What You Built Is Worth Less Than You Think

Almost every reader of this publication owns something they made. A firm, a product, a process, a client list. This article is about a well-measured tendency to overvalue exactly that, and about what the overvaluation costs at the one moment it becomes a number.

The Verdict, Stated First

Five claims, in descending order of confidence.

One. This is one of the best-evidenced findings in the series. A 2012 paper with four studies, a successful conceptual replication in 2017, and a 2026 meta-analysis across 55 studies and 5,454 participants reporting the effect robust to moderators.

Two. The magnitude is above this publication's own median. At 0.57 it sits well clear of the 0.41 median the fiftieth article computed across this series' first forty-nine pieces.

Three. The boundary condition is the most useful part and is almost never quoted. The effect dissipated when participants failed to complete the task or destroyed what they built, so the inflation attaches only to finished work.

Four. The commercial consequence is a deal that does not happen. On our own arithmetic, a 0.57 inflation in the seller's valuation raises the probability of no agreement from 50 to 65.7 percent.

Five. There is no partial fix. On the same arithmetic, restoring an even chance requires giving up the entire effect, because conceding two thirds of it still leaves the odds meaningfully against a deal.

Our Grades For These Claims

Applying the scheme from the first article in this series.

Grade A for the 2012 findings, from an abstract obtained verbatim from the publisher and corroborated against a copy of the paper hosted by the third author's own university.

Grade A for the meta-analytic figure, from an abstract obtained verbatim from the publisher, though truncated before it finishes listing secondary outcomes.

Grade A for the conceptual replication, from an abstract obtained verbatim from two independent sources.

Grade B for the mechanism, which the replication attributes to psychological ownership and which we did not examine in any paper.

Grade A for our own arithmetic, on assumptions we state, one of which is invented.

Our position: the finding is solid, current, and quantified, which is rare in this series, and the commercial translation is entirely ours.

A Note On Method

Everything here is verified to August 2026.

We obtained the 2012 paper's abstract verbatim from the publisher[3], plus body text and its reference list from a copy hosted by the third author's own university department[1], and further text from a copy on the first author's institutional repository[4]. We did not obtain the paper's data or results tables.

We obtained the 2026 meta-analysis's abstract from the publisher, in two renderings which agree[2]. It is truncated mid-sentence in both, and we did not obtain the paper.

We obtained the 2017 replication's abstract verbatim from the publisher and from a repository copy[5]. We did not obtain the paper.

We did not obtain the 2012 companion paper, the 1959 precursor, the finance study, or any other work cited here, and report all from citation records and titles.

All arithmetic is ours, assumes normality and equal spreads, and uses one entirely invented parameter which we flag at each use.

This article discusses research on valuation. It is not valuation, transaction or pricing advice, and nothing here should inform the price of any actual business.

The 2012 Paper

The source.

Norton, M. I., Mochon, D., and Ariely, D. (2012), The IKEA effect: When labor leads to love, Journal of Consumer Psychology, 22(3), 453–460, July, DOI 10.1016/j.jcps.2011.08.002[3].

The paper explains its own name: "we suggest that labor alone can be sufficient to induce greater liking for the fruits of one's labor: even constructing a standardized bureau, an arduous, solitary task, can lead people to overvalue their (often poorly constructed) creations. We call this phenomenon the 'IKEA effect,' in honor of the Swedish manufacturer whose products typically arrive with some" assembly required[4].

Three observations, ours.

The phrase "labor alone can be sufficient" is the claim being staked. Not that customisation raises value, which would be unremarkable, but that effort on a standardised object does it.

The parenthetical "often poorly constructed" is the authors' own, and it forecloses the obvious defence before anyone raises it. The creations were not better; they were theirs.

And the provenance is unusually good. We obtained the paper's text from a copy hosted by the third author's own university department, which is better than most sources in this series.

Four Studies

What was done, from the abstract.

"In four studies in which consumers assembled IKEA boxes, folded origami, and built sets of Legos, we demonstrate and investigate boundary conditions for the IKEA effect, the increase in valuation of self-made products."[3]

Four observations, ours.

The three tasks are deliberately unlike each other. Flat-pack furniture, origami and Lego differ in skill, duration and physical form, and a result holding across all three is harder to attribute to any one of them.

All three are assembly rather than creation. Nobody designed anything, which is what makes the finding about labour rather than about authorship.

The abstract says the paper investigates boundary conditions, and boundary conditions are the part this series usually finds missing from popular retellings. Here they are in the abstract's second sentence.

