The Competition and Markets Authority (CMA) is chasing ghosts. By launching investigations into Just Eat, Auto Trader, and others over "fake and misleading reviews," regulators are performing a theater of protection that misses the fundamental shift in how humans actually buy things in 2026. They are fighting a 2010 battle with a 2026 map.
Regulators want you to believe that if we just scrub the bots and the "incentivized" five-star gloss, the internet will return to a pristine state of objective truth. This is a delusion. The "misleading review" isn't a glitch in the system; it is the inevitable byproduct of a system that tried to turn subjective human experience into a standardized metric.
We have reached Peak Review. And the crackdown won't save us.
The Algorithmic Incentive to Lie
Why do companies like Just Eat or Auto Trader find themselves in the crosshairs? Because we built a digital economy where a 4.2-star rating is a death sentence. When the difference between the first page of search results and total obscurity is a fraction of a decimal point, fraud becomes a rational business expense.
The CMA’s "crackdown" treats fake reviews like a criminal outlier. In reality, they are a structural necessity for survival in a platform-dominated world. If you run a local kebab shop and the platform’s algorithm buries you because a rival’s cousin left three one-star rants, you don’t wait for "regulatory intervention." You buy a pack of reviews to keep your lights on.
The regulators are punishing the symptoms while ignoring the disease: the tyranny of the aggregate score. By forcing platforms to police these reviews more strictly, they are simply driving the deception deeper into the shadows, making it harder to spot and more expensive to execute.
The Myth of the Objective Five-Star Rating
The premise of the investigation is that there is a "true" review out there being obscured by "fake" ones. This is fundamentally flawed.
I have spent a decade watching brands manipulate sentiment. Here is the reality: even a "genuine" review is often useless. Consumer psychology is a chaotic mess of mood, bias, and context. A customer might give a restaurant one star because it rained on their way there, or five stars because the waiter looked like their ex.
When you aggregate this noise into a single score, you aren't getting truth. You're getting a smoothed-out, mediocre average that tells you nothing about the actual quality of the service.
The Hidden Value of the "Fake" Review
Counter-intuitively, the presence of obvious fluff or bot-generated praise often acts as a signal to savvy consumers. We have developed a "cynical literacy." We look for the 3-star reviews because they are the only ones with enough nuance to be credible.
By attempting to "clean" the platforms, the CMA is actually making it harder for consumers to use their own bullshit detectors. A perfectly manicured feed of verified, positive reviews is far more dangerous than a messy feed where the fakes are easy to spot.
The Platform Paradox
Auto Trader and Just Eat are being grilled for not doing enough to stop the rot. But platforms are in a double bind.
- Friction is the Enemy: If a platform makes it too hard to leave a review (identity verification, proof of purchase, video evidence), the volume of reviews drops.
- Volume is the Product: A car listing with 500 reviews looks more authoritative than one with 5 "verified" ones, even if half of the 500 are garbage.
The platforms have zero incentive to be perfect. They only have an incentive to look "clean enough" to avoid a fine. This leads to a "security theater" approach to moderation—deleting the obvious bots while ignoring the sophisticated, high-level manipulation that actually sways markets.
Why "Verified Purchases" Won't Save You
The gold standard of the regulatory dream is the "Verified Purchase" tag. It sounds foolproof. It isn't.
In the world of high-stakes e-commerce, "brushing" schemes involve actual products being shipped to actual addresses to create a paper trail for a fake review. In the service industry, it’s even easier. You "buy" the meal, the platform takes its cut, and the five-star review is logged as legitimate.
The CMA is trying to regulate a process that has already evolved past their ability to track it. We are moving toward a post-review world, and the regulators are still checking the receipts.
The Rise of the Dark Social Reference
If you want to know what’s actually happening in the market, look at where people are fleeing to. They aren't looking at Auto Trader’s aggregate star rating. They are going to niche Discord servers, private WhatsApp groups, and Reddit threads where the "review" isn't a metric, but a conversation.
Trust is migrating from platforms to people. This is "Dark Social." It’s untrackable, unrankable, and unmaskable. While the CMA wastes millions of pounds and thousands of man-hours trying to fix the star-rating system, the market is simply abandoning it.
Imagine a scenario where a car dealership has a 4.9-star rating on a major platform, but a 10-minute search on a specialized forum reveals a pattern of transmission failures they’ve been burying with "incentivized" fluff. The regulator sees a compliance success; the consumer sees a trap.
Stop Trying to Fix the Feedback Loop
The obsession with "authenticity" in reviews is a dead end. We should be asking why we rely on these centralized feedback loops at all.
Instead of demanding that Just Eat "fix" their reviews, we should be demanding interoperability of reputation. Why is my "trust score" locked inside a single ecosystem? If I am a reliable reviewer, that reputation should follow me across the web, independent of the platform's desire to sell more burgers or cars.
But the platforms don't want that. And the regulators don't understand it. They would rather play a game of Whac-A-Mole with bot farms than address the centralized power that makes review manipulation so profitable in the first place.
The Real Cost of Regulation
Every time the CMA or a similar body "cracks down," the cost of compliance for small businesses goes up. A massive corporation like Auto Trader can afford an army of moderators and AI filters. The independent dealer or the small family restaurant cannot.
By raising the bar for what constitutes a "compliant" review system, we are inadvertently entrenching the biggest players. We are making it harder for new, disruptive platforms to emerge because they can't afford the legal overhead of policing every single comment to the government's satisfaction.
We are sacrificing competition at the altar of a "safety" that doesn't actually exist.
The Actionable Pivot for the Modern Consumer
If you are still looking at the aggregate score to make a decision, you are the problem. You are the one being manipulated.
- Ignore the Extremes: The 1s and the 5s are almost always noise.
- Look for the Specificity: If a review doesn't mention a specific, idiosyncratic detail about the product or service, discard it.
- Check the Velocity: A sudden spike in reviews is a red flag that no algorithm will ever catch as effectively as a human eye.
- Follow the Discourse, Not the Rating: If the platform doesn't allow for a back-and-forth between the reviewer and the business, the information is static and likely useless.
The CMA investigation is a press release disguised as progress. It will result in some fines, some updated "Terms of Service," and a lot of self-congratulation in Whitehall.
But it won't make the internet more honest. It will just make the lies more sophisticated.
Stop waiting for the government to tell you who to trust. They are the last ones to know.
Trust is not something that can be regulated into existence. It is earned through transparency, consistency, and the freedom to fail—none of which are served by a bureaucrat with a checklist. The review is dead. Long live the reputation.
The investigation into Just Eat and Auto Trader isn't the beginning of a cleaner internet. It's the funeral for the idea that a star rating ever meant a damn thing.
Stop looking at the stars and start looking at the incentives. That's where the truth is hiding.
