Sort of. It's more because of how ML models work. If you have an audience of 100mn then it's much easier to get enough conversions to optimize the models. It's much harder to do it with an audience of 3mn.
By that logic Indonesia and Bangladesh would have higher ad-spend per head than France, because they are larger markets and it doesn't matter how much money people have to spend?
It's clearly both. But if I have a 1/100000 outcome, I can get maybe 500 conversions in France or the UK, but 3000 in the US. That makes a really large difference in terms of how many times your ad is shown to likely users.
But yeah, ML models do in fact work better in Indonesia and Bangladesh, but as you noted they have less money to spend.
