IKEA Effect

011 min

Where the IKEA effect shows up

An illustrative case, using a role rather than a person. A staff engineer spent three weeks building an internal deployment tool. It works. In a planning meeting somebody proposes replacing it with a commercial product that costs $400 a month and covers about 90 percent of what the internal tool does.

The engineer argues against it, and the arguments are specific and technically sound. The commercial tool does not handle one of their deployment targets. Migration would cost a sprint. There would be a vendor dependency. Everyone in the room can follow the reasoning.

Change one input and run it again. Suppose the internal tool had been built by a contractor who left, and the engineer had merely been maintaining it. Same tool, same gaps, same migration cost, same vendor risk. In that version the engineer is usually the one proposing the switch.

The technical facts did not decide the meeting. Who assembled the thing did.

The moment the effect takes hold is earlier than the meeting, and it is worth being able to locate. It is the point where the work is finished and it functions. Before that point people are often sharply critical of what they are building. Afterwards the same object is valued well above what anyone who did not build it would pay for it, and the reasons arrive already phrased as engineering judgement.

022 min

Why the IKEA effect happens

Three mechanisms have been proposed, and the honest position is that they overlap and are not fully separated by the evidence.

Effort justification. The oldest account, going back to Elliot Aronson and Judson Mills in 1959 and cited as the foundation by the researchers who named the IKEA effect. The claim is that effort spent on something creates pressure to see that something as worthwhile, because the alternative is to conclude the effort was wasted. Liking is adjusted to fit the labour already spent.

Psychological ownership. Assembling an object makes it feel like yours in a way that being handed it does not. Norton, Mochon and Ariely designed against the obvious objection here: they showed the increase in value holds over and above what comes from merely being given the product, or merely handling it.

Feelings of competence. Completing a build is evidence that you are capable. The object becomes the record of that evidence, and valuing it highly is partly valuing what it says about you.

The boundary that tells you which part is load-bearing

The most useful finding in the original research is the one that draws the limit. Across four experiments, the researchers report that participants who built their creations and then took them apart, or who were not allowed to finish, showed no increase in willingness to pay.

That result constrains all three mechanisms. If effort alone produced the effect, unfinished builders should still show it, because they still spent the effort. They do not. So successful completion, rather than labour as such, is what the effect runs on.

One further finding closes off a simpler explanation. The effect appeared with plain IKEA storage boxes and Lego sets, which allow no customisation at all. So it is not that people value products tailored to their own preferences. There was nothing to tailor.

031 min

Where the IKEA effect comes from

Michael Norton, Daniel Mochon and Dan Ariely named it in The IKEA effect: When labor leads to love, published in the Journal of Consumer Psychology in 2012.

Their first experiment is the plainest. Fifty-two participants at a university in the southeastern United States were paid $5. Half assembled a plain black IKEA Kassett storage box from its unassembled parts and instructions. The others were handed an identical box already assembled and given time to inspect it. Everyone then bid for the box they had in front of them, under an incentive-compatible procedure that made underbidding pointless.

Builders bid an average of $0.78. Non-builders bid $0.48. Asked how much they liked the box on a seven-point scale, builders averaged 3.81 and non-builders 2.50.

The box was chosen deliberately. Participants rated it toward the utilitarian end of a hedonic scale, and it comes in one colour and one configuration. A plain storage box is the hardest case for any explanation involving self-expression, which is exactly why they used it.

The paper opens with a story about instant cake mixes in the 1950s, which manufacturers are said to have reformulated to require adding an egg. The authors present it as an anecdote and say plainly that several reasons probably contributed. It is an illustration in the paper, not evidence, and it is worth repeating with the same caution they used.

042 min

The origami experiment

The second experiment is the one worth knowing in detail, because it measures how large the effect is rather than only that it exists.

