If you want people to behave well, rating them looks like an efficient shortcut. In practice, rating humans with five little stars mostly builds a thin layer of anxiety on top of already messy systems. And things become dystopian’ish fast.
The core problem is simple: we try to compress complex behavior, unequal power, and shifting moods into a single metric, then pretend it is neutral. That is good for dashboards, bad for dignity.
The illusion of accountability
The pitch for ratings is straightforward: more data, more accountability. The reality is closer to mutual surveillance with no clear standard.
On ridesharing platforms, both driver and passenger know they are being scored. So they shape their behavior for the metric, not for the moment. Drivers are told to smile, chat, offer water, and keep their average above a mostly invisible cutoff. Passengers get nudged to provide “honest feedback” in under ten seconds.
The result is not real accountability. It is a choreographed interaction where both sides are quietly defensive. Each knows a stray three star rating from the other can drain income or raise prices. What looks like a shared safety mechanism behaves more like a network of small, mutual threats.
The five star scale makes this worse. In theory it offers nuance. In practice, everyone learns the same rule: five or you’re in trouble. Four stars becomes a warning, not a mild “could be better.” The lower end of the scale may as well not exist, except as a weapon.
A few things follow:
- Ratings drift toward all fives, which kills their informational value
- Small deviations become alarming, because they are rare and often unexplained
- People who depend on the platform for income carry most of the risk
You can call that “feedback”, but it functions more like a fuzzy penalty system without an appeals process.
The absurdity of customer ratings
Some environments show the crack in the logic more clearly.
Imagine a Michelin starred restaurant rating its diners, or a mountain lodge spa where massage therapists quietly score clients on a tablet afterwards. The whole point of these places is that the guest can relax inside a well run system. If guests knew they were being evaluated, the chillaxing vibes would break. You cannot build ease and constant judgment into the same room.
Hospitality should be simpler: staff will try to serve well, guests will try not to be jerks, and management will intervene when either side drifts too far off. That is already a rating system of sorts, but it is human, local, and uses judgment instead of a global score following you forever.
Having said that, try checking out 3 star restaurants. I think some of these are actually real winners. More real. Less instagram.
But overall, the more we paste formal ratings onto these interactions, the more we swap trust for theatre. People start behaving as if they are always auditioning.
When ratings replace trust
Healthcare shows how dangerous this swap can be.
A doctor already navigates a dense web of constraints: clinical evidence, patient history, time limits, and the blunt fact that not every outcome will be good. Add public star ratings from patients, and you introduce a new pressure: keep the mood pleasant, or risk a visible scar on your record.
That can pull in the wrong direction:
- Being honest about a hard diagnosis may feel riskier than soft pedaling
- Saying “no” to an inappropriate request for antibiotics may invite a bad review
- Spending longer with one complex case can lower “throughput” metrics and anger others
The relationship is supposed to rest on candor and trust, not satisfaction scores. Turning every visit into a micro performance review bends the incentives toward comfort and away from clarity.
Even in more transactional settings like ridesharing, the pattern holds. People get low scores for traffic they did not cause, for being too quiet, for being too talkative, for not sharing the small talk script in someone’s head. The metric records mood as if it were fact.
Philosophically, this is where things get odd. Kant’s idea that we should treat people as ends in themselves sits uneasily next to a system where your worth as a driver, guest, or patient is frozen into a number on someone’s phone. Utilitarian arguments about “maximizing happiness” do not help much if the measurement tool is this blunt and this entangled with power.
Our discomfort with ambiguity
The technology here is not the root problem. The underlying habit is older: we do not like ambiguity, so we invent fast proxies.
A written note, a real conversation, or a local complaint process can all handle nuance. They are slow and sometimes awkward. A rating button is instant, low friction, and easy to average. It feels scientific without asking much of the rater.
So we get:
- People using ratings to vent about parking, weather, or their own bad day
- Companies optimizing for the appearance of satisfaction instead of deeper quality
- Individuals who depend on these scores trying to guess which invisible line they crossed
Instead of wrestling with messy feedback, we route everything through a single number and tell ourselves it is “data driven.”
The irony is that these systems are marketed as trust machines. In practice they often hollow trust out. Real trust comes from repeated interactions, honest repairs after mistakes, and a sense that both sides can afford to be human. A five star average is not a substitute for that; it is a summary statistic sitting on top of whatever is actually going on.
Who rates the raters?
One more asymmetry matters: the system itself is nearly unrateable.
The people most exposed to star ratings rarely designed the rules. They just live under them. Product managers and executives choose the thresholds, the prompts, and the ways ratings affect pay or access. Users are then asked to grade one another inside that scaffolding.
The Walmart example makes this concrete. You go in to buy a $30 LEGO set for your kid. It is locked in a plastic cabinet. You scan a code, wait around, and eventually an overworked associate sprints over to unlock it. Later, the app asks you to rate the “experience” on a five star scale, complete with options like “not friendly” or “not helpful.”
The real bug here is the process design: locking small, low risk items, under staffing the floor, and pushing the annoyance downward. The associate is just trying to keep up. You are effectively being asked to punish the person at the end of the chain for a system level decision they did not make and cannot change.
There is no parallel control to give the process itself a low score in a way that carries weight. The architects of the incentive structure and store layout sit behind the glass. The people you see are the only ones you are allowed to rate.
The same pattern shows up across sectors:
- Drivers carry the weight of routing, pricing, and matching algorithms
- Nurses and doctors feel pressure from patient scores shaped by wait times and insurance foibles
- Frontline staff absorb anger that belongs to policy, not to them
If we were serious about accountability, we would route more of the feedback up the chain instead of concentrating it on the nearest human.
Choosing where ratings belong
None of this means we should torch every feedback mechanism. Some form of structured review is useful for catching real patterns of abuse, discrimination, or negligence. People want a way to warn one another about bad actors.
The question is narrower: where do star ratings help, and where do they quietly corrode the very trust they claim to support?
A more careful design would:
- Use ratings sparingly, for behavior that is repeatable and clearly within someone’s control. Try sitting in an Uber quietly and politely 100 times. You won’t get 5 stars each time, but you should.
- Combine them with qualitative feedback and local judgment, not treat them as a final word.
- Make system level design choices visible and open to critique, not hide them behind worker scores .
- Give people who are rated meaningful ways to contest, contextualize, or repair their record.
Treat ratings like a sharp tool in a crowded workshop: useful in the right context, dangerous if left lying around everywhere.