levelsio says hotel bookings made $3,240 in two weeks
His first affiliate revenue that actually worked, and he thinks he knows why.
levelsio posted his first hotel bookings revenue, $3,240 in two weeks, which he annualizes to roughly $6,480 a month. His own words on it are "Much more than I expected!"
$3,240 in first 2 weeks So about $6,480/mo Much more than I expected!
The part worth sitting with is not the number, it is why he likes it. He says he is happy to make money with a project that is not reselling AI tokens again, and that it seems more moat-ey to him. That is a founder who has shipped a lot of AI products saying the token resale game feels thin.
The affiliate caveat, from him
He tempered it himself in the replies. Asked about affiliate revenue, he said it almost never made him money before, and that hotels might work because the value of a sale is much higher. Same operator, same tactic, different order value, different outcome. That is the whole mechanism here, and it is old. Commission businesses only pay when the transaction is big enough for a percentage of it to matter.
So if you are looking at this as a template, the transferable bit is not "build a hotel site". It is that a small slice of a large purchase beats a large slice of a small one, and that levelsio's own track record with affiliates on his other projects was bad until the ticket size changed.
The tooling detour
He also tried to squeeze the commissions higher. He says he signed up to Stay22, installed the script, and then it started rewriting his affiliate tags to the complete wrong hotels and places. He deleted it the same day. No word from Stay22 in the sources, and no explanation of what went wrong, so treat it as one founder's experience rather than a verdict on the product.
Traffic, he says, is mostly from X and Google. Which is to say a large existing audience is doing some of the work here, and that is not a variable most people can copy.
The business read. Two weeks of revenue annualized to a monthly figure is a projection, not a run rate, and levelsio is the one reporting all of it. But the underlying point holds regardless of whether the number sticks. If you are tired of thin margins on inference, the alternative he is pointing at is not a cleverer AI product, it is finding a transaction large enough that a commission on it is real money.
Very very happy to make money with a project that's not reselling AI tokens again!
It almost never made me money, with hotels it might work because the value of sale is much higher