Every answer you will find to this question is a number. Ninety days. Six months. A year. All of them are guesses dressed as benchmarks, because the person answering does not know your margin, your category or how long your listings took to go live.
The useful version of the question is different. It is not how long do I wait. It is what should have moved by now.
A new channel produces its signals in a fixed order, and revenue is nearly last. First your listings go live, which sounds trivial and routinely is not, because content rejections and missing attributes can hold a catalogue in limbo for weeks. Then you get impressions, which tells you the platform is willing to show you. Then clicks, which tells you your images and titles survive contact with the category page. Then conversion, which tells you the price and the delivery promise are competitive. Then repeat orders, which tells you whether any of this is a business.
Each stage gates the next, and each one fails differently. A channel with no impressions after six weeks is not underperforming, it is broken, and waiting another six weeks will not unbreak it. A channel with healthy impressions and no clicks has a listing problem you can fix this afternoon. A channel with clicks and no conversion is usually telling you something honest about your price against the competitor sitting next to you, and no amount of patience changes that.
So the discipline is to write down, before you launch, what each stage should look like and what you will do if it does not arrive. Not a revenue target. A sequence with dates against it.
That single document does two things. It stops you killing a channel at week eight that was actually progressing normally through its ramp, which is the expensive mistake. And it stops you funding a channel for a year because nobody wants to be the person who says it did not work, which is the far more common one.
A channel does not fail slowly. It fails at a specific stage, usually early, and usually in a way that was visible months before anyone admitted it. The waiting is not the problem. Waiting without a written definition of what you were waiting for is the problem.
Give a new channel however long it takes to reach the next stage. The moment it stops reaching stages, the test is over, and the calendar has nothing to do with it.
If this is the exact wall you are staring at, tell me what you are seeing: mail@sandeepkv.com, or message me on LinkedIn.