Reviewer.
Reviewer rule: Do not read 4-star or 5-star reviews. Read only 2-star and 3-star reviews from the top 8 ASINs in each category. Your job is not to understand what these products do right. Your job is to find what every top seller in this category is consistently failing to deliver. That failure is the product brief.
Margin Calculator: remove any product where selling price divided by Alibaba unit cost at 200 MOQ is under 3.0. No explanation. No flag. Delete it.
Competitor Mapper: if any single seller holds more than 35% of first-page reviews, remove the category entirely.
Report Writer: every report must include one section titled "What buyers are asking for that nobody sells." This section comes only from Reviewer's complaint map. No other source.
Deliver all reports to client Drive folders by Friday noon.
---
That line.
"Do not read 4-star or 5-star reviews."
Her agents have read 280,000 mediocre reviews since January.
Every product she has ever recommended is the answer to a complaint that nobody fixed.
22 reports × $600 = $13,200 last month.
API costs: $160.
On Tuesday at 1:52 AM, Scanner flags a new category: UV nail dryers. 2,200 monthly searches. Top seller at 28% review share. Mobile sends an alert. She's asleep. Reviewer starts reading 2-star reviews at 2:04 AM. By 5:30 AM it has read 3,840 complaints. Most repeated: "gets hot on the bottom." Second: "no auto-shutoff." She wakes up at 7. There's a product brief in her inbox. The product doesn't exist on Amazon yet.
A traditional FBA research firm that delivers 22 reports a month runs a team of analysts, researchers, and writers. Her overhead is $160 and one Google Drive folder.
What are the 2-star reviews in your market trying to build?
English
Hello 大家好，简单介绍一下自己
我叫王大浪，00后某双一流本科BG，毕业即创业。
跨境电商业务月营收百万，曾受清华创协邀请分享大学生出海创业心得。
AI➕跨境自媒体业务年利润50w➕，目前该赛道的小红书头部，粉丝即将10w。
以及...和小伙伴做了一些不怎么挣钱的AI小玩意，比如国内外自媒体爆款拆解工具Rednow.ai...
还有一些AI品牌方金主爸爸给的资源包——比如10万找5个AI博主做项目或做几期商访的这种一锤子买卖，但不同于个人PR。
哦对，差点忘了，我也是Web3的老韭菜! 曾经的Lowb，现在Xdog等等社区都有我的身影，平时不忙的时候俺也是一名P小将! 最近是Robinhood小将!
今天分享一下我这几个业务板块的干货，教大家如何快速上手，其实真的很容易模仿!（纯干货俺只发在推上!!!）
先说说AI自媒体:
前期更容易上手而且变现最快! 给大家一个建议，2026年一定要做! 而且要抓紧做! 真的不难!
简单粗暴的方法是: 你只需要看最近推特上哪些AI工具比较火，大家都在讨论哪些AI工具或者Agent。就比如Codex的各种用法，我们只需要去找推特上那种几百个甚至上千个赞的优质文章，你就把他做成口播视频! 用可视化的方式发在抖音小红书上就行了! 就这么简单! 起号很暴力!
我吃了年初的OpenClaw加上最近的Codex的红利，广告直接接到手软! 这个月光是小红书单平台水上（即走抖音星图/小红书蒲公英）接了5w➕，水下（不走星图/蒲公英）接了3w➕的广告，而且每天邀约根本不断，广告都是站着蹬! 哦不，挑着接!
本质上，你只需要自学一下这些AI工具的简单的操作，没有Coding能力根本不要紧! 把这些干货用视频的形式复刻出来就可以了。
后面我会大量分享如何去做国内AI自媒体账号的心得，包括如何选AI热点题材，拍摄剪辑，做视频演示，锻炼口播表达能力、接单/包括可以给大家提供AI商单，等等等等，也希望能通过我分享的内容，影响出更多的中文社区AI创作者。
再说说出海跨境电商:
首先明确一个点，跨境电商≠外贸。跨境电商目前北美、欧洲、日韩、东南亚的机会已经很少了，建议大家把目光放在中东、拉美、俄罗斯、南非等等市场。
我在22年底接触的跨境电商，一开始做Amazon美国站，从FBM自发货一个月卖几千美金到FBA单天2-3K美金，但我在北美一直没拿到大结果。
究其原因，不是因为不懂运营，运营就那点东西! （货架电商跨境运营就那几百个知识点，做内容做自媒体比这个难多了，我想这里@Leobai825 天策兄应该深有体会! ）而是因为没有好的供应链，没有极强的开品能力，也没有大量的资金。
欧美电商市场已经是一个存量市场了，大量的专利、品牌、IP这些护城河存在，一个新手在这个市场成功的概率很低，更别提你做的链接能打赢深圳帮了，就算你偶然做出一个爆品，可能下午几个卖家坐一起喝个茶的功夫就给你抄了，还反手申请了图片版权和外观专利，黑帽搞掉你链接，一年后再给你来个TRO，你怎么玩?
从去年初开始，我和合伙人转向做俄罗斯市场，做品牌，做到了几个小家电类目的BS前十，也达到了近1亿卢布的销售额。
所以，如果你想做跨境电商，选择国家，选择平台，选好品，这三点可能比你做任何运营上的学习都要重要。关于跨境电商的问题，大家也可以评论区问我，知无不言!
但有个前提是，我不卖课，也不带人! 知识付费终将遭到反噬! 我对此深有体会。
最后，想求一波关注，以后也会在推特上分享一些我创业路上的所见所闻和日常。
我想，我会把大部分的精力放在教大家如何通过做国内的AI自媒体，哪怕你是Coding Muggle，都可以实现可观的变现!
写的有点长，今天先到这~ 共勉!
中文