And the abstract closes with a scope claim: "we show that labor increases valuation for both 'do-it-yourselfers' and novices."[3] So a taste for building things is not required.

As Valuable As An Expert's

The result, and the comparison chosen makes it sharper than a bare valuation gap would.

"Participants saw their amateurish creations as similar in value to experts' creations."[3]

Four observations, ours.

The benchmark is not an identical object made by someone else. It is an object made by someone better, and participants rated their own as comparable.

The word "amateurish" is the authors'. They are not describing a small gap in objective quality; they are describing a large one that the maker did not price.

That structure is what makes the finding commercially serious. An owner comparing their firm to a better-run competitor and seeing rough parity is the business version of the same judgment.

And it explains why the effect survives outside advice. The comparison that should correct you is the one the effect distorts, so looking at a better operator does not supply the correction it should.

And They Expected You To Agree

The clause that connects this to the last two articles.

"...and expected others to share their opinions."[3]

Four observations, ours.

This is projection in the sense the seventieth article examined, applied to a valuation rather than a preference.

Its practical form is a negotiation failure rather than a private error. Believing your price is high is survivable; believing others will see it as fair is what produces a stalled sale.

The two halves compound. An inflated valuation you know is personal can be discounted before you quote it, and one you believe is shared cannot.

And that pairing is why we treat the deal-failure arithmetic below as the article's main output rather than the valuation gap itself.

The Boundary Condition

The most useful sentence in the paper and the one nobody quotes.

"We show that labor leads to love only when labor results in successful completion of tasks; when participants built and then destroyed their creations, or failed to complete them, the IKEA effect dissipated."[3]

Four observations, ours.

The word "only" is doing serious work. This is not a moderator that weakens the effect; the abstract says it dissipated.

Two separate conditions produced that: failing to finish, and finishing then destroying. The second is the more interesting, because the labour was expended in full and the valuation still did not attach.

So the effect is about the artefact rather than the effort. Labour is necessary and something must survive it, which is a narrower claim than the popular version.

And it produces a prediction about a founder that we develop below, and which as far as we can tell nobody has tested.

A Conceptual Replication

The check, five years later.

Sarstedt, M., Neubert, D., and Barth, K. (2017), The IKEA Effect. A Conceptual Replication, Journal of Marketing Behavior, 2(4), 307–312, DOI 10.1561/107.00000039[5].

Its abstract: "We replicate and extend Norton et al.'s (2012) and Mochon et al.'s (2012) studies on the IKEA effect, according to which consumers show a higher willingness-to-pay when they assemble products themselves. Our results support the robustness of the original effect and indicate that psychological ownership acts as a psychological mechanism that underlies the IKEA effect."[5]

Four observations, ours.

A successful conceptual replication is not the outcome this series usually reports, and we note it plainly.

It is conceptual rather than direct, meaning the design differed. A repository copy records the materials as loom bands rather than furniture or origami[5], so this is the effect surviving a change of stimulus.

The mechanism proposed is psychological ownership, which is a different account from effort justification and connects the finding to the endowment literature this publication covered separately.

And the same copy records that the authors also examined whether the effect dissipates when individuals disassemble their creations[5], which is the 2012 boundary condition being re-tested. We did not obtain what they found.

The 2026 Meta-Analysis

The current quantitative synthesis, and it is unusually recent for this series.

Pelled, A., Demetriades, S. Z., and Walter, N. (2026), Labor Leads to Love, Right? A Meta-Analysis of the IKEA Effect, Psychology and Marketing, DOI 10.1002/mar.70064[2].

Its abstract: "The IKEA effect refers to people's tendency to overestimate the value of items they personally assembled or assisted in creating. The current meta-analysis (k = 55, N = 5,454), synthesizes data from empirical studies assessing the relationship between self-assembly and product valuation to estimate the average effect size of the phenomenon, test its main and secondary outcomes, and assess potential boundary conditions."[2]

Three observations, ours.

Fifty-five studies and 5,454 participants is a substantial base, and the authors are from communication rather than marketing departments, which means the synthesis was not performed by the finding's own field.

The stated aims include boundary conditions, which matters because the 2012 paper's boundary is the part we think most usable.

And the publication date makes this the most current source in this series, which is worth noting given how often these articles report literatures settled decades ago.

Point Five Seven

The number.

"The results point to a significant moderate impact of self-assembly labor on valuation (d = 0.57, p < 0.005), as well as a host of secondary outcomes, including liking, self-concept, and sense of", at which point the abstract is truncated in both renderings we obtained[2].

And on robustness: "the observed effects remained quite robust in the face of various moderators such as product customization and tangibility."[2]

Four observations, ours.

