Anchoring & Adjustment: An S-Tier Behavioral Designer’s Guide to Decision Design

In 1974, Amos Tversky and Daniel Kahneman spun a rigged wheel of fortune in front of their research subjects. The wheel was gimmicked to stop on either 10 or 65. Then they asked a question that had absolutely nothing to do with the number the wheel had just shown: “What percentage of African countries are members of the United Nations?” Subjects who saw 10 gave a median estimate of 25%. Subjects who saw 65 gave a median estimate of 45%. A meaningless number, pulled from a spinning wheel, dragged the answer almost twenty points in whichever direction it landed.

That experiment is the hidden physics behind every pricing page you have ever seen, every negotiation you have ever won or lost, every leaderboard that made you feel fast or slow, and every “originally $199, now $79” tag that made a product feel like a bargain. It is the single cheapest design lever in behavioral science — and it is the lever most designers never consciously pull, which is why their competitors beat them.

The principle is called the anchoring and adjustment heuristic. I’ve spent over two decades teaching designers how Core Drives motivate behavior. What almost nobody outside the lab understands is that anchoring is the reference-point mechanism that decides which Core Drive fires, how hard it fires, and whether your user feels like they are winning or losing. Get anchoring wrong and every other Octalysis move you make is fighting a tide.

This is the full operator’s manual — what the experiments actually show, why the “adjustment” in “anchoring and adjustment” is a lie most of the time, where the effect breaks, and how to design with it without crossing into manipulation.

⚡ Speed Run Notes

• Anchoring is the human tendency to rely too heavily on the first number that enters a decision, then “adjust” from there — usually not enough.
• The anchor does not need to be relevant, or even plausible, to move the estimate.
• The “adjustment” half is usually insufficient, and for anchors you hand the user it may not be adjustment at all.
• Experts are not immune: real-estate agents and criminal judges both tracked numbers they knew were arbitrary.
• In Octalysis terms, anchors reset the reference point that Core Drive 8 (Loss & Avoidance) uses to decide what counts as a loss and Core Drive 6 (Scarcity & Impatience) uses to decide what counts as scarce.
• Common design anchors: the strikethrough “original price,” the highest-tier plan on a pricing table, the first question in a survey, the first suggested tip percentage, the top-of-leaderboard score, the round-number goal (10,000 steps, 100,000 followers).

Table of Contents

Author Credibility: Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

What Is Anchoring & Adjustment?

Anchoring and adjustment is the cognitive heuristic that describes how people estimate unknown values: they start from an initial value (the anchor), then adjust from there to reach a final answer. The problem — and the reason this matters — is that the adjustment is almost always too small, and the anchor can come from anywhere, including places that have nothing to do with the decision.

Amos Tversky and Daniel Kahneman introduced the heuristic in their 1974 Science paper, “Judgment Under Uncertainty: Heuristics and Biases.” In that paper, they argued that human judgment under uncertainty does not proceed by the calm, Bayesian calculation that classical economics assumes. Instead, it proceeds by a small set of mental shortcuts — heuristics — that usually work well enough but produce predictable, systematic errors. Anchoring was one of three heuristics they named (alongside availability and representativeness), and more than fifty years later, it is still the one that shows up most consistently in applied design.

The structure of an anchoring effect looks like this. A decision-maker faces a question they cannot answer precisely — how much is this house worth, how much should I pay for this wine, how likely is this patient to have a heart attack, what is a fair salary for this role. Rather than building the answer up from first principles, the mind grabs the nearest plausible number, treats it as a tentative answer, and then edits. The editing is usually conservative. It stops when the answer feels “about right,” not when the answer is right.

The critical implication is that the nearest plausible number does not have to be a good number. It can be the listing price in a real-estate ad. It can be the first salary offer on a hiring call. It can be the MSRP on a product box. It can be a random number pulled from a spinning wheel or the last two digits of a Social Security card. The mind does not know the anchor is irrelevant until the anchor has already done its work.

In Octalysis language, this is why anchoring is so powerful: it sets the reference point that the rest of the motivation system measures itself against. Core Drive 8 — Loss & Avoidance — only fires when there is a loss relative to some baseline, and the anchor is that baseline. Core Drive 6 — Scarcity & Impatience — only fires when a resource feels scarce relative to some supply, and the anchor sets the expected supply. Core Drive 2 — Development & Accomplishment — only fires when progress happens relative to some goal, and the anchor names the goal. Move the anchor, and you move what counts as a gain, a loss, scarce, or accomplished. Everything downstream bends with it.

The Core Findings Behind Anchoring

The Wheel of Fortune Experiment (Tversky & Kahneman, 1974)

The foundational anchoring study is a thing of beauty. Tversky and Kahneman sat subjects in front of a wheel of fortune numbered 0 to 100. The wheel was secretly rigged to land only on 10 or 65. After the wheel stopped, subjects were asked whether the percentage of African countries in the United Nations was higher or lower than the number the wheel had just shown — and then were asked to give their best estimate of the actual percentage.

The median estimate from subjects who saw 10 was 25%. The median estimate from subjects who saw 65 was 45%. A random number from a spinning wheel moved answers about the United Nations by twenty percentage points. The subjects knew the wheel was random. They could see it spinning. And it still worked.

That study is not just a cute party trick. It is the reason anchoring is categorized as a heuristic rather than a bias of ignorance. Bias-of-ignorance effects go away when people are informed or paying attention. Heuristic effects do not. Informing subjects about the randomness of the wheel weakens anchoring somewhat. It does not eliminate it.

