Why You Love the Thing You Assembled
People value what they built above what somebody else built. The boundary condition nobody quotes is that it only works when the building works, which makes this a finding about completion and not about effort.
Key takeaways
- Norton, Mochon and Ariely ran four studies using IKEA boxes, origami and Lego, and found people valuing their own creations close to experts' creations.
- Participants also expected other people to share that valuation, so the misjudgement comes with a prediction about everybody else.
- The effect dissipated when participants destroyed their creations or failed to complete them, which makes completion rather than effort the active ingredient.
- Marsh, Kanngiesser and Hood found the bias emerging around age five and, in five and six year olds, not moderated by the amount of effort invested.
- Nobody in this literature appears to have measured how long the inflated valuation lasts.
Four studies, three materials, one result. IKEA boxes, folded origami and sets of Lego, and in each case people valued what they had built above what somebody else had built. The finding has a name, the IKEA effect, and it is one of the few things in consumer psychology that you can test on yourself this afternoon.
Our position before the evidence, because it is the useful half. The effect is real and the thing that makes it interesting is its boundary, which the original paper reports and almost nobody quotes. Labour leads to love only when the labour works. Build something and fail to finish it, or finish it and then destroy it, and the valuation goes back to normal. That makes this a finding about completion rather than about effort, and the commercial version has the emphasis in the wrong place.
What did they actually measure?
Michael Norton, Daniel Mochon and Dan Ariely, in the Journal of Consumer Psychology, in the 2012 volume, published online in September 2011. Participants assembled IKEA storage boxes, folded origami and built Lego, then valued their own output against output made by somebody else.
> People rated their own amateurish output as worth about what an expert's was worth, and expected everybody else to agree.
Two findings in the first half. People saw their amateurish creations as similar in value to experts' creations, which is the part that gets laughed at, and they expected other people to share that opinion, which is the part that should worry anybody selling something. The misjudgement is not private. It comes with a prediction about everyone else.
Then the boundary. When participants built and then destroyed their creations, or failed to complete them, the effect dissipated.
Why does that boundary change the advice?
Because the folk version of this says effort creates attachment, and the paper says successful effort creates attachment.
Those produce opposite design decisions, and they can be set out side by side.
| If the ingredient is | Then more friction | And an abandoned setup |
|---|---|---|
| Effort | Helps, because effort is the point | Still produces attachment |
| Completion | Helps only until somebody finishes | Produces none at all |
| If effort is the active ingredient, more friction is better, and a harder setup flow should produce a more committed user. If completion is the active ingredient, friction is only useful up to the point where somebody finishes, and every extra step past that is a chance to not finish, which the paper says removes the effect entirely. |
Nir Eyal's Hooked builds a product model on the second half of the loop, which he calls investment, and it is the one phase of his four that deliberately asks the user for effort: small pieces of work that make the product more valuable to the person who did them. Read against this paper, the model is right to ask for work and would be wrong to ask for difficult work.
### Why the destroyed-creation condition matters
The destroyed-creation condition deserves its own paragraph, because it is the cleanest part of the design.
If valuing your own work were simply a matter of having spent time on it, then time spent and then wasted should still produce the attachment. The sunk effort is identical. What the participants lost was not the effort but the object, and the valuation went with the object rather than with the hours.
That is the difference between an effort account and a completion account, and it is testable in one condition. Few findings in this area separate their own mechanisms that cleanly.
### What this does to the incomplete-signup problem
It also predicts something commercially specific that nobody seems to have tested.
If completion is the active ingredient, then an abandoned setup is worse than no setup. The person has spent effort, acquired nothing they made, and on this evidence should feel no attachment at all. A long onboarding flow with a high drop-out rate is therefore not a flow that works less well. It is a flow that produces two populations, one attached and one with nothing.
Whether that is true of real products is a question these studies cannot answer, because nobody in them abandoned a subscription. We raise it because the prediction is unusually clear and the test would be cheap.
What happens when somebody tests the mechanism?
Here is the result that unsettles the mechanism, and it comes from an unusual direction.
Lauren Marsh, Patricia Kanngießer and Bruce Hood took a developmental approach in Cognition, published online in 2017 and in the 2018 issue, asking when the bias appears and what drives it. Three studies.
The bias emerges at around age five, which they connect to milestones in how a self-concept forms. That is interesting on its own. The second study is the one that matters here: in five and six year olds, the IKEA effect was not moderated by the amount of effort invested in the task.
If the effect does not scale with effort, effort is not doing the explanatory work, at least not there.
Their third study goes after the next obvious candidate, ownership, and rules it out too: ownership alone cannot explain why children value their creations more. Their own summary of the series is blunter than anything we would have written on their behalf. The results, they say, are incompatible with existing theories of the IKEA bias.
What they propose instead is a link between our creations and our self-concept, emerging at age five, which is why the thing you made is valued above the identical thing somebody else made. That is a different kind of explanation from either effort or ownership, and it is the paper's actual contribution rather than a loose end.