0.57 sits well above the median of 0.41 that the fiftieth article computed across this publication's first forty-nine pieces. This is a larger effect than most of what this series has covered.

The authors call it "moderate", which is the correct word and is worth keeping. It is not a dominant effect and it is not a rounding error.

Robustness to customization confirms the 2012 framing. If the effect held whether or not the object was personalised, then labour rather than authorship is doing the work.

And robustness to tangibility is the clause that licenses everything commercial below. If the effect does not depend on the thing being physical, it can apply to a business, a process or a client relationship, which a furniture experiment on its own would not support.

We would put weight on that clause carefully, because it is doing more work in this article than any other single word. Every experiment we have described involves an object a person can hold, and the leap to a business is large.

What the moderator test establishes is that tangibility did not change the size of the effect across the studies pooled, which is weaker than establishing the effect in an intangible setting. The strongest form of that evidence would be a study on something with no physical form at all, and the finance title named below suggests one exists.

The Question Mark In The Title

A small note on our own expectations, since we got them wrong. Ours.

Three observations.

The meta-analysis is titled "Labor Leads to Love, Right?", and we went into it expecting a debunking. A question mark appended to a famous finding usually signals one.

It is not. The meta-analysis confirms the effect at a moderate magnitude and reports it robust to the moderators tested, and the punctuation is rhetorical.

We record that because this series has spent seventy-one articles warning against reading confidence into presentation, and the same caution applies to reading doubt into it. A question mark is not a finding either.

What It Does To A Sale

The commercial translation, which is the point of this article. Our own arithmetic, assuming normality and equal spreads.

Suppose a seller values a business at the true figure plus 0.57 standard deviations, and a buyer values it at the true figure, both with the same spread of uncertainty. No deal happens when the seller wants more than the buyer will pay.

With no effect, the probability of no deal is 50.0 percent, since either party is equally likely to value it higher.

With d equals 0.57, the probability of no deal is 65.7 percent.

Four observations.

That is an increase of 15.7 percentage points, turning an even chance into roughly two in three against.

The effect operates entirely through the seller in this construction, which is the situation of a founder selling to a buyer who did not build the thing.

And the framing matters. The cost is not a lower price; it is a transaction that does not occur, which does not appear in any ledger and which the seller will attribute to the buyer being unrealistic.

Note what this does not model. Real sales involve negotiation, multiple bidders and information exchange, all of which should reduce the failure rate, so this is an upper bound on a single-shot encounter rather than a forecast.

Two further things the construction assumes, both of which we would want a reader to weigh. It puts the whole effect on the seller, which is right for a founder selling to a financial buyer and wrong for a sale to a competitor who built something similar themselves.

And it assumes the same spread of uncertainty on both sides, which is generous to the seller. A buyer with better comparables and a narrower spread would decline more often than this figure suggests, not less.

So the honest reading of 65.7 percent is directional. The number is ours, the assumptions are ours, and what the reported effect size supports is that a moderate seller-side inflation moves the odds of agreement materially, from roughly even to clearly against.

Translating Into Dollars

What 0.57 standard deviations is worth, which depends on a quantity we had to invent. Our own arithmetic, and the spread figure is entirely ours.

An effect measured in standard deviations means nothing in money until you know how much independent valuations of the same business actually vary. On a two million dollar business:

If valuations spread by 10 percent: the implied inflation is 5.7 percent, or $114,000.

At 20 percent: 11.4 percent, or $228,000.

At 30 percent: 17.1 percent, or $342,000.

At 40 percent: 22.8 percent, or $456,000.

At 50 percent: 28.5 percent, or $570,000.

Four observations.

No source we obtained states how much private company valuations vary, and the entire table therefore rests on a number we chose. We flag that rather than presenting a single figure as though it were derived.

What the table does establish is the sensitivity. The dollar consequence of the same measured effect ranges over a factor of five depending on a quantity the literature does not supply.

Which means a reader should take the shape and not the numbers. A moderate effect on a business with a wide valuation range is worth a great deal of money, and businesses with narrow valuation ranges are rare.

And there is a check available to anyone who wants one. Obtain two independent valuations and see how far apart they are, which measures the spread directly for your own firm rather than borrowing ours.

There Is No Partial Correction

An unwelcome result from the same arithmetic. Ours.

How much of the effect must a seller give up to restore the original odds of agreement?

To return to a 50 percent chance of no deal: the entire effect, all 0.570.

To reach 55 percent: a reduction of 0.392, which is 69 percent of the effect.