The Social Security Digits Experiment (Ariely, Loewenstein, Prelec, 2003)

Nearly thirty years after Tversky and Kahneman, Dan Ariely, George Loewenstein, and Drazen Prelec published “Coherent Arbitrariness” — a study that is probably the most economically consequential replication of anchoring ever done.

Fifty-five students were asked to write down the last two digits of their Social Security number, then asked whether they would buy each of six products at that many dollars: two bottles of wine, a cordless computer keyboard, a cordless trackball, a design book, and a box of Belgian chocolates.

Only then were they asked the question that mattered, which was the most they would actually pay. Real money was on the line, because a random draw decided whether the sale went through at a randomly generated price — the standard Becker-DeGroot-Marschak procedure rather than an auction against other bidders.

Students whose Social Security digits landed in the top fifth were willing to pay roughly three times what students in the bottom fifth would pay for the same items. That is the authors’ own summary of the quintile pattern.

On the cordless keyboard the gap was $56 against $16, a ratio of about three and a half to one.

What stayed stable was the ordering: subjects ranked the six products against each other coherently, so the wine still beat the chocolates by a sensible margin in every group. The absolute level of what the whole set was worth to them was tethered to a number they had written down sixty seconds earlier, which is the paper’s title in one sentence.

The implication for pricing design is surgical. Consumers do not walk into a market knowing what things are “really” worth. They infer value from whatever anchor is nearby when the question is asked. If your pricing page places the enterprise tier first and your self-serve tier next to it, the self-serve tier feels cheap. If you hide the enterprise tier, the self-serve tier is just however much it costs.

The Real-Estate Agents Experiment (Northcraft & Neale, 1987)

In 1987, Gregory Northcraft and Margaret Neale ran a study that put anchoring’s reputation for “it only works on naïve subjects” permanently to rest. They took practicing real-estate agents to an actual house in Tucson, Arizona.

Every agent got a ten-page information packet with the same real data: the MLS listing sheet, six months of neighborhood and city-wide sales, comparable properties currently listed, recently sold, and previously listed-but-unsold. Agents then toured the house for up to twenty minutes with a calculator available.

One number differed between packets, and only one: the listing price. In the second of the two experiments, 47 agents saw the same house listed at $119,900, $129,900, $139,900, or $149,900, around a property whose real listing price was $134,900 and whose appraisal a year earlier was $135,000.

The agents’ mean appraisals ran $114,204, $126,772, $125,041, and $128,754 across those four conditions. A $30,000 swing in the anchor moved professional appraisals by $14,550, on the same house, with identical underlying data, and the trend was statistically significant on all four of the study’s estimate measures.

Worth noticing that the two middle conditions came out in the wrong order, which is what a real effect with real noise looks like rather than the clean straight line the result usually gets drawn as.

Then the part that makes this study famous. The questionnaire asked each subject to check which of sixteen factors they had used, and 81% of the agents did not check the listing price at all — against 63% of the student amateurs, who were measurably more willing to admit it. The paper is careful about what that does and does not prove: it says plainly that it remains an open question whether the agents were unaware of using the anchor or simply unwilling to say so.

One more finding from this study, because it contradicts the advice designers usually take from it. Northcraft and Neale expected that an anchor further from the true value would bite less, and they tested it directly. It got only weak support in the first experiment and no support at all in the second. The agents themselves claimed a 5% credibility window, and they were significantly influenced by listing prices sitting 11% outside it.

The Judges and the Dice Experiment (Englich, Mussweiler, Strack, 2006)

If agents get anchored, maybe judges don’t. Birte Englich, Thomas Mussweiler, and Fritz Strack tested that. They gave experienced German criminal judges a case file for a hypothetical shoplifting case, asked them to roll a pair of dice that had been secretly loaded to land on either 3 or 9, and then asked them to recommend a prison sentence in months.

Judges who rolled 3 recommended an average of 5 months. Judges who rolled 9 recommended an average of 8 months. A pair of dice, openly rolled by the judges themselves, with no pretense of relevance to the case, shifted sentences by roughly three months on average. These are trained legal professionals. They know the law, the precedents, the sentencing guidelines. The anchor still got them.

The Negotiation First-Offer Effect (Galinsky & Mussweiler, 2001)

Adam Galinsky and Thomas Mussweiler demonstrated in a series of negotiation studies that the party who makes the first offer — and the size of that offer — systematically shifts the final settlement in their favor. Across laboratory negotiations over goods, services, and salaries, the correlation between first-offer anchor and final price is consistently strong. Galinsky, Ku, and Mussweiler’s review of this work reports that on a single quantifiable issue such as price, the extremity of the first offer can account for up to half of the variance in individual outcomes.

The moderator in their original experiments is the one worth knowing, because it is the counter-move. Whichever side made the first offer did better, and that advantage disappeared when the other party deliberately focused on information inconsistent with the offer: their own alternatives, their own reservation price, or their own target. Three separate focus manipulations, same result each time.

So if someone anchors you, re-anchoring is not the only defence available. Attending hard to your own walk-away number erases most of their advantage.

That finding alone has probably transferred more money in the last two decades than any other behavioral-economics result. It is why every “how to negotiate your salary” guide worth reading tells you to anchor first with a high, specific number, not a round one, and not a range.

What Tversky & Kahneman Got Right

The reason anchoring outlived the 1970s and still shows up in product decisions in 2026 is that Tversky and Kahneman got four things correct that most subsequent theories got wrong.

1. They described a mechanism, not a folk observation

“People are bad at estimates” is a folk observation. It is true but useless. What Tversky and Kahneman gave you instead is a mechanism: start from a nearby number, then adjust insufficiently. Once you know the mechanism, you can predict where anchoring will happen and where it won’t. You can predict that strangers will anchor more than experts on unfamiliar topics. You can predict that round anchors produce sloppier adjustments than specific anchors. You can predict that high-confidence anchors (from authorities) will produce smaller adjustments than low-confidence anchors (from random wheels). None of those predictions are available if your model is “people are bad at estimates.”

