This plan is written for one person, working with tooling, based in Vietnam, with no existing audience. It does not promise USD 1,000 a month in 90 days — per the economics in the earlier post, USD 1,000 a month in a good category needs roughly 51,000–106,000 views, which is month six territory. Ninety days buys you the answer to whether to continue.
Day 0: three blockers, before any content
- Settle the payout route. Test it with a small amount. With no payout route everything below is meaningless.
- Join Amazon Associates US, get a real tag. Complete the tax interview declaring services performed outside the US. Details in the first post of this series.
- Register as a household business. The VND 1 billion exemption sounds distant, but at USD 3,224/month you reach it, and registering after crossing the threshold costs more than registering before.
Days 1–30: the first three orders
Amazon closes the account with fewer than three sales in 180 days. That number is smaller than it looks: at a 1% conversion rate, three sales is 300 clicks, which is 15,000 views, or about 19 videos. At 3% it is only 5,000 views and about 6 videos. Month one's goal is not revenue, it is clearing probation — so pick a category that converts easily, not one with the highest rate.
- Pick exactly one category. Kitchen, tools, or small home electronics — things with existing demand priced USD 40–120.
- 20 videos, one product each, a distinct sub-tag per video so you learn which one pays.
- Checkpoint: 3 sales in the first 30 days. If you miss it, the problem is product selection, not video count — change category, do not add videos.
Days 31–60: find what converts
This phase has one question: of last month's 20 videos, which produced clicks and which produced nothing. The sub-tag in the link exists to answer it. Without it, every later decision is a guess.
- Compute revenue per 1,000 views for each video. It is the only comparable metric — views do not pay.
- Cut the 80% of videos with zero revenue per 1,000 views. Make 40 more using the template of the surviving 20%.
- Checkpoint: revenue per 1,000 views in the good group must be at least 3x the bad group. If it is not, the content is not differentiated — change the hook, not the product.
Days 61–90: scale volume, not quality
By now you know which template pays. What remains is volume, because the gap in the economics table is 2–20x in views. At 800 views per video, reaching 106,000 views takes about 132 videos — roughly 4.5 a day. That is why the work has to be machine-assisted: each video renders in 13–23 seconds, but product selection and hook writing remain human work.
- Target 120–150 videos in month three, one template.
- Checkpoint: if after 90 days revenue per 1,000 views is still under USD 1, this model does not work in the chosen category. Change category, or change offer.
- Above USD 1 per 1,000 views: that is the rate where 106,000 views yields USD 1,000. From there it is only a volume problem.
Three things not to do in 90 days
- No dashboard, no multi-channel management system, no analytics. There is no data to display yet.
- Do not open a second channel before the first has sales. Two half-channels are worse than one that sells.
- Do not buy views. Policy verbatim: "will not artificially generate clicks or impressions ... whether by way of a robot or software program". It is an account-closing offence, and it also corrupts the very data you use to decide.
This plan deliberately has no "USD 1,000 in 90 days" target. What 90 days buys is certainty about whether the model works — and that is the only thing worth trading 90 days for. Source of the numbers: the economics script at /mnt/blockstore/ebay-aff/amazon_blog_numbers.py.