To reach 60 percent: a reduction of 0.212, which is 37 percent.

Three observations.

Conceding a third of the effect moves the failure probability from 65.7 to only 60 percent, which is very little for a large concession.

So the usual advice to allow for some sentiment in your asking price is arithmetically weak. A partial adjustment buys a small change in the odds and costs the whole adjustment.

And this is the opposite shape from the sixty-eighth article's finding on bargaining, where the payoff surface was flat and concession was cheap. Here the surface is steep and partial measures do not pay, which is why we would treat the two situations differently.

The Completion Boundary, Applied

A testable prediction that follows from the 2012 boundary condition. Ours, untested, and we have found nobody who has examined it.

Four observations.

The effect dissipated when the task was not completed or the creation destroyed. So the inflation attaches to finished work only.

Applied to a founder, that predicts an unevenness. The parts of the business that were completed carry the premium, and abandoned projects, discontinued lines and half-built systems do not.

If that holds, it has a practical use in the opposite direction from the rest of this article. A founder's valuation of what they abandoned should be closer to honest than their valuation of what they finished, which makes the abandoned material a calibration point.

And the same boundary suggests why an ongoing business is the worst case. A firm that is still trading has not been completed and has not been destroyed, so which side of the boundary it falls on is genuinely unclear, and nothing we obtained addresses it.

The Precursor

Where the idea comes from, from the 2012 paper's own reference list.

It cites Aronson, E., and Mills, J. (1959), The effects of severity of initiation on liking for a group, Journal of Abnormal and Social Psychology, 59, 177–181, and Axsom, D., and Cooper, J. (1985), Cognitive dissonance and psychotherapy: The role of effort justification in inducing weight loss, Journal of Experimental Social Psychology[1].

We obtained neither and report the citations.

Three observations, ours.

The tradition is effort justification, which says people revalue things upward to justify what they spent getting them, and it predates the IKEA work by more than fifty years.

That account differs from the psychological ownership account the 2017 replication proposes. Justification is backward-looking and ownership is not, and they predict differently for something you acquired effortlessly and then owned for years.

And we did not obtain either paper, so we can name the tradition and not evaluate it.

In Finance

The most directly relevant extension we found, named only.

The meta-analysis's reference list identifies Brunner, F., Gamm, F., and Mill, W., MyPortfolio: The IKEA Effect in Financial Investment Decisions, Journal of Banking and Finance, 154, article 106529[2].

We did not obtain it and report the title and citation only.

Three observations, ours.

The title asserts the effect appears in investment decisions, which would mean a portfolio you assembled yourself is valued differently from an identical one you did not.

If so, the implication for an owner is direct and unpleasant. The concentrated position in your own company is the most self-assembled asset you hold, which this publication's home bias article approached from a different direction.

And a title is not a finding. We report it so a reader knows where to look and make no claim about what it concludes.

The Reverse Effect, And What It Means Now

A 2026 finding running the other way, which we report because it is timely and because we could obtain it.

A study on work assisted by generative artificial intelligence reports: "Across three experimental studies (N = 932) involving management students and full-time employees, we find that although AI-assisted participants achieved significantly higher objective performance than unassisted participants, they systematically underestimated their performance, rating it lower than their objectively achieved outcomes."[6]

It draws on "the effort heuristic, which links perceived effort to perceived value", and notes that AI "allows users to produce high-quality results with minimal effort."[6]

Four observations, ours.

This is the IKEA effect inverted. Less labour, less felt value, and here the output was measurably better.

The two findings together give a single rule. Perceived value tracks effort rather than quality, which mis-prices in both directions and by the same mechanism.

The commercial consequence in 2026 is specific and we have not seen it stated. Work that took you less time will be under-priced by you and valued normally by the client, which is the reverse of the founder's problem and equally expensive.

And we obtained this study's abstract from a repository page and not the paper, and it is one study rather than a literature.

One Mechanism, Two Directions

What unites the founding finding and its reverse, and it is older than either.

A reference page identifies Kruger, J., Wirtz, D., Van Boven, L., and Altermatt, T. W. (2004), The effort heuristic, Journal of Experimental Social Psychology, 40(1), 91–98[7], and the AI study we obtained describes it as "the effort heuristic, which links perceived effort to perceived value."[6]

We did not obtain the 2004 paper and report the citation and that one-line characterisation.

Four observations, ours.

If perceived value tracks effort rather than quality, both findings follow from one rule. Assembling furniture raises felt value because it took work. AI-assisted output lowers it because it did not.

That unification is worth more than either finding alone, because a single rule predicts both directions rather than requiring two separate biases.