2. They showed the effect survives awareness

Most cognitive biases weaken dramatically when subjects are warned about them. Anchoring does not. The Tversky-Kahneman subjects could watch the wheel spin. The Northcraft-Neale agents could see the listing price and knew it was just a listing price. The Englich judges rolled their own dice. In each case, the subjects had every reason to know the anchor was irrelevant — and the anchor still worked. That is the finding that makes anchoring a design problem rather than an education problem. You cannot educate your users out of it. You can only choose, as a designer, whether to use it ethically or not at all.

3. They generalized across domains

Some behavioral findings are domain-specific — they work for snack foods but not for surgeries, or for laboratory games but not for real-stakes decisions. Anchoring has replicated across pricing, negotiation, appraisal, sentencing, diagnosis, salary, auctions, forecasting, and general knowledge. Furnham and Boo’s 2011 literature review walked the first thirty-seven years of this work across dozens of domains and found the effect robust in almost all of them. The pooled estimates came later, and they are discussed further down. Few behavioral-economics results are that portable.

4. They made it operational

The anchoring literature turned into a design toolkit faster than almost any other behavioral-economics finding. Within a decade of the 1974 paper, salary-negotiation coaches, pricing strategists, and ad copywriters were teaching “anchor high,” “strikethrough the old price,” and “start with the most expensive option.” The academic community occasionally complains that the practitioners oversimplified the theory. They did. But the practitioner version works, which is more than can be said for many academically precise findings that never made it out of the journals.

Where Anchoring & Adjustment Falls Apart

Anchoring is robust, but it is not universal. A designer who treats it as a universal lever will eventually push it into a context where it breaks, and the product will underperform in ways that look mysterious until you realize you were counting on an anchor that failed to anchor.

1. The “adjustment” metaphor is probably wrong

This is the critique that has gained the most ground in the psychology literature in the last twenty years. Tversky and Kahneman described the process as a two-step: anchor first, adjust second. The “adjustment” language implies a deliberate mental operation — starting from the anchor and walking toward a better answer. But in a series of studies starting with Mussweiler and Strack in 1997 and continuing through Epley and Gilovich in 2005, researchers have shown that the “adjustment” part of anchoring only happens for self-generated anchors (like “the boiling point of water on Everest must be lower than 212°F, so… maybe 190°?”). For external anchors — the kind that dominate real-world design — the mechanism is not adjustment at all. It is selective accessibility: the anchor activates a network of features and facts consistent with the anchor being correct, and those features then get used to construct the answer.

The design consequence is that the old advice of “make the anchor feel close to the desired answer, so there’s less adjustment needed” is partly wrong. What actually matters is how plausible the anchor makes a particular answer feel. A $10,000 anchor for a coffee mug does not produce a $10,000 sale. It produces a weird, confusing experience that triggers suspicion. The anchor has to be plausible enough to activate the right features, not just high enough to drag the adjustment upward.

2. Accountability and direct feedback partially erase the effect

Accountability and feedback both weaken anchoring in several studies, and no condition yet reported makes it disappear. The size of that weakening is where the literature gets interesting, and where this page used to be wrong.

Paying people for accuracy mostly does not work. Röseler and Schütz’s 2022 meta-analysis found no reliable effect of incentives on the magnitude of anchoring at all.

Simmons, LeBoeuf, and Nelson (2010) explained why, and their answer is more useful to a designer than a percentage would have been. People fail to adjust because they do not know which way to adjust, and that uncertainty makes them feel they have already gone too far. When the direction of the correct adjustment was made obvious, accuracy motivation did move estimates away from the anchor, for anchors the experimenter provided as well as anchors the subject generated.

So the lever is not effort. It is whether the user can tell which side of the anchor the truth is on. A pricing page where the user has no idea whether the fair price is above or below your first number is a page where no amount of care on their part will unstick them.

The practical implication is that anchoring is strongest in one-shot, low-accountability, no-feedback situations. That is most consumer pricing. It is most first-impression UX. It is most cold-call negotiation. It is far weaker in recurring B2B transactions where the buyer runs comparisons, tracks vendor pricing over time, and has to defend the purchase to a procurement committee. Designing for professional buyers as though you were designing for impulse shoppers is a classic failure of over-generalizing anchoring.

3. Extreme anchors trigger suspicion rather than assimilation

Here is where this page had the literature backwards, and the correction is more interesting than the error.

The obvious intuition is that a wild number gets thrown out and resets the judgment. Mussweiler and Strack tested that directly in a paper titled “Considering the Impossible: Explaining the Effects of Implausible Anchors,” and they found the opposite. Implausible anchors still anchored, often more strongly than plausible ones, and subjects compared themselves against the impossible anchors faster.

Their earlier work had already shown it: asked whether Gandhi died before or after the age of 9, or before or after 140, subjects’ estimates still moved with the anchor. Nobody believes Gandhi died at 9.

The honest 2026 version has two parts. Very extreme offsets do reduce the magnitude, so there is a decay curve rather than a cliff, and a preregistered 2024 pair of experiments by Mayer and Rebholz found more extreme anchors producing stronger effects with no detectable effect of whether the anchor was relevant at all.

What collapses when you put a transparently fake number in front of someone is not the anchoring. It is the trust, and in some jurisdictions the legality.