We are being careful about the scope. This is a developmental sample and the original was adults valuing boxes and Lego, so none of it establishes the adult mechanism. What it does is remove the two explanations everybody reaches for and put a third on the table.
### The age-five finding on its own terms
The ontogeny is worth a moment even if you care only about adults.
A bias that is absent at three and present at five is a bias that arrives with something, and Marsh and colleagues connect it to the formation of a self-concept. That is a claim about what kind of thing the effect is: not a calculation about effort, and not a property of objects, but something to do with where a person draws the line around themselves.
If that is right, then the reason your own shelf is worth more to you is closer to the reason your own handwriting is, which is a stranger and more interesting answer than effort justification.
Where does that leave the commercial version?
Narrower than it is sold, and more useful.
The onboarding flow that asks you to pick preferences, name things, upload a photo and arrange a dashboard is doing something the evidence supports: it is getting you to produce an output you can look at. The flow that is merely long is not, and a setup that asks you for small pieces of work is doing something different from one that merely takes a while. The difference is whether the user ends up with something finished that they made, which is the question to ask about what a setup flow is actually buying.
The same test separates the honest version from the manipulative one. A setup that leaves you with a configured tool has given you the thing you now overvalue. A setup that leaves you with nothing but sunk time has used the mechanism and given you no object.
What would change our mind?
An adult test of the self-concept account. Marsh and colleagues removed effort and ownership in children and proposed a creation-to-self link in their place, and the obvious next study is whether that holds in the population the original used. If adult valuation does scale with effort, the mechanism is effort after all and the developmental result is about children rather than about the effect.
We also do not know how long any of this lasts. Every measurement here is a valuation taken shortly after building. Whether the thing you assembled in March is still overvalued in September is a question nobody in this literature appears to have asked, and it is the one that decides whether this is a fact about products or a fact about afternoons.
So the concession we would make to a sceptic is real, and larger than we expected when we started. The headline finding is well established and replicated across materials. This is now a finding whose mechanism is not settled, since both of the explanations usually offered for it have been tested and found wanting in children, the duration is unmeasured in anybody, and the commercial advice drawn from it rests on the half of the original paper that the original paper itself qualifies.
Common questions
What is the IKEA effect in psychology?
The IKEA effect is the finding that people value things they assembled themselves above identical things assembled by somebody else. Norton, Mochon and Ariely demonstrated it across four studies using IKEA storage boxes, origami and Lego, and found that participants rated their own amateurish output close to experts' output while expecting other people to agree with them.
Does the IKEA effect depend on how much effort you put in?
Not on the available evidence. Marsh, Kanngiesser and Hood tested effort directly in five and six year olds and found the effect was not moderated by the amount invested. The original paper points the same way from a different angle: the effect disappeared when participants destroyed what they had built or failed to finish it, so successful completion rather than exertion is the condition.
Is the IKEA effect the same as the endowment effect?
They are related and not identical. The endowment effect concerns owning something, and the IKEA effect concerns having made it. The distinction matters because Marsh and colleagues tested ownership as the explanation for the IKEA effect and found that ownership alone cannot account for it.
How do companies use the IKEA effect?
Through setup flows that ask you to configure, name, arrange or upload something, so you finish with an output you made. On this evidence the useful version is one you complete, and length on its own is not the mechanism.
Sources
Norton, M. I., Mochon, D., & Ariely, D. (2011). 'The IKEA effect: when labor leads to love.' Journal of Consumer Psychology, 22(3), 453-460. Four studies using IKEA boxes, origami and Lego; participants valued self-made products close to experts' products and expected others to agree; the effect dissipated when creations were built and then destroyed, or not completed. Issued online 9 September 2011 in the volume 22 issue. Verified at Crossref and the published record 2026-10-05. doi:10.1016/j.jcps.2011.08.002
Marsh, L. E., Kanngiesser, P., & Hood, B. (2017). 'When and how does labour lead to love? The ontogeny and mechanisms of the IKEA effect.' Cognition, 170, 245-253. Three studies; the bias emerges at age five alongside self-concept milestones; it was not moderated by the amount of effort invested in five to six year olds; the third study examines whether feelings of ownership moderate it. Verified at the published record 2026-10-05. doi:10.1016/j.cognition.2017.10.012
Eyal, N. (2014). Hooked: How to Build Habit-Forming Products. Portfolio. Ch. 5, Investment, pp. 135-162, the phase that deliberately asks for effort. Read in full in this vault, 2026-08-30.
Alter, A. (2017). Irresistible. Penguin Press. The six-ingredient account of behavioural addiction, which includes escalating difficulty. Read in this vault, 2026-09-01.
A note on scope: the developmental result is a developmental sample, so a non-moderation by effort in six year olds does not establish non-moderation in adults. It moves effort from an assumption to a hypothesis.