It also predicts the error is symmetric in mechanism and asymmetric in salience. The overvaluation of what you built is famous; the undervaluation of what came easily has no name in general circulation.

And it identifies where the rule fails, which is the useful part. Effort and quality correlate in most settings, so the heuristic usually works, and it breaks precisely where a tool or a talent decouples them.

What Actually Survives

Our reading, stated directly.

Five statements.

The effect is real, moderate and current. Four studies in 2012, a successful conceptual replication in 2017, and a 2026 meta-analysis across 55 studies putting it at 0.57.

It does not require customisation or physicality, on the meta-analysis's own moderator tests, which is what licenses applying it to a business.

It attaches to finished work only. The founding paper reports it dissipating when the task was not completed or the creation destroyed.

The mechanism is contested between effort justification and psychological ownership, and we obtained neither account in full.

And the commercial consequence is a deal that does not happen, which on our own arithmetic rises from 50 to 65.7 percent and admits no partial fix.

Who This Hits Hardest

A consequence of the completion boundary combined with the effect size, which identifies the worst-affected group. Ours.

Four observations.

The premium attaches to finished work, and to work the person did themselves. That combination describes a specific kind of owner: one who built rather than bought, and who completed rather than delegated.

Which is close to a description of a successful founder-operator. The people most likely to have a business worth selling are the people most likely to overvalue it, and that correlation is not a coincidence but a consequence of the same effort.

It follows that the effect should be weakest in the cases where advisers see it least: inherited businesses, acquired businesses, and firms where the current owner arrived after the building was done.

And we would flag that as our own inference, untested and easy to test. An adviser with a book of transactions could check whether founder-sellers hold out longer than inheritor-sellers at comparable multiples, which would measure the effect directly on the population that matters.

Valuing Your Business

The first application. Ours, untested, and not valuation or transaction advice.

Four points.

The finding predicts that your own figure is high, and the founding paper's second clause predicts you will expect others to agree. The second is what stalls the sale.

The practical instrument is external and cheap. Obtain two independent valuations and record your own before you see them, which measures both your gap and your firm's valuation spread at the same time.

Recording your figure first is the part people skip and it is the part that works. An estimate written before you see the answer is behavioural evidence about yourself, which is the instrument the previous article argued is the only one available.

And the boundary condition gives a second calibration point at no cost. Value something you abandoned, and compare how confidently you price it against how confidently you price what you finished.

Two cautions on that instrument, since we are recommending it. An abandoned project is abandoned for a reason, and the reason is usually that it was not working, so its low valuation is partly deserved rather than purely uncontaminated.

What the comparison measures is therefore confidence rather than level. The useful signal is not that you price the dropped project lower, which you should, but whether you can state a range for it while giving a single number for the business you finished.

Pricing Your Work

The second application, and it runs in both directions. Ours.

Four points.

The effect predicts that work you laboured over will feel more valuable than work you did quickly, independent of the result.

Which means effort-based pricing is not merely a billing convention. It is a mechanism for turning a perceptual bias into an invoice, and it will systematically overprice slow work and underprice fast work.

The 2026 AI finding sharpens this considerably. If a task that took a day now takes an hour, the effort heuristic says you will value it at an hour's worth, and the client's valuation has not changed at all.

And that is an argument for pricing on outcome rather than on input which does not depend on any theory of value. It depends only on the observation that your own sense of what work is worth tracks how hard it felt.

What You Built Against What You Bought

The third application, and the one with an immediate test. Ours.

Four points.

Most firms contain both. Systems you built and systems you bought, processes you designed and processes you inherited, and the finding predicts you value the first category higher for reasons unrelated to their merits.

That produces a specific and checkable prediction about replacement decisions. You will be more reluctant to replace what you built than what you bought, holding performance constant.

The test costs nothing. List the systems you have replaced in five years and mark which were built in-house, and if the built ones survived longer at equal performance, you have measured the effect on yourself.

And this compounds with the fifty-fifth article's finding on status quo. Two separate reasons to keep what you have, one of which applies only to what you made, which predicts in-house systems outlive bought ones by more than either effect alone would explain.

The Other Side Of The Same Finding

Where the effect works for you rather than against you. Ours.

Four points.

Everything above treats the effect as a cost to the maker. The same finding says a customer who assembles, configures or contributes will value the result more.

The 2012 abstract supports the direction: labour increased valuation for "both do-it-yourselfers and novices"[3], so a customer need not enjoy the work for it to raise their valuation.