The Groupon-style “originally $799, now $29” anchor was already losing force by the mid-2010s because consumers started noticing that the original prices were fictional. When an anchor becomes transparently false, it stops being an anchor and starts being a reason to distrust the whole offer. The ceiling for plausible anchors is higher than intuition suggests but much lower than unlimited.

The Brain on Anchoring

What is actually happening in the brain when an anchor is set? The neuroscience of anchoring arrived long after the behavioral work and it is much thinner, so three strands are worth knowing and each one is narrower than what tends to get claimed for it.

The anchor is processed as evidence, not as noise. The best-established imaging result here is Tamir and Mitchell’s 2010 PNAS study, which watched people estimate another person’s preferences from their own and found dorsal medial prefrontal cortex activity tracking the adjustment away from the self-anchor. Qu, Zhou, and Luo reported the electrophysiological signature the year before that, in an EEG study of the adjustment process rather than an fMRI one.

What both lines agree on is the thing that matters for design. The brain treats a supplied number as information about the answer rather than as a stimulus to be filtered out, which is exactly what the selective-accessibility account predicts: the anchor cues memory toward the features that make the anchor look correct.

Be careful with anyone who tells you anchoring “lives” in a particular structure. The hippocampal-formation and ventral-striatum evidence that gets quoted for that claim comes from a 2024 preprint that has not yet been peer reviewed, and no settled neural model of anchoring exists today.

Anchoring is fast and cheap; debiasing is slow and expensive. The cognitive load of forming an anchored judgment is significantly lower than the load of consciously overriding it. Studies using cognitive-load paradigms (adding a secondary task to the main estimation task) show that anchoring strengthens under load and weakens slightly under deliberate thought. This is the System 1 / System 2 signature Kahneman would later make famous in Thinking, Fast and Slow: anchoring is System 1 doing its job at high speed, and it takes deliberate System 2 effort to pull the answer away from where System 1 placed it.

An anchor changes the feeling, not only the estimate, and the mechanism is less settled than it is usually sold as. Reward-prediction systems are reference-dependent, which is the foundation Prospect Theory rests on and the reason a price shown against a higher price feels different from the same price shown alone.

Reaching from that to a specific dopaminergic response to a specific strikethrough price goes further than any published study supports, and this page used to make that reach. Treat the felt-value shift as a well-replicated behavioral finding with an unfinished neural story.

Anchoring vs Other Theories

Anchoring gets filed alongside availability, representativeness, framing, and loss aversion as though the five were one family. They are not one family. Tversky and Kahneman’s 1974 paper named three heuristics, and anchoring was one of them; framing and loss aversion arrived with Prospect Theory in 1979 and the framing work that followed, and they describe how value is evaluated rather than how a quantity is estimated. The five are related and designers routinely confuse them, so here is a crisp comparison:

Anchoring vs Framing. Framing is about the wording of a choice: “90% survival rate” versus “10% mortality rate” describe the same fact, but the wording changes the emotion. Anchoring is about the number near the choice. The two often stack — a frame can name a high anchor, and a high anchor can bias how a frame is received — but they are mechanistically distinct. I wrote a full pillar on framing; the quick version is that framing is verbal; anchoring is numerical.

Anchoring vs Availability. The availability heuristic says we judge probability by how easily examples come to mind. A recent plane crash makes air travel feel more dangerous. The anchor there is not numerical — it is an example. Anchoring is the numerical cousin of availability: we judge quantities by the nearest number, the way we judge probability by the nearest example.

Anchoring vs Loss Aversion. Loss aversion says we weight losses roughly twice as heavily as gains. But loss aversion only works relative to a reference point, and the reference point is set by the anchor. Anchoring is the scaffold; loss aversion is what loads onto it. Prospect theory — which Kahneman and Tversky published five years after the anchoring paper — formalized this relationship, and it is no accident that the same two researchers produced both.

Anchoring vs Default Effects. Default effects are about the option that is pre-selected when the user does nothing. Anchors can be defaults, but not all defaults are anchors. A checkbox that says “Yes, subscribe me” defaulted to checked is a default effect. A strikethrough $199 next to a $79 price is an anchor. Nudge theory bundles both under “choice architecture,” but the levers are distinct: defaults work through inaction; anchors work through comparison.

Anchoring in the Real World

The anchoring literature is enormous, but four domains keep coming back because they show the effect operating on large populations, in real money, under competitive conditions. These are the places to look first if you want to learn from the best (and the worst) anchoring design.

Pricing and E-Commerce

Every pricing page on the internet is an anchoring exercise. The most common pattern — and it is common because it works — is the three-tier pricing table with the highest tier shown first or in the center with a “Most Popular” label. The highest tier sets the anchor; the middle tier then feels like the reasonable compromise; the low tier looks like a budget sacrifice. Drop the highest tier and the middle tier no longer feels like a compromise. It just feels expensive.

Restaurants are where this tactic is most visible. William Poundstone’s Priceless (2010) describes the pattern from the industry side: a box around an outrageously expensive item, a $125 plate of oysters at Balthazar, and then every steak further down the menu reading as reasonable by comparison. The expensive item is not there to be ordered. It is there to set what “expensive” means on that page.

One honest caveat, because this page previously got it wrong. Poundstone offers that as an industry observation rather than a measured result, and the published menu research by Sybil Yang and Sheryl Kimes tested something else: how prices are formatted, with 201 diners seeing the same lunch menu priced with a dollar sign, as a bare numeral, or spelled out in words. Yang’s eye-tracking work went on to undercut the menu “sweet spot” folklore rather than confirm it. The premium-anchor effect on a menu is plausible, widely practised, and not something anyone has published a clean field experiment on.

