If online reviews matter less than they used to, what’s actually replacing them? Most operators reach for an answer without thinking hard about it: nothing has replaced them. A strong star rating still decides who gets the callback. That was true for a long time. The current, more rigorous data says something narrower and less comfortable. The way people discover and evaluate a company has shifted toward social and community signal, and it has shifted faster than trust in that particular channel has caught up. Customers are not moving toward something more credible than a review. They are moving toward something more native to how they already spend their day. The space between those two facts is the actual problem.
“Reviews still matter most. If our star rating is good, we’re covered.” That assumption sits behind most operator marketing budgets right now, and it survives because it used to be a safe bet. It is getting less safe every quarter, not because a more trustworthy alternative appeared to take the reviews’ place, but because a more convenient one did.
What’s actually replacing reviews
McKinsey published “From likes to buys” in July 2026, drawing on its own State of the Consumer Survey fielded that March and April. The chart lands on the finding this whole article is built around. Social media is Gen Z’s single most important channel for discovering and buying almost anything. Gen Z shoppers say it plays a key role in a purchase decision at roughly double the rate baby boomers do: 34% versus 16%. In the identical survey, the same generation ranks social media among the least trusted channels they actually use, behind online reviews, behind even a generative AI answer.
Read that pairing straight and it says something narrower than the version usually making the rounds in marketing decks. It is not that social discovery has become more trustworthy than a review. McKinsey’s own data says closer to the opposite: reviews are still trusted more. What changed is where people go first, not what they believe once they get there. A generation can use a channel constantly and rank it near the bottom on trust among its own options at the exact same time. That combination, not a trust upgrade, is the shift worth planning a marketing budget around.
Picture what that looks like on the ground, not in a slide deck. It is not a stranger reading a five-star review and booking. It is a comment thread under a loading-day video, a friend tagging a company in a group chat because they used them last spring, a short clip that shows up while someone is scrolling for something else entirely and happens to stick. None of that carries the deliberate, audited quality of someone sitting down to read twelve reviews before making a call. It carries something else: it is simply where the person already was.
For an operator, the practical consequence lands on the exact marketing line item most likely to sit unfunded: a real, sustained presence on whatever platform carries this kind of casual discovery in the local market, not a boosted post twice a year. A company might treat this channel as a lower priority because it has not “proven” itself the way a review page did. That company is optimizing for a discovery moment fewer of its future customers will ever pass through.
What years of calm actually proved
For most of the last fifteen years, a strong review average looked like a stable, durable kind of trust. Nassim Taleb’s own point about stability is worth borrowing directly here: a system that has not been tested by real stress is not evidence that the system is sound. It is only evidence that the test has not happened yet. Reviews got their test, and it did not go well.
The Federal Trade Commission’s final rule banning fake reviews and testimonials has been in effect since October 21, 2024. It exists because the agency had already seen enough evidence that the review system needed a government fix. The rule bans paid five-star reviews, undisclosed insider reviews from a company’s own staff, fabricated testimonials for people who never used the product, and the outright suppression of negative reviews through threats or intimidation. Violators face civil penalties up to $51,744 per violation, for anyone who knew or should have known better. A regulator does not build a rule with teeth that sharp over a channel that is working fine.
The economics behind the rule are just as concrete. In 2023, the economists Akesson, Hahn, Metcalfe and Monti-Nussbaum ran an incentive-compatible experiment with 10,000 real shoppers for a paper published by NBER, “The Impact of Fake Reviews on Demand and Welfare.” Fake reviews, they found, steer people toward the worse product on the shelf. In the setting they tested, that costs roughly twelve cents of every dollar spent. The people who paid that cost were not the platforms hosting the fake reviews, and they were not the sellers who bought them. The businesses with real skin in the game absorbed the damage a shared trust signal inflicted on everyone once enough of it turned out to be fake. Those are the ones with an honest track record and no fabricated reviews propping up their average.
That is the honest reason reviews are losing ground. Not because a better, more verified channel came along and beat them fairly. Because the channel itself took real, documented damage, at the exact moment an easier alternative happened to be sitting one tap away.
People go where it’s easy, not where it’s earned
None of that explains why the traffic moved specifically to social media instead of moving toward some more rigorously verified review platform. The honest answer has less to do with trust than with plain psychology, and Rory Sutherland’s own habit of reading behavior instead of logic is useful here: people rarely choose the option a spreadsheet would recommend. They choose the option already open in front of them.
A teenager deciding where to eat lunch is not running a trust audit across five platforms and picking the one with the strongest verification standards. They are opening the app that is already open, and that app happens to be built for exactly this kind of casual, high-frequency decision: a friend’s video, a comment thread, a location tag someone else already attached. A review site was built to be searched, deliberately, with intent. A social platform was built to be scrolled, and scrolling wins by default because it asks nothing of the person doing it.