But the boundary condition constrains the design severely. The effect dissipated when participants failed to complete the task, so a configuration step a customer abandons partway is worse than no step at all.

And we would state the professional caution plainly. Deliberately inserting effort to raise a customer's valuation is a technique for making someone pay more for the same thing, and whether that is service design or manipulation depends on whether the effort produces anything they wanted.

One test separates the two cases and we would apply it before designing anything. Would you tell the customer why the step is there? Configuration that genuinely tailors a product survives being explained; a step inserted solely to create attachment does not.

And the 2026 finding on effortless work supplies the commercial argument against manipulation, for anyone unmoved by the ethical one. If perceived value tracks effort, then a competitor who removes the effort will be undervalued by their own customers too, and the firm that wins on convenience will have to price against a heuristic rather than with it.

When The Premium Is Real

The section this article needs. Ours.

Four observations.

Something you built can genuinely be worth more to you than to a buyer, and not through any error. It fits your operation, you can run it without documentation, and replacing it has switching costs a buyer would not face.

That is a real difference in value and the literature does not deny it. The experiments compare valuations of identical standardised objects, which is precisely the design that removes genuine fit from the picture.

So the honest formulation is a question rather than a correction. Is your figure higher because the thing suits you, or because you made it? The first is a legitimate reason to decline a sale and the second is not.

And there is a test that separates them. Ask what you would pay for an identical system built by someone else, since genuine fit survives that question and the IKEA effect does not.

What To Do

Write your own valuation down before you see anyone else's. That converts an unmeasurable gap into a recorded one, and it is the only instrument the previous article left standing.

Get two independent valuations and note the spread between them. That measures your firm's valuation variance directly, which is the quantity this article had to invent.

Expect no deal rather than a lower price. On our own arithmetic the failure probability rises from 50 to 65.7 percent, and the cost appears as a transaction that does not happen.

Do not bother with a partial adjustment. Conceding a third of the effect moves the odds by less than six points, and restoring an even chance requires giving up all of it.

Ask what you would pay for the same thing built by someone else. Genuine fit survives that question and the effect does not.

Use what you abandoned as a calibration point. The founding paper reports the effect dissipating on uncompleted work, so your valuation of a dropped project should be closer to honest.

Watch the reverse effect on fast work. A 2026 study reports people underestimating AI-assisted performance that was objectively better, so work that took less time will be underpriced by you and not by the client.

Notice which systems you built when you decide what to replace. The prediction is that in-house systems outlive bought ones at equal performance, and your own replacement history tests it.

The Limits Of This Analysis

Several caveats matter. This article discusses research on valuation and is not valuation, transaction or pricing advice; nothing here should inform the price of any actual business, and the applications are our own reasoning and untested. Everything is verified to August 2026. We did not obtain any of the papers in full. The 2012 paper is reported from its published abstract plus body text and reference list from copies hosted by two of its authors' institutions, and we obtained none of its data or results tables, so we report no willingness-to-pay figures from the original studies. The 2026 meta-analysis is reported from an abstract truncated mid-sentence in both renderings we found, so its secondary outcomes are listed only partially and we have none of its moderator analyses, publication-bias tests or confidence intervals. The 2017 replication is reported from its abstract, and we did not obtain what it found on the disassembly condition it examined. We did not obtain the 2012 companion paper, the 1959 or 1985 precursors, or the finance study, and report all from citation records and titles; a title is not a finding. All arithmetic is ours. The deal-failure calculation assumes normality, equal spreads for both parties, and a single-shot encounter with no negotiation, no competing bidders and no information exchange, all of which would reduce the failure rate, so 65.7 percent is an upper bound rather than a forecast. The dollar table rests on a valuation spread we invented, because no source we obtained states how much independent valuations of a private company vary; the resulting figures range over a factor of five and should be read for shape only. The prediction about abandoned projects, the prediction about in-house systems outliving bought ones, and the reading of the AI finding as a reverse effect are all our own inferences and untested. And the experiments compare identical standardised objects, which by design removes the genuine fit that makes something you built legitimately more valuable to you.

Related Reading

No. 73 The One-Third Threshold Trust pays if the other party returns more than a third. But a firm on a thirty percent margin needs a seventy percent payment probability, and that is... Behavioural Finance The Question This Series Has Been Avoiding Why the only usable correction here is a number written down in advance. Behavioural Finance The Cost Nobody Writes Down The second reason your in-house systems outlive the ones you bought.

Frequently Asked Questions

What is the IKEA effect?