Amazon and every major e-commerce site now uses the strikethrough anchor — “List Price $49.99, Our Price $22.99” — so aggressively that reference pricing has drawn regulatory attention. In July 2017 the FTC made informal inquiries into whether Amazon’s discount claims mislead shoppers, after the advocacy group Consumer Watchdog sampled about a thousand products, found reference prices on roughly 46% of them, and reported that 61% of those sat above what Amazon itself had charged in the previous ninety days. No enforcement action followed, and Amazon called the study deeply flawed.

The detail worth carrying into your own design review is which format drew the complaint. Consumer Watchdog’s own report found that bare strikethrough prices generally did match prices Amazon had recently charged; the contested formats were “was” prices, “sale” prices, and the List Price field. A former-price claim is on safe ground under the FTC’s guides when the former price is one at which a substantial number of sales actually occurred, which is a cheap thing to check and an expensive thing to be wrong about. None of which stops the anchor from working: the mind assimilates the number without auditing whether any unit ever sold at $49.99, so deceptive pricing is a legal and trust problem rather than an effectiveness one.

Negotiation and Salary

The single most valuable application of anchoring for most working professionals is in salary negotiation. The meta-analytic evidence is overwhelming: the candidate who names a specific, defensible, somewhat-aggressive number first captures more of the final settlement than the candidate who lets the employer anchor. The difference averages out to thousands of dollars per negotiation and compounds across a career.

Specific numbers anchor better than round numbers. An ask of “$144,000” produces a higher final offer than an ask of “$145,000” or “$150,000,” because the specificity suggests a researched, informed position that the other side is reluctant to argue down as aggressively. Mason, Lee, Wiley, and Ames (2013) documented this “precision effect” in negotiation contexts with consistent results.

Healthcare and Medical Decision-Making

Physicians anchor. Surveys and experiments have shown that when a previous clinician records a tentative diagnosis in a patient’s chart, the next clinician’s independent assessment tracks that initial diagnosis far more than the underlying clinical data would predict. This is one of the drivers of diagnostic momentum, a chain of clinicians all anchored to the first provisional read.

The cost is not academic, and the defensible numbers are these. Chart-review work by Hardeep Singh and colleagues estimates that about one in twenty US adult outpatients experiences a diagnostic error each year, on the order of twelve million adults. Diagnostic error accounts for roughly a fifth of malpractice claims, with the majority of those arising in ambulatory care. In a series of a hundred diagnostic-error cases in internal medicine, 74% involved cognitive factors, of which anchoring is one among several.

No published figure isolates anchoring’s own share of patient harm, and this page used to carry one that does not exist.

In patient decision-making, anchoring also drives risk perception. A patient who reads that a procedure has a 95% success rate makes very different decisions than a patient who reads that the same procedure has a 5% failure rate, even though the two framings carry identical underlying anchors. When medical consent forms pick the 95% frame, they are not lying. They are selecting which anchor the patient uses.

Gamification, Leaderboards, and Progress Design

Every leaderboard is an anchor. The number at position 1 tells every other player what the ceiling is. If that ceiling is plausible and aspirational, it motivates. If it is so high that it feels unreachable, it crushes. The classic failure mode of poorly designed fitness apps is a leaderboard whose top five positions are held by obsessive users with 50,000 steps per day. For a typical user starting at 4,000 steps per day, that anchor does not motivate. It makes them feel like the game is not for them, and they quit.

Good anchoring design in gamification picks anchors the user can see themselves crossing. A progress bar that starts pre-filled at 30% (“you’re already 30% of the way there — just add a few things to complete your profile”) is a better anchor than one that starts empty, because it reframes the task as a smaller remaining distance. Endowed Progress, documented by Nunes and Drèze in their 2006 car-wash field experiment, is anchoring applied to progress bars. They handed out 300 loyalty cards: one needed eight stamps and started blank, the other needed ten and arrived with two already affixed. Both demanded the same eight washes. 34% of the pre-stamped cards were completed against 19% of the blank ones, and the pre-stamped customers also came back faster as they approached the reward.

The condition that gets dropped from every retelling is the one you need: the effect disappeared when no reason was given for the head start. A pre-filled progress bar that cannot explain itself is decoration, and users treat it as such.

The Elephant in the Room

The ethical conversation about anchoring is harder than the ethical conversation about almost any other behavioral-economics principle, because anchoring is genuinely unavoidable. You cannot not anchor. The moment you put any numerical information in front of a decision-maker, you have set an anchor, and the decision has been shifted accordingly. A pricing page without anchors is not a neutral pricing page. It is a pricing page where you have surrendered the anchor-setting to the user’s random background thoughts — in which case they will anchor on whatever the last product they saw cost, or whatever their spouse told them this kind of thing should cost, or whatever their Social Security digits happen to be.

The ethical question, then, is not “should I use anchoring?” You have no choice. The question is: which anchor, and in whose interest?

Three tests separate ethical anchoring from manipulation. First, the accuracy test: does the anchor give the user a more accurate mental model of the product’s value, or less accurate? A pricing page that anchors on the genuine enterprise price of a product is helping users calibrate what this category costs. A strikethrough price that was never real is inventing value that does not exist. Second, the symmetry test: would I be comfortable if my best customers knew exactly what anchor I was setting and why? If the anchor is so sensitive that the design only works while the user is unaware, it is probably exploiting rather than informing. Third, the long-term test: does this anchor help the user make a decision they will be glad they made a year from now, or only glad for the 30 seconds during which they clicked “buy”?

None of these tests are clean. They are judgment calls. But they are closer to the truth than the two extreme positions of “anchoring is always manipulation” and “anchoring is always fine as long as it’s legal.” The honest answer is that anchoring is a lever, like any lever — powerful, morally neutral in itself, and meaningful only in the context of what the designer is trying to do with it.

