The Anchor You Didn't Choose: How the First Number or Review a Buyer Sees Shapes the Rest

What anchoring actually is, and why it is not a metaphor

In 1974 Amos Tversky and Daniel Kahneman published a synthesis in Science arguing that people do not judge uncertain quantities by full calculation. They rely on a small set of heuristics, mental shortcuts that are efficient but produce systematic, replicable errors. The paper named three: representativeness, availability, and anchoring.

Anchoring is the most counterintuitive of the three. In their demonstrations, participants were shown an arbitrary starting number, generated in front of them by a spin of a wheel of fortune, and then asked to estimate an unrelated quantity such as the percentage of African nations in the United Nations. Estimates were pulled toward the arbitrary anchor even though everyone could see it was meaningless. The mind starts from the first value it is given and adjusts insufficiently away from it.

The finding matters here precisely because it holds when the anchor is known to be irrelevant. A buyer does not need to believe that the first review they read is representative for it to shape their judgment. The pull operates below deliberate reasoning, which is why order of exposure is a genuine variable and not a presentation nicety.

Anchoring bias in reviews: why order matters, not just the average

The common intuition is that a buyer forms an impression from the aggregate: the star average, the total count, the general drift of sentiment. Anchoring complicates that picture. If judgment starts from the first value encountered and adjusts insufficiently, then the first review surfaced, and the rating attached to it, does work that the average alone cannot undo.

There is corroborating evidence that buyers over-weight the visible summary number in the first place. In a study of 1,272 products across 120 categories, de Langhe, Fernbach, and Lichtenstein found that average user ratings did not reliably track independent quality scores, were often built on too few ratings to be informative, and ran higher for pricier items independent of actual quality, yet buyers leaned heavily on the star average when forming quality judgments. The authors call this a perceived-validity bias: the number feels more diagnostic than it is.

Put the two findings together and the operational reading is straightforward. The rating a buyer meets first becomes the anchor, and the average they see is trusted more than it deserves. Which review the platform shows first, and whether the visible summary is accurate, are therefore decisions with consequences, not defaults to ignore.

Why the first reviews matter more than the ones that follow

Anchoring explains why a single early impression sticks. A separate, rigorous body of work explains why early reviews compound. In 1992 Bikhchandani, Hirshleifer, and Welch formalized the theory of informational cascades: once enough people have visibly chosen an option, it becomes individually rational for the next observer to follow that choice and discount their own private signal.

The mechanism is sequential and order-dependent by construction. The first few visible choices carry weight far beyond their number because they set the direction the cascade runs. A business with two early, credible, well-detailed reviews is easier for the next buyer to choose than one with two thin or negative ones, not because the sample is large but because it anchors the sequence. The same logic warns that cascades are fragile and can flip on small new information, which is why the recency and accuracy of what shows first is not a one-time fix.

The asymmetry: a bad first signal pulls harder than a good one

First exposure is not neutral in sign. In one of the earliest causal studies of online reviews, Chevalier and Mayzlin compared relative book sales across Amazon and Barnes and Noble and found that an improvement in a book's average rating correlated with higher relative sales, with a one-star improvement associated with up to a 9.9 percent increase. Critically, they found the impact of one-star reviews was larger in magnitude than the impact of five-star reviews, an asymmetry consistent with loss aversion.

Anchoring and this asymmetry interact in a way that raises the stakes of order. If the first review a buyer reads is negative, it does double duty: it anchors the sequence, and negative signals already weigh more heavily than positive ones. A stale, unrepresentative one-star review sitting at the top of a profile is therefore not a minor blemish; it is a heavily weighted anchor doing more damage than a proportional five-star review would repair.

Anchoring effect pricing: the first number frames the next

The same mechanism governs numbers other than ratings. Because anchoring was demonstrated on arbitrary numeric anchors, the first price a buyer encounters frames how every subsequent price is judged. A quoted figure, a starting-at number, or a competitor's visible rate becomes the reference point against which the buyer decides whether the next number is high or low. The judgment is relative to the anchor, not absolute.

This is the evidence-backed core of what popular marketing calls price anchoring. The caution is equally important: the finding licenses attention to which number a buyer meets first, not manipulation. Presenting a first number that is accurate and fairly framed is defensible; engineering a misleading anchor to distort judgment is the dark-pattern version the same literature warns against.

When the engine names one business, the anchor concentrates

A reasonable extrapolation, and it should be labeled as one, is that AI answer surfaces intensify anchoring rather than remove it. When a search engine returned ten blue links, the buyer scanned several options and the anchor competed with alternatives in view. When an engine answers a query by naming a single business, that name becomes the anchor with no visible competitor beside it to adjust against.

This is an emerging thesis, not an established finding. The causal anchoring and cascade literatures were built on lists, prices, and sequential choices, not on single-answer AI surfaces, so the extension is an inference from mechanism, not a measured result. Even so, it still has a clear implication: as the presented consideration set narrows toward one, the value of being the first, and only, business the buyer encounters rises rather than falls.

What this changes for the profiles a business controls

The evidence narrows to a practical surface. A business cannot dictate a buyer's cognition, but it can influence what that buyer meets first, and the research says first exposure is where the pull is strongest. That makes the controllable inputs specific: which review and listing profiles carry the business's name, whether those profiles are owned and verified, whether the facts on them agree, and whether the rating and reviews shown first are current and representative.

Two failure modes follow directly. A profile the business does not control can surface a stale or unrepresentative review as the buyer's first impression, planting an anchor the business cannot correct. And conflicting facts across duplicate or unclaimed listings give an engine no clean signal to name confidently, so it anchors the buyer on a competitor whose details resolve cleanly instead.

None of this justifies fabricating or reordering reviews to deceive. Reviews are earned from real customers only, and selectively suppressing negative ones is both dishonest and, since the 2024 FTC rule on consumer reviews, unlawful. The legitimate work is narrower and durable: own every profile, make them agree, keep what shows first accurate and current, and respond to what is there. That is reputation groundwork, and the anchoring evidence is why the order it produces matters.