The increase in valuation of self-made products. In four studies involving flat-pack furniture, origami and Lego, participants saw their amateurish creations as similar in value to experts' creations and expected others to share their opinions. A 2026 meta-analysis of 55 studies puts the effect at d equals 0.57.

How well established is it?

Unusually well, by this series' standards. The 2012 paper, a successful conceptual replication in 2017 using different materials, and a 2026 meta-analysis across 5,454 participants reporting the effect robust to moderators including customisation and tangibility. At 0.57 it sits above this publication's own median of 0.41.

What is the boundary condition?

The one part almost never quoted. The founding paper reports that labour leads to love only when it results in successful completion: when participants failed to finish, or built and then destroyed their creations, the effect dissipated. So the inflation attaches to finished work only.

What does it cost when selling a business?

On our own arithmetic, a 0.57 inflation in the seller's valuation raises the probability of no agreement from 50 to 65.7 percent. The cost is not a lower price; it is a transaction that does not happen, and the seller will attribute it to the buyer being unrealistic.

Can I just discount my price a bit?

Not usefully. On our own arithmetic, conceding a third of the effect moves the failure probability from 65.7 to only 60 percent, and restoring an even chance requires giving up the whole thing. This is the opposite shape from ordinary bargaining, where partial concessions are cheap and effective.

Does it ever work in reverse?

Yes, and it matters in 2026. A study of AI-assisted work reports participants achieving significantly higher objective performance yet systematically underestimating it, drawing on the effort heuristic. Perceived value tracks effort rather than quality, which mis-prices in both directions by the same mechanism.

Is my higher valuation ever justified?

Often. Something you built can genuinely suit your operation better than a buyer's, and the experiments deliberately compare identical standardised objects, which removes that. The test that separates them: ask what you would pay for the same thing built by someone else. Genuine fit survives that question and the effect does not.