How to Apply Anchoring & Adjustment with the Octalysis Framework

The Octalysis Framework names the eight Core Drives that cause real human motivation. Anchoring is not a Core Drive — it is a perceptual primitive that decides how every Core Drive’s input is read. You can think of it as the lens that sits in front of the Octalysis octagon, setting the reference point that each Core Drive uses to decide what counts as meaningful, developmental, creative, social, owned, scarce, unpredictable, or lost.

The most direct interaction is with Core Drive 8: Loss & Avoidance. CD8 only fires against a reference point, and the anchor is the reference point. A “90% off today only” anchor makes the ordinary purchase price feel like a loss tomorrow. A “your score dropped from #7 to #12” anchor makes a previously acceptable ranking feel like a loss. Move the anchor, move the loss. The enterprise-tier price on a pricing page is doing this job whether or not anyone chose it deliberately, which is why it is the first thing worth testing when a middle tier underperforms.

Core Drive 6 (Scarcity & Impatience) also runs on anchors. “Only 3 left in stock” is an anchor set against an implicit supply anchor of “plenty.” “Sale ends in 14:47” is an anchor set against an implicit time anchor of “as long as you want.” The scarcer the anchor, the stronger CD6 fires. The anchor is not just a motivator — it is a meter by which the motivator is measured.

Core Drive 2 (Development & Accomplishment) depends on goal anchors. “10,000 steps per day” is an anchor. “Complete 3 courses this month” is an anchor. Change the numerical goal and you change the identity of the accomplishment. The round number (“10,000”) and the over-round number (“100,000 followers,” “1,000,000 views”) are cultural anchors that have trained consumers to treat those thresholds as accomplishments in themselves — which is why achieving 10,000 steps feels meaningful but achieving 9,500 does not. This is pure anchoring. There is no meaningful physiological difference between 9,500 and 10,000 steps. The meaning is in the anchor.

Core Drive 4 (Ownership & Possession) interacts with anchoring via the endowment effect and reference-point ownership. A user who reaches “Gold Tier” anchors to being a Gold Tier customer. Any loyalty-program redesign that moves the threshold up (“Gold Tier is now 10,000 points, not 8,000”) will feel like a loss to existing Gold Tier holders, even if their benefits are unchanged. The anchor is not just what they achieved. It is who they are in the system.

Two of the Game Techniques in the Octalysis library carry the word in their names, which is a fair signal of how central this is. Game Technique #69 (Anchored Juxtaposition) puts two options side by side where one costs money and the other costs labor, so each one prices the other. Game Technique #58 (Conformity Anchor, or Social Norming) shows the user how people like them are behaving, which anchors a quantity without naming a price at all.

Four more do the same work from different angles. Game Technique #68 (Magnetic Caps) caps how often a Desired Action can be taken, which sets the supply the user measures scarcity against. Game Technique #65 (Countdown Timer) anchors the full duration so the remainder reads as shrinking. Game Technique #1 (Status Points) gives standing a number, and Game Technique #3 (Leaderboard) publishes the ranking, so the top position anchors the ceiling for everyone below it.

Used well, anchors help users understand what is normal, what is scarce, and what is achievable. Used poorly, they manipulate users toward purchases and behaviors that serve the designer and not the user. The Octalysis framework is fundamentally White Hat — biased toward designs that users would endorse on reflection. Anchoring can be deployed in either direction, which is exactly why conscious anchor design is a core Octalysis skill and why unconscious anchor design usually leaves money on the table or, worse, burns trust for short-term numbers.

Practical Steps to Apply Anchoring & Adjustment

If you are designing a product, pricing page, leaderboard, onboarding flow, or negotiation and you want to use anchoring deliberately rather than by accident, here is the operational sequence I run with clients.

Step 1: Map every number a user sees in order. Open your product and write down, in sequence, every numerical value the user encounters in the first three minutes. Prices, scores, progress percentages, timers, follower counts, ratings, quantities. The first number on that list is the primary anchor. The second is the secondary anchor. Most product teams have never made this list. You cannot design anchors you have not inventoried.

Step 2: Ask what answer each anchor implies. For every number on the list, ask: what quantity is this anchor going to bias an estimate of? If the first price on the pricing page is $499, the user will anchor their “this kind of product costs about X” estimate in the vicinity of $499. If the first follower count they see on a social feed is 8.4M, their “how big is this person” estimate anchors there. Anchors do not act on “the user’s perception of the product in general.” They act on specific numerical quantities the user is estimating.

Step 3: Decide whether each anchor serves the user’s decision. A good anchor helps the user make a decision they would endorse on reflection. A bad anchor shifts the decision in the designer’s favor at the user’s expense. Go through each anchor and ask: does this make the user’s mental model more accurate, or less? If less, you have either a manipulation problem or a user-education problem.

Step 4: Choose the numerical form. Specific numbers anchor more powerfully than round numbers (Mason et al. 2013). “$144,000” anchors harder than “$150,000.” “94% customer retention” anchors harder than “95%.” Do not round down the anchors that matter unless the round number has independent cultural meaning (10,000 steps, 1M followers, etc.).

Step 5: Test the anchor for plausibility. If the anchor is implausibly high or low, the mind will reject it and suspicion will spill onto the rest of the experience. Ask a few unrelated users to look at the anchor in context and tell you, without prompting, whether it feels real. If more than one person says “that can’t be right,” your anchor is past its ceiling and needs to come down.

Step 6: Stress-test for professional buyers. If your users include repeat buyers, procurement professionals, or analysts who compare pricing across vendors, anchoring weakens considerably. For those users, the anchor should still be there, but it should be defensible — the enterprise-tier price should reflect a real enterprise feature set, not a fictional premium. Unethical anchoring costs the most in exactly the professional segments where the dollar amounts are highest.

