The IKEA Effect

What it is

Effort is supposed to be a cost. In most accounts of consumer choice, the work required to obtain something is subtracted from what the thing is worth, which is why so much of product design is spent removing steps. The IKEA effect describes the opposite arithmetic, in a narrow set of circumstances: the work performed on an object is added to the value of the object, by the person who did the work, and by nobody else.

Norton, Mochon and Ariely reported four studies in the Journal of Consumer Psychology in 2012, under the title "The IKEA Effect: When Labor Leads to Love". Participants assembled IKEA storage boxes, folded origami and built Lego sets. The origami study is the one that reaches most retellings, including ours. Amateur builders bid substantially above what their folds were worth to anyone else, valuing their own creased efforts near the level of pieces folded by experts, and well above what outside bidders would pay for the same objects. The gap between what the maker will pay and what a stranger will pay is the effect.

The paper's boundary condition deserves as much space as the result. Labour led to love only where the labour finished. Participants who built something and then dismantled it, or who did not complete the task, showed no inflation at all. So the mechanism is not simply that sweat gets capitalised. Something about a completed object, an object that exists because of you, is doing the work, and the 2017 conceptual replication points at psychological ownership as the candidate.

That replication, by Marko Sarstedt, Doreen Neubert and Kati Barth in the Journal of Marketing Behavior, reproduced the higher willingness to pay for self-assembled products and supported the robustness of the original. A meta-analysis followed in Psychology & Marketing in 2026: 55 studies, 5,454 participants, a pooled effect of d = 0.57 on valuation, with secondary effects on liking, self-concept and sense of ownership, and with the result surviving moderators including product customisation and tangibility.

The vault records this concept as primary-cited, which means the primary citation is present and correct and the wider literature had not been assessed when the note was written. That assessment has now been started rather than completed. The replication and the meta-analysis above were checked at search level on 2 September 2026, at abstract and bibliographic-record level, and the tag stays as it is until the papers themselves have been read.

Nir Eyal imports the finding into product design in Hooked, at pp. 136-138, and the import is the reason most people in technology have heard of it. His argument is that the effort a user puts into a product, a completed profile, a curated playlist, a board of saved images, is not friction to be engineered away. It is what makes the product theirs, and things that feel like ours are harder to abandon.

In effect

Origami its makers would not sell

Participants folded paper cranes and frogs, badly, then bid against strangers for the pieces. Their own bids ran far above what the same objects fetched from anyone else, and close to what people bid for expert work. The builders were not confused about the quality of the folding. They were valuing a different object from the one everybody else could see.

The bookcase you will not throw out

A Billy bookcase assembled over an afternoon, with one shelf slightly out of true and a spare cam lock left in the bag, outlives several better bookcases. It moves house. It gets defended. The version of it that arrived pre-built and was carried in by two men in a van goes on the pavement without ceremony when something nicer turns up.

The failure case is in the same room. A flat-pack half-built and abandoned in the spare bedroom, still in its plastic, produces no attachment whatever, which is exactly what the original paper found when it stopped people before completion.

Investment stages, and effort as the product

Pinterest boards, Spotify playlists, a LinkedIn profile filled in field by field, a Duolingo streak. Each of these asks the user for work that pays no immediate reward, and the work is the point. Eyal's fourth Hook stage is built on it, and the design instruction that follows is unusually blunt: ask the user to do something, and make sure they finish.

The commercial version has a feature the laboratory version does not, which is that the accumulated work is also a switching cost. Whatever the user believes about the value of their playlist, they cannot take it with them, and the platform benefits either way.

What it does not say

It does not say that effort creates value in general. The 2012 paper found the effect only where the task was completed. Incomplete or destroyed work produced nothing. A product that asks users for effort and then frustrates them is not banking goodwill, it is running the condition under which the effect disappears.

It does not say the object is better. Nothing in this literature touches quality. What moves is the builder's valuation, and the fact that outside bidders do not follow it is half of the evidence for the effect existing at all.

It does not establish that this is why people stay with software. Bidding on origami in a laboratory and abandoning a service after eight years are different behaviours, and no study we have read tests the second. Eyal's extension of the mechanism to products is a good hypothesis about switching costs, offered as an argument rather than a finding, and it should be quoted as his.

The pooled figure carries its own limits. A d of 0.57 across 55 studies is an average over Lego, origami, flat-pack furniture and hypothetical assembly tasks, most of them run on small samples, and it does not license a prediction about any particular onboarding flow. We have that number from an abstract, not from the paper.

Who developed it

Michael I. Norton, of Harvard Business School, is first author, with Daniel Mochon and Dan Ariely, in the Journal of Consumer Psychology in 2012. He has no page on this site yet.

Our vault credited "Ariely, Norton and Mochon" until 1 September 2026, fronting the famous name, which is the most reliable failure in this genre and the same error the site has had to correct in seven other notes. Naming the first author is the rule here regardless of who is better known. There is a second reason in this case, recorded in the vault's note on Norton: questions have been raised about data integrity in other work Ariely co-authored, none of it this paper, and an attribution that puts his name first invites a doubt the finding has not earned.

---