Why most reviews are only ever checked by software

That single fact, most reviews pass without a human ever seeing them, is why fakes slip through and genuine ones occasionally get caught. It is not a moral failing of any one platform; it is the predictable shape of policing speech at scale.

The bottleneck, and what sets the floor

Every platform sits on the same knife-edge: block too aggressively and you delete legitimate reviews and silence real customers; block too little and you leave the fakes in place. So the default is that most reviews go up on automated checks, with post-publication systems sweeping for patterns afterwards. What stops platforms from simply doing the cheapest possible job is enforcement. Regulators now set a floor for what those automated and after-the-fact processes have to achieve over time, backed by real teeth, fines reaching ten per cent of global turnover for fake-review failings in one regime, and sweeps that have found a significant share of businesses engaged in misleading review practices in another.

What it means for you, and the limit

For a small business, the practical upshot is to set your expectations honestly. A fake aimed at you may not be caught quickly, and a real review of yours may occasionally be filtered, and neither is a sign that the game is rigged against you personally. The limit is that scale also means your individual case rarely gets bespoke attention, which is exactly why documentation matters when you flag something, and why the durable answer is volume of genuine reviews, not a fight with the filter over any single one.