Step 7: Monitor the anchor over time. Anchors age. An anchor that was aspirational two years ago may feel modest today, or suspicious today. Pricing anchors drift with inflation. Leaderboard anchors drift with user growth. Goal anchors drift with what the culture treats as a meaningful threshold. Build a quarterly review of your anchors into your design process. The number that worked in 2024 may be silently hurting you in 2026.

Closing Thoughts

Anchoring is the cheapest, fastest, and most underused design lever in behavioral science. Every pricing page, negotiation, leaderboard, and progress bar on the planet is already anchoring — whether by accident or on purpose. The teams that pull ahead are the teams that take the anchor away from accident and put it under deliberate, ethical control.

The Octalysis frame is useful here because it reminds you that anchors do not work in isolation. An anchor sets the reference point, and then the Core Drives — Loss Aversion, Scarcity, Development, Ownership, Epic Meaning — do the motivational work against that reference point. A great anchor with no Core Drive behind it is a wasted opportunity. A Core Drive design with no deliberate anchor is half-built.

If you only remember one sentence from this post: the first number your user sees is doing more work than anything else on the screen, and you have probably never audited it. So audit it, change it on purpose, and measure what happens. You will be surprised how often the cheapest real improvement available to a product is the first number on the page.

If you want the full operator’s manual, the Octalysis Framework is the larger system this post sits inside, and my book Actionable Gamification walks through how to design the Core Drives around anchors like this for real products. If your team is sitting on a pricing page, loyalty program, leaderboard, or negotiation flow where the anchors feel accidental rather than chosen, that is exactly the problem my advisory work is built around. Reach out through the contact page — I take a small number of engagements each quarter, and a first-number audit is usually the cleanest way to start.

Frequently Asked Questions

What is the anchoring and adjustment heuristic in simple terms?

Anchoring and adjustment is the tendency to rely on the first number that enters a decision and then “adjust” from there — usually not enough. If you see a $200 price first and then a $79 price, the $79 feels cheap because the $200 became your reference point, even if the $200 was arbitrary.

Who discovered anchoring and adjustment?

Amos Tversky and Daniel Kahneman introduced the heuristic in their 1974 paper “Judgment Under Uncertainty: Heuristics and Biases,” published in Science. It was one of three core heuristics (along with availability and representativeness) they proposed to explain how people make estimates under uncertainty. Kahneman won the 2002 Nobel Memorial Prize in Economic Sciences partly for this body of work.

Does anchoring work on experts?

Yes. Northcraft and Neale (1987) found real-estate agents’ appraisals tracking arbitrary listing prices while 81% of those agents did not list the listing price among the factors they had used. Englich, Mussweiler, and Strack (2006) found experienced German criminal judges recommending sentences that tracked dice they had rolled themselves, five months against eight.

Whether expertise helps at the margin is still unsettled. A preregistered 2024 study found experts anchoring less than novices, while a 2019 meta-analytic thesis found domain knowledge making it worse. What no study has found is a group that is immune.

What is the difference between anchoring and framing?

Framing is about the wording of a choice (“90% survival rate” vs “10% mortality rate”). Anchoring is about the number near the choice. The two often stack in real design, but the mechanisms are distinct. Framing changes which concept the user activates; anchoring changes which numerical answer the user lands on.

How strong is the anchoring effect?

Robust in the classic paradigm, and genuinely contested in magnitude. The standard version, where the experimenter supplies a relevant anchor, is among the most reliably replicated findings in the field: it came through the Many Labs replication project in 2014 as one of the strongest effects tested, across dozens of samples.

On size, there are two live positions. The largest synthesis to date, Schley and Weingarten’s working meta-analysis of more than 2,600 effect sizes, puts it large at a Hedges’ g of about 0.83 and reports it surviving publication-bias correction; Röseler and Schütz’s 2022 meta-analysis reaches similar conclusions about robustness while finding several popular moderators, incentives among them, to be null.

Pushing the other way, a 2025 replication in Economic Inquiry argues this literature routinely runs at under 30% statistical power, and its own higher-powered replication produced estimates roughly a ninth the size of the originals.

The practical reading for a designer is narrower than either camp. Explicit, relevant, plausible anchors move behavior reliably. Incidental and subliminal anchors, the ones that make for better dinner-party stories, sit close to zero once properly powered, so do not build a product on them.

Can you resist anchoring if you know about it?

Only partially. Knowing about anchoring and being warned in advance weakens the effect by some amount, but no study has ever shown that awareness eliminates it. This is the main reason anchoring is treated as a design problem rather than an education problem: you cannot train your users out of being influenced, so the ethical burden falls on the designer.

How does anchoring relate to Prospect Theory?

Tightly. Prospect Theory’s core claim is that value is measured relative to a reference point, with losses weighted more heavily than gains. Anchoring is the mechanism that sets the reference point. Move the anchor, and you move what the mind treats as the neutral baseline — which means you move what counts as a gain and what counts as a loss. Anchoring is the scaffold Prospect Theory loads onto.

What is a real-world example of anchoring in pricing?

The classic example is a three-tier pricing page where the highest tier is shown first and anchors the user’s price expectations for the category. When the expensive tier is removed, the middle tier feels more expensive, even though its price has not changed. Restaurant menus use the same move with a deliberately expensive item near the top, which sets what “expensive” means before the reader reaches anything they might order. That one is standard industry practice rather than a published field experiment.

Is anchoring ethical?