One hundred and six people were approached in a student centre at a university in the northeastern United States. Some, the builders, were given origami paper and an instruction sheet and asked to fold a crane or a frog. They then bid for their own creation, between 0 and 100 cents. A second group, who had not built anything and did not know whose work it was, bid on those same amateur creations. A third group bid on origami folded by two research assistants with a great deal of experience.

Read the first two rows together: builders priced their own work at roughly four and a half times what a stranger would pay for the identical object. Then read the first and third rows together. Builders valued their amateur origami at close to what people paid for expert work, and the difference between those two figures was not statistically significant.

The obvious objection is that builders were overbidding to avoid losing something they had made, rather than genuinely believing it was worth that much. The researchers tested it. They asked a further 14 people to fold cranes, bid, and then estimate what the average student would bid. The two figures came out at $0.19 and $0.21. Builders were not defending their origami against the market. They expected the market to agree with them.

That is the finding that makes this more than a curiosity about valuation. The effect distorts not only what you think your work is worth, but what you predict others will think.

051 min

Individual effects

For one person, the distorted decision is almost always the same one: whether to keep their own work or replace it with something better.

The direction of the error is consistent. You overvalue what you built, so you keep it too long, defend it too readily, and discount alternatives you would have chosen on the merits if somebody else had built the incumbent.

The specific cost is not the artefact. It is the comparison that never happens honestly. A fair evaluation of a build-or-buy decision requires pricing your own option at what it is worth, and the original research says you cannot do that reliably for something you made.

There is a second cost that is easier to miss. Because builders expect others to share their valuation, they are repeatedly surprised when colleagues, users or buyers do not. The $0.05 the strangers bid is the number the market was always going to produce. The $0.23 felt like a fact from the inside.

061 min

Systemic effects

Across a team the effect compounds through a specific mechanism: the people who build a thing are usually the people asked whether to keep it.

That is an ordinary and sensible arrangement, because builders know the most about what they built. It also routes every keep-or-replace decision through exactly the group whose valuation of that thing is inflated, and it does so repeatedly rather than once.

The visible result is an organisation with more internal tooling than it needs and more of its own frameworks than it can maintain. Each individual decision was defended on technical grounds by the best-informed person available. Not-invented-here syndrome is often described as pride or as a cultural problem. A large part of it is this effect operating inside a reporting line that puts builders in charge of build-or-buy reviews.

The compounding is worse than for a single decision, because maintained code accumulates. A tool kept for a year has a year of integrations attached to it, which raises the cost of replacing it later and makes the next defence of it genuinely stronger than the last one.

072 min

How the IKEA effect shows up in product, design, and AI

Deliberately used, and named in the original paper. Norton and colleagues point at Build-a-Bear, where customers pay a premium for the opportunity to assemble a toy themselves, and at farm stays where guests harvest the food they then eat. The general pattern in software is onboarding that asks the user to configure a workspace, import data, name things and set preferences before they see value. Those steps are usually defended as necessary setup. Part of what they buy is attachment to the result.

The design judgement is about where the labour goes. Effort that produces something the user can see working is on the mechanism. Effort spent on forms that produce nothing visible is friction, and the research is specific that an unfinished or destroyed build produces no effect at all.

Accidentally suffered: design critique. A designer presenting work they have finished is not in a neutral position to hear that it should be scrapped, and neither is a reviewer defending a component in a design system they authored. The practical guard is to bring alternatives into the room before either one is finished.

Accidentally suffered: prompts and agent scaffolding. This is where the effect is currently most expensive in AI work. Teams build elaborate prompt chains, retrieval pipelines and agent harnesses, and those artefacts represent weeks of visible, successful labour. When a simpler approach or a newer model matches them, the comparison is made by the people who built the scaffolding. Fix the evaluation set before building, and re-run the simplest baseline on a schedule rather than when somebody suggests it.

In user research. A participant who has configured something in a session will rate it higher afterwards. If you need a clean preference measurement between two options, do not let the participant build one of them.

