Credence Goods and the Trust Gap: Why Legal, Dental, and Medical-Aesthetic Buyers Search Differently Than Retail Buyers

The trichotomy that explains the whole problem

Economists sort products by when a buyer can judge their quality. Phillip Nelson drew the first line in 1970: a search good can be assessed before purchase (you can read a laptop's specification, compare prices, confirm a lamp is in stock), while an experience good can only be judged after using it (a meal, a haircut, a hotel stay). Michael Darby and Edi Karni added the harder third category in 1973: a credence good is one whose quality the buyer cannot confidently verify even after consuming it, because doing so would require expertise the buyer does not have.

Most of local commerce is a mix of the first two. A hardware store sells search goods. A restaurant sells an experience good, which is why a synthesized rating carries so much weight there. But legal, dental, and medical-aesthetic services fall squarely in the third category. A patient cannot tell, from the outcome alone, whether a dental crown was necessary or oversold, whether an injectable was administered under proper supervision, or whether a settlement was the best that skilled counsel could have reached. The information gap does not close at purchase. It often never closes at all.

That single distinction reorganizes everything downstream, including how these buyers search, what they look for, and what visibility work actually has to accomplish.

Why the credence buyer cannot do what the retail buyer does

The retail buyer runs a verifiable checklist. Is the price fair, is the item in stock nearby, does the specification match the need. Those questions have answers a search result can supply directly, which is why local retail discovery is dominated by concrete, checkable intent such as a product name plus "near me" and an in-stock signal.

The credence buyer has no equivalent checklist for the thing that matters most. The evidence is sharpest in aesthetics, where the credential that actually governs safety is frequently invisible to the person choosing. A peer-reviewed study of plastic-surgery provider selection found that patients routinely cannot verify the credentials that govern procedure safety, and that non-board-certified practitioners performing aesthetic procedures is a documented patient-safety concern. The buyer standing in front of two clinics often cannot distinguish a supervising-physician-run practice from an unsupervised one on sight.

The regulatory reality underneath makes the gap concrete. In the US, medical-aesthetic procedures are legally the practice of medicine, and physician-supervision requirements vary sharply by state, from close on-site oversight in some jurisdictions to no supervision requirement at all where nurse practitioners hold full independent scope of practice. The very fact that decides whether a treatment is being delivered safely is a legal structure the buyer has almost no way to read from a listing, a website, or an answer engine.

The regulator stands in for the verification the buyer cannot do

If buyers in these verticals cannot verify quality themselves, someone has to substitute for that verification. That is the most useful way to read the regulatory apparatus around credence goods: it is society's attempt to install institutional guardrails where individual verification is impossible.

Attorney advertising is the clearest case. The American Bar Association's Model Rule 7.1 imposes an affirmative truthfulness standard, not merely a rule against lying. A statement that is literally true is still prohibited if it omits a fact necessary to keep the whole communication non-misleading, or if it would lead a reasonable person to a conclusion for which there is no reasonable factual foundation. Rule 7.2 separately bars paying for referrals except through qualified, unbiased not-for-profit lawyer referral services. These rules exist precisely because a prospective client cannot independently check a lawyer's claims before hiring, so the profession polices the claims instead.

The same logic runs through the other two verticals. State medical-board supervision law governs who may perform an aesthetic procedure and under what oversight. And the Federal Trade Commission finalized a rule in 2024, effective October 21, 2024, making fake and deceptive consumer reviews and testimonials a specified unfair-or-deceptive act, applying to every reviewed local business rather than only to platforms. Read together, Rule 7.1, state supervision law, and the FTC review rule are three different regimes solving the same problem: they vouch for signals the buyer cannot verify on their own.

So the buyer searches for proxies, not products

Denied the ability to inspect the product, the credence buyer does something rational: they hunt for observable proxies of an unobservable quality. Instead of price and specification, they look for credentials, board certification, a supervising physician, a verifiable license, third-party corroboration, and the accumulated narrative of people who came before them.