References

1. Copy of Norton, M. I., Mochon, D., & Ariely, D. (2012), The IKEA effect: When labor leads to love, Journal of Consumer Psychology, 22(3), 453–460, hosted by the third author's own university department, reproducing the abstract, the paper's running head and page range, and its reference list, which includes Ariely, D., Kamenica, E., and Prelec, D. (2008), Man's search for meaning: The case of Legos, Journal of Economic Behavior and Organization, 67; Ariely, D., and Simonson, I. (2003), Buying, bidding, playing, or competing? Value assessment and decision dynamics in online auctions, Journal of Consumer Psychology, 13, 113–123; Aronson, E., and Mills, J. (1959), The effects of severity of initiation on liking for a group, Journal of Abnormal and Social Psychology, 59, 177–181; and Axsom, D., and Cooper, J. (1985), Cognitive dissonance and psychotherapy: The role of effort justification in inducing weight loss, Journal of Experimental Social Psychology. Note: a copy hosted by a co-author's own university department, which is the best provenance short of the publisher. We obtained the abstract, front matter and reference list, and not the data or results tables. people.duke.edu
2. Pelled, A., Demetriades, S. Z., & Walter, N. (2026). Labor Leads to Love, Right? A Meta-Analysis of the IKEA Effect. Psychology and Marketing, DOI 10.1002/mar.70064. Publisher record reproducing the abstract: on the IKEA effect referring to people's tendency to overestimate the value of items they personally assembled or assisted in creating; on the meta-analysis, with k equal to 55 and N equal to 5,454, synthesizing data from empirical studies assessing the relationship between self-assembly and product valuation, to estimate the average effect size, test main and secondary outcomes, and assess potential boundary conditions; on the results pointing to a significant moderate impact of self-assembly labor on valuation at d equal to 0.57 with p below 0.005, as well as a host of secondary outcomes including liking, self-concept, and sense of, at which point the reproduction is cut off; and on the observed effects remaining quite robust in the face of various moderators such as product customization and tangibility. The same record carries a reference list confirming Brunner, F., Gamm, F., and Mill, W., MyPortfolio: The IKEA Effect in Financial Investment Decisions, Journal of Banking and Finance, 154, article 106529; Buechel, E. C., and Janiszewski, C. (2014); Mochon, D., Norton, M. I., and Ariely, D. (2012), Bolstering and Restoring Feelings of Competence via the IKEA Effect, International Journal of Research in Marketing, 29(4), 363–369; and Nguyen, P., and Wang, X. (2024), The Influence of Non-Physicality of Goods on Disparities in Seller-Buyer Valuations: A Meta-Analysis, Journal of Consumer Psychology, 34(3), 445–465. Note: the publisher's record, obtained in two renderings which agree. The abstract is truncated mid-sentence in both, so we report no moderator analyses, publication-bias tests or confidence intervals; we obtained none of the works in its reference list. onlinelibrary.wiley.com
3. Publisher record for the 2012 paper, reproducing the abstract in full: on four studies in which consumers assembled IKEA boxes, folded origami, and built sets of Legos demonstrating and investigating boundary conditions for the IKEA effect, being the increase in valuation of self-made products; on participants seeing their amateurish creations as similar in value to experts' creations and expecting others to share their opinions; on labor leading to love only when labor results in successful completion of tasks, with the effect dissipating when participants built and then destroyed their creations or failed to complete them; and on labor increasing valuation for both do-it-yourselfers and novices. Note: the publisher's record. Our source for the abstract verbatim, corroborated word for word against the copy hosted by a co-author's department. sciencedirect.com
4. Copy of the same paper on the first author's institutional repository, reproducing the citation as Norton, Michael I., Daniel Mochon, and Dan Ariely, 2012, Journal of Consumer Psychology, 22(3), July, 453–460, together with body text in which the authors state that labor alone can be sufficient to induce greater liking for the fruits of one's labor, that even constructing a standardized bureau, an arduous and solitary task, can lead people to overvalue their often poorly constructed creations, and that they name the phenomenon the IKEA effect in honour of the Swedish manufacturer whose products typically arrive requiring some assembly. Note: the first author's institutional repository. Our source for the paper's own explanation of its name and its framing of the central claim. dash.harvard.edu
5. Sarstedt, M., Neubert, D., & Barth, K. (2017). The IKEA Effect. A Conceptual Replication. Journal of Marketing Behavior, 2(4), 307–312, DOI 10.1561/107.00000039. Publisher record reproducing the abstract: on the authors replicating and extending Norton and colleagues' and Mochon and colleagues' 2012 studies on the IKEA effect, according to which consumers show a higher willingness-to-pay when they assemble products themselves; and on the results supporting the robustness of the original effect and indicating that psychological ownership acts as a psychological mechanism underlying it. Together with a repository copy reproducing the same abstract and recording that the study examined whether the effect dissipates when individuals disassemble their creations, that a supplementary table compares the differences and similarities among the three sets of studies, and that participants in the first experimental group were asked to choose one of three predesigned loom bands. Note: the publisher's record plus a repository copy. We obtained the abstract only and not the paper, so we did not obtain what the authors found on the disassembly condition. emerald.com
6. Repository page for the 2026 meta-analysis, which carries alongside it the abstract of a separate study on performance evaluation in human and artificial intelligence collaboration, reproduced there as related work: on generative artificial intelligence improving employee performance and productivity across a wide range of tasks while little is known about how users subjectively evaluate their AI-augmented work; on the authors drawing on the effort heuristic, which links perceived effort to perceived value, to hypothesise that individuals using AI may paradoxically underestimate their performance; and on three experimental studies with 932 participants, involving management students and full-time employees, finding that although AI-assisted participants achieved significantly higher objective performance than unassisted participants, they systematically underestimated their performance, rating it lower than their objectively achieved outcomes. The same page records that AI allows users to produce high-quality results with minimal effort. Note: a repository page for a different paper, on which this abstract appears as related work; we did not obtain the AI study itself, we do not have its authors or its publication venue, it is one study rather than a literature, and our reading of it as a reverse IKEA effect is our own inference. We report it at that low strength. researchgate.net
7. Business school faculty research record for the 2012 paper, confirming the citation as Norton, Michael I., Daniel Mochon, and Dan Ariely, The IKEA Effect: When Labor Leads to Love, Journal of Consumer Psychology, 22, number 3, July 2012, pages 453–460; together with an academic preprint reference list confirming the same citation with DOI 10.1016/j.jcps.2011.08.002, and a practitioner reference page confirming the same and additionally identifying Kruger, J., Wirtz, D., Van Boven, L., and Altermatt, T. W. (2004), The effort heuristic, Journal of Experimental Social Psychology, 40(1), 91–98. Note: three further independent confirmations of the citation and DOI. We obtained none of the works named beyond those cited elsewhere here. hbs.edu

This article discusses research on valuation and is not valuation, transaction or pricing advice; nothing here should inform the price of any actual business. None of the underlying papers was obtained in full. The 2026 meta-analysis abstract is truncated mid-sentence in both renderings consulted, so no moderator analyses or confidence intervals are reported. All arithmetic is the authors' own; the deal-failure calculation assumes a single-shot encounter with no negotiation or competing bidders and is an upper bound, and the dollar table rests on a valuation spread the authors invented because no source consulted supplies one.