082 min

How to work with the IKEA effect

This concept is unusual among biases: it is both something to guard against in your own judgement and something legitimate to design with. Both directions follow from the same mechanism.

Guarding against it. Awareness is not enough, and nothing in the research suggests it would be. Change who decides or what is being asked.

• Have build-or-buy reviews led by somebody who did not build the incumbent, and give them the requirements rather than the history.
• Write down what the internal option must deliver before building it, then evaluate against that list rather than against the finished artefact.
• State the price. "Would we pay $400 a month for what we have?" is a sharper question than "is what we have good enough?", because it forces the valuation into the open where the $0.23 and the $0.05 can be compared.

Designing with it. The conditions are strict, and they come straight from the boundary in the original experiments.

So the effort you ask of a user must be finishable in the session, must visibly succeed, and must produce something they can see. An onboarding step that ends in a working dashboard qualifies. A seven-field form that ends in a confirmation message does not.

The line between using this and exploiting it is the same one that applies to any attachment mechanism. Building attachment to something the user genuinely wants is design. Building attachment in order to make leaving feel like a loss is a retention tactic wearing design's clothes, and users notice the difference at the moment they try to leave.

092 min

Common misunderstandings

"Any effort makes people value something more." The single most common misreading, and the original research rules it out. Participants who built and then dismantled their creations, and those prevented from finishing, showed no increase in willingness to pay. Effort without completion buys nothing.

"It is the same as sunk cost." Sunk cost is about continuing to invest in something because of what you already spent. The IKEA effect is about valuing a finished object more highly. A builder under the IKEA effect would price their origami at $0.23 to a buyer today, with no further investment in question.

"It is about customisation." The experiments used plain black storage boxes and Lego sets that permit no personalisation. Customisation raises valuation too, and it is a separate literature. The IKEA effect was demonstrated with it removed.

"It means people who build things have bad judgement." It means their valuation of one specific object is inflated, and that they expect others to share it. Their technical judgement about that object is usually the best available. Both can be true, which is what makes the effect hard to manage: you need the builder's knowledge and you cannot use their price.

Where guarding against it would be a mistake. Attachment to your own work is also what sustains maintenance, documentation and care over years. An organisation that successfully stripped this out would find fewer people willing to look after anything.

101 min

IKEA effect vs. nearby concepts

The quickest test in a live discussion: ask whether the person defending the thing built it. If they merely own it, you are looking at the endowment effect. If they built it and finished it, this one is in play as well, and the two add.

112 min

Where the evidence is contested

The effect itself has held up better than most findings of its era, and the interesting disagreement is about the mechanism rather than the existence.

On existence, the strongest evidence available is a meta-analysis by Ayellet Pelled, Stefanie Demetriades and Nathan Walter, published in Psychology & Marketing in 2026. Pooling 55 studies and 5,454 participants, they report a moderate effect of self-assembly on valuation, d = 0.57, and find it holds up against moderators including customisation and tangibility. They also find the secondary outcomes the original paper described: liking, self-concept and sense of ownership all move.

On mechanism, the sharpest challenge comes from developmental work. Lauren Marsh, Patricia Kanngiesser and Bruce Hood published When and how does labour lead to love? in Cognition in 2018. They found the effect emerging around age five, which fits its link to self-concept. The part that complicates the standard story is what they found next: among five- and six-year-olds the effect was not moderated by how much effort was invested, and evaluations did not differ according to whether the object was shown to others.

Stated at its strongest, that is a real problem for the effort-justification account. If labour raises value because effort has to be justified, more effort should justify more value. In that sample it did not. The authors propose self-extension instead: the object is valued because it has become part of how the child sees themselves, and a small amount of making is enough to achieve that.

Where this leaves a practitioner is clearer than the mechanism debate. The effect is real and moderate in size, so plan for it. But if you are designing with it, do not assume more work produces more attachment. The evidence that completion matters is strong and the evidence that quantity of effort matters is not.

?8 questions

Questions people ask

§5 sources

Sources