That is the reframe worth sitting with, not the comfortable one about a ratings problem. Something more convenient replaced reviews, not something more trustworthy, and convenience does not need to win an argument about credibility to win the traffic. A company cannot out-trust a channel its customer was never fully trusting to begin with. It can only show up where the customer already is. That is a genuinely different project from the one most operators think they are running when they add one more review widget to the homepage.
The gap the data doesn’t cover
Every figure in the last two sections comes from research on browsing, eating out, or buying something small enough that a bad choice barely matters. Moving a household is none of those things. It happens once every several years, involves real money, and carries a real cost if the company turns out to be wrong for the job. Honestly: no Tier 1 source was found that specifically measures social-driven discovery for a considered, infrequent purchase like a household move. Forcing a restaurant-recommendation statistic to carry that specific claim would be a worse mistake than simply leaving the gap visible.
What does exist is a real, verified data point about search behavior itself, not moving specifically. In July 2022, TechCrunch reported comments from Prabhakar Raghavan, the senior vice president running Google’s own Knowledge and Information group. Raghavan was relaying internal research at a conference, not a published study, when he said: “almost 40% of young people, when they’re looking for a place for lunch, they don’t go to Google Maps or Search. They go to TikTok or Instagram.” That is Google’s own internal data about search habits broadly, not a moving-industry finding. It should be read as exactly that: evidence that an entire generation’s discovery reflex has moved, not proof that the same reflex governs a five-figure household decision.
The honest extension is a reasoning claim, not a data claim, and the difference matters. If the discovery reflex for something as low-stakes as lunch has already moved this far, the same generation is not reverting to an older search habit the moment the decision in front of them gets bigger. What changes for a bigger decision is not which app the person opens first. It is how much verification they demand before they act on what they found there. That demand is exactly where a business still has real room to earn the trust the discovery channel itself was never built to supply.
Showing up is not the same job as earning trust
None of this argues for abandoning reviews, or for chasing every platform where a customer might conceivably be scrolling. Hamilton Helmer’s own discipline is worth applying directly here: before spending a dollar or an hour on a channel, ask what it actually protects, not just whether it is popular this quarter. A presence on a platform nobody else in the local market has bothered building can function like a real advantage. A presence assembled by copying whatever the loudest competitor did last week cannot, because anyone can copy it by next Tuesday for the same effort it took the first time.
For an operator, that discipline usually points somewhere unglamorous. It is not five half-maintained accounts spread across every platform a teenager mentioned once. It is one habit, kept: real loading-day footage, unedited, posted consistently on whichever single platform the local target audience already treats as a search engine, with a real name attached to it. That is not a trust play. It is a presence play, and it only starts doing trust work once someone actually finds it. The same footage does double duty: it also wins over the second decision-maker who never heard the original call, not just the algorithm.
Helmer names seven real sources of durable advantage: scale economies, network effects, brand, switching costs, a cornered resource, process power, and counter-positioning. None of them cares which app a company happens to post on. Each one asks whether a competitor can copy the position by next Tuesday. Maintained consistently, long enough to become recognizable, a single channel edges toward the one power on that list an unfunded marketing line item can build: brand. Five neglected accounts, refreshed whenever someone remembers, build nothing a competitor could not copy by opening the same five apps.
That framing answers the objection an experienced operator will already be forming: does any of this mean reviews stop mattering? No. The NBER research above only makes sense because real reviews still carry real evidentiary weight, enough that faking them was worth the fraud in the first place. The point is not to replace a review strategy with a social one. It is to stop assuming the review page is where a customer’s decision actually begins, when the data increasingly says it is where the decision gets confirmed, several steps after the discovery already happened somewhere else.
This is a different question from who owns a review once a customer leaves one, which already has its own honest answer elsewhere: why your reviews belong to the platform covers what happens to that review the moment it lives on someone else’s site instead of yours. This piece is about an earlier moment. It is about which door the customer walked through before a review ever entered the picture, and whether that door currently has any version of the business standing in it.
The search you haven’t tried
A company waiting for social discovery to become respectable enough to invest in is waiting for a permission slip that may never arrive. The audience it’s trying to reach has already moved on, regardless of whether the channel ever earns their full trust. The real business risk here is not that the newer channel is unproven. It is that waiting for proof is itself a decision, and it is the decision that leaves the door unattended while a competitor, intentionally or not, stands in it instead.
When was the last time you searched for your own company the way an 18-to-24-year-old actually would? Not a Google search for the business’s own name. The other one: open TikTok or Instagram, search the city and “movers” or “moving company,” and see who actually shows up. If the honest answer is nobody, including the business asking the question, then a real share of its next customers are already using that channel to decide who gets the call, and the business currently has no version of itself standing in it at all, trusted or otherwise.