Anchoring itself is morally neutral — you cannot avoid setting an anchor whenever you show a user a number. The ethical question is which anchor you choose and in whose interest. Anchors that help users understand the product category accurately are legitimate design. Anchors that invent fictional reference points to make bad deals feel like bargains are manipulation and generally destroy trust in the long run.

How do I use anchoring in gamification design?

Anchor the user’s sense of what is possible. Leaderboards anchor the ceiling of achievement. Pre-filled progress bars anchor the remaining distance. Round-number goals anchor what counts as accomplishment. Status tiers anchor the cost of losing standing. In Octalysis terms, every Core Drive has an anchor it measures against, and conscious anchor design is one of the highest-leverage moves in the framework.

References

1. Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124–1131.
2. Ariely, D., Loewenstein, G., & Prelec, D. (2003). “Coherent arbitrariness”: Stable demand curves without stable preferences. Quarterly Journal of Economics, 118(1), 73–106.
3. Northcraft, G. B., & Neale, M. A. (1987). Experts, amateurs, and real estate: An anchoring-and-adjustment perspective on property pricing decisions. Organizational Behavior and Human Decision Processes, 39(1), 84–97.
4. Englich, B., Mussweiler, T., & Strack, F. (2006). Playing dice with criminal sentences: The influence of irrelevant anchors on experts’ judicial decision making. Personality and Social Psychology Bulletin, 32(2), 188–200.
5. Galinsky, A. D., & Mussweiler, T. (2001). First offers as anchors: The role of perspective-taking and negotiator focus. Journal of Personality and Social Psychology, 81(4), 657–669.
6. Mussweiler, T., & Strack, F. (1999). Hypothesis-consistent testing and semantic priming in the anchoring paradigm: A selective accessibility model. Journal of Experimental Social Psychology, 35(2), 136–164.
7. Epley, N., & Gilovich, T. (2005). When effortful thinking influences judgmental anchoring: Differential effects of forewarning and incentives on self-generated and externally provided anchors. Journal of Behavioral Decision Making, 18, 199–212.
8. Chapman, G. B., & Johnson, E. J. (1999). Anchoring, activation, and the construction of values. Organizational Behavior and Human Decision Processes, 79(2), 115–153.
9. Strack, F., & Mussweiler, T. (1997). Explaining the enigmatic anchoring effect: Mechanisms of selective accessibility. Journal of Personality and Social Psychology, 73(3), 437–446.
10. Simmons, J. P., LeBoeuf, R. A., & Nelson, L. D. (2010). The effect of accuracy motivation on anchoring and adjustment. Journal of Personality and Social Psychology, 99(6), 917–932.
11. Mason, M. F., Lee, A. J., Wiley, E. A., & Ames, D. R. (2013). Precise offers are potent anchors. Journal of Experimental Social Psychology, 49(4), 759–763.
12. Nunes, J. C., & Drèze, X. (2006). The endowed progress effect: How artificial advancement increases effort. Journal of Consumer Research, 32(4), 504–512.
13. Furnham, A., & Boo, H. C. (2011). A literature review of the anchoring effect. Journal of Socio-Economics, 40(1), 35–42.
14. Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
15. Poundstone, W. (2010). Priceless: The Myth of Fair Value (and How to Take Advantage of It). Hill and Wang.
16. Mussweiler, T., & Strack, F. (2001). Considering the impossible: Explaining the effects of implausible anchors. Social Cognition, 19(2).
17. Galinsky, A. D., Ku, G., & Mussweiler, T. (2009). To start low or to start high? The case of auctions versus negotiations. Current Directions in Psychological Science, 18(6), 357–361.
18. Tamir, D. I., & Mitchell, J. P. (2010). Neural correlates of anchoring-and-adjustment during mentalizing. Proceedings of the National Academy of Sciences, 107(24), 10827–10832.
19. Qu, C., Zhou, L., & Luo, Y. (2008). Electrophysiological correlates of adjustment process in anchoring effects. Neuroscience Letters, 445(3). (EEG, not fMRI — see the brain section.)
20. Röseler, L., & Schütz, A. (2022). Hanging the anchor off a new ship: A meta-analysis of anchoring effects. PsyArXiv preprint.
21. Schley, D., & Weingarten, E. 50 years of anchoring effects: A theoretical re-integration and meta-analysis (2,601 effect sizes; working paper, not yet peer reviewed).
22. Li et al. (2025). Underpowered studies and exaggerated effects: A replication and re-evaluation of the magnitude of anchoring effects. Economic Inquiry.
23. Klein, R. A., et al. (2014). Investigating variation in replicability: A “many labs” replication project. Social Psychology, 45(3), 142–152.
24. Mayer, M., & Rebholz, T. R. (2024). Navigating anchor relevance skillfully: Expertise reduces susceptibility to anchoring effects. PsyArXiv preprint, not yet peer reviewed.
25. Lee, A. J., Loschelder, D. D., Schweinsberg, M., Mason, M. F., & Galinsky, A. D. (2018). Too precise to pursue: How precise first offers create barriers-to-entry in negotiations and markets. Organizational Behavior and Human Decision Processes, 148.
26. Yang, S. S., & Kimes, S. E. (2009). Menu price presentation influences on consumer purchase behavior in restaurants. International Journal of Hospitality Management, 28(1), 157–160.

Related Reading

• Prospect Theory: An S-Tier Behavioral Designer’s Guide to Loss Aversion
• The Framing Effect: Gain vs Loss Framing
• Dual Process Theory: System 1 vs System 2
• The Octalysis Framework
• Core Drive 8: Loss & Avoidance
• The Goal-Gradient Hypothesis: Why We Speed Up at the End
• Core Drive 6: Scarcity & Impatience

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?