Trust online used to be a matter of instinct: a professional-looking website, a familiar brand, a friend’s recommendation. Increasingly, it is being turned into numbers. Platforms rate sellers and drivers, services score reviewers for reliability, identity systems assign confidence levels to accounts, and researchers attempt to measure the trustworthiness of information sources themselves. What was once a social judgement is becoming an infrastructure, calculated continuously and used to decide what people see, buy and believe.
The drivers are easy to understand. At the scale of modern platforms, human judgement cannot review every transaction, listing or post, so measurable proxies take over: completion rates, dispute histories, verification checks, behavioural signals. The World Economic Forum and other bodies have examined how such systems might support safer digital economies, particularly where formal institutions are thin. Follow the discussion through our homepage and the appeal is clear: measured trust lets strangers cooperate with a confidence that once required personal knowledge.
From Gut Feeling to Metrics
Measurement, however, changes the thing it measures. Once a score affects income or visibility, people optimise for the score, sometimes by improving their service and sometimes by gaming the metric. Ratings inflate as reviewers fear retaliation or feel social pressure to be generous. Newcomers face a cold start, distrusted by default until they accumulate history, while established accounts can coast on reputations earned under different conditions. A number that began as a summary of behaviour becomes a target, and then, gradually, a distortion of it.
Some designers argue for a different ambition: not scoring people, but verifying specific claims. A system might confirm that a seller shipped previous orders, that a reviewer actually made a purchase, or that an account belongs to a real person, without compressing a human being into a single grade. Claim-by-claim verification is narrower, but it travels better between contexts and leaves less room for quiet discrimination. The technical tools for such selective disclosure now exist; the argument is over business models and habits, not feasibility.
Who Gets Measured, and Who Decides
Questions of control follow close behind. Who designs the measure, who can see it, and who can appeal when it is wrong? A low score with no explanation can function as a punishment without a hearing, affecting livelihoods in ways that feel arbitrary to those on the receiving end. Regulators are beginning to ask whether significant trust and reputation systems should meet standards of transparency and contestability similar to those applied to credit scoring, where the consequences of error have long been recognised.
The Limits of a Trust Score
There is also the matter of context. A person may be a meticulous seller, a careless commentator and a scrupulously honest reviewer all at once, yet systems prefer a single portable figure. Efforts to make trust credentials portable between services raise the stakes further, promising convenience while risking a world in which one platform’s judgement follows a person everywhere. Designers of these systems increasingly acknowledge the tension, even when they disagree about where the balance should sit.
None of this means measurement is a mistake. Informal trust had its own injustices, favouring the well-connected and the familiar. The mature position is to treat trust metrics as tools with known failure modes: useful signals, deserving of audit, explanation and appeal, and never a complete substitute for judgement. As more of economic and civic life moves online, getting that balance right becomes a first-order question. Our technology coverage continues to follow it.
Whatever form it takes, measured trust will need governance proportionate to its power. Audit trails, human review of consequential decisions, and clear routes to correct bad data are becoming the expected minimum rather than gold-plating. Users, for their part, are learning to read scores sceptically, asking what was actually measured and what was left out. Trust was never really a feeling that technology replaced. It is a relationship that measurement can support, distort or damage, depending on how carefully the measuring is done.