Reviews function differently here than in retail, and the difference is measurable. A large topic-modeling study of online health-community reviews found that narrative reviews measurably shift which provider a patient picks, and that the content of the review matters: clinical-skill narratives and service narratives differentially predict choice. The patient is not scanning for a star average the way a shopper scans a product page. They are reading the reviews as testimony about a competence they cannot assess themselves, which is why a credible narrative about clinical skill carries weight that a generic five-star burst does not.

This is also where search-engine trust frameworks intersect the economics. Google's quality guidelines treat health, legal, and financial topics as Your Money or Your Life subjects held to a higher bar, and its E-E-A-T framing (experience, expertise, authoritativeness, trustworthiness) is, in effect, an attempt to score the same proxies the buyer is hunting for. The engine and the buyer are looking for the same thing: evidence that the provider is who they claim to be and competent at what they claim to do.

How the asymmetry changes the visibility craft

Once you accept the trichotomy, the job of visibility work changes shape by vertical. For a search good, the work is largely about presence and availability: be findable, be nearby, show the price and the stock. For a credence good, presence is necessary but nowhere near sufficient. The core work is making the unverifiable verifiable, taking the credentials, supervision structure, licensure, and real client outcomes that a buyer cannot see and rendering them legible, consistent, and corroborated across every surface where the buyer now looks.

Why black-hat tactics fail twice in these verticals

In a search-good market, a deceptive tactic risks one penalty: a platform sanction. In a credence-good vertical it risks two at once, because the same act violates platform policy and professional-conduct or consumer-protection law simultaneously.

A fabricated review breaches the platform's terms and, since October 2024, the FTC's fake-review rule. A misleading practice-area claim breaches search-engine spam policy and ABA Rule 7.1. An overstated aesthetic outcome breaches ad policy and, potentially, state medical-board advertising rules. Substantiated claims and real reviews are not an optional style choice in these verticals. They are the only method the law and the platforms leave standing, because the shortcuts are barred twice over and the buyer, unable to verify quality directly, is unusually exposed to being misled.

The retail contrast, made concrete

It helps to see how differently the same reputation signal behaves outside the credence category. The foundational quasi-experimental study of online ratings, Michael Luca's work on Yelp, found that a one-star increase in rating produced a 5 to 9 percent increase in restaurant revenue, and that the effect was concentrated in independent restaurants while chain-affiliated restaurants showed no rating effect, plausibly because a recognized brand already supplies consumers with a quality prior.

A restaurant is an experience good: the synthesized rating substitutes for the meal you have not yet eaten, and it works because after one visit you can verify the experience yourself. In a credence good the substitution is only partial. A patient can read that a clinic has excellent reviews, but the reviews cannot fully certify the thing that matters, whether the clinical judgment was sound, because most reviewers could not verify that either. This is why credence-good visibility leans so hard on credential and corroboration signals that a review average cannot carry alone, and why the independent-versus-chain lesson from Luca points the same way: the smaller, owner-operated practices that make up this market have the most to gain from making their trust signals legible, because they have no national brand supplying the prior for them.

How to read this

Two things in this framework are firmly established, and one is our own synthesis, and it is worth keeping them separate. The economics of search, experience, and credence goods is settled theory dating to Nelson in 1970 and Darby and Karni in 1973. The vertical evidence anchoring it, the credential-verification gap in aesthetics, the state-by-state supervision law, ABA Rule 7.1, the FTC review rule, and the reviews-and-choice literature, is real and attributable.

What is RavenEye's own contribution is the application: mapping that economic trichotomy onto how these verticals should be made visible. That mapping is analysis, not adjudicated fact, and where it touches legal-advertising exposure it is a framework for thinking, not legal advice. The layer that complicates it further, how the newer AI answer engines weigh these same trust proxies, is still emerging: vendor studies report that ranking well in the classic local pack is a weak predictor of being named in an AI local recommendation, but those figures come from single-vendor proprietary data and should be treated as directional. The reliable posture is to act firmly on the established economics and the documented regulation, and to measure the emerging AI-answer layer rather than assert it.