The IKEA effect is the finding that people value a thing more when they built it themselves.
Designers quote it to defend long signup flows. Make them work for it and they will stick around. That is half the finding, and it is the wrong half to build on.
Here is the part that gets skipped. Norton, Mochon and Ariely had people assemble a plain IKEA storage box. Builders bid $0.78 to keep the box they made. People handed an identical pre-built box bid $0.48. In a later experiment, one group was stopped two steps before the end. They bid $0.59. The group allowed to finish bid $1.46.
Effort is not what creates attachment. Finishing something is.
Your onboarding asks for plenty of effort. Count how much of it ends in something the user can point at.
1. A profile form is effort with nothing at the end
What teams ship: a progress ring, six fields, a nudge email on day two.
Why it fails: the user does the work and no object exists afterwards. There is no box on the table, just a filled row in your database they cannot see.
Ship instead: make the first task produce a thing with a name. A board, a draft, a page. Something they could show a colleague.
2. The thing has to stay theirs
What teams ship: a guided tour that builds a sample project, then clears it at the end.
Why it fails: in the same paper, people who built a Lego set and then took it apart showed no extra value at all. The hands moved. The object went away. Demo data you wipe does the same thing.
Ship instead: let them build the real one on day one, with their own words in it, and keep it.
3. New users are the right people to ask
What teams ship: a skip button, because asking for work feels rude.
Why it fails: the same authors later showed the effect runs on feeling capable. After an easy test, 33 percent of people chose to assemble furniture themselves. After a test designed to make them feel stupid, 58 percent did. A new user is already in the second state. Building something small is how they climb out of it.
Ship instead: one task they will clearly finish, early, before the feature tour.
4. More steps is not more love
What teams ship: nine screens, because the effect is real.
Why it fails: a meta-analysis of 55 studies and 5,454 people puts the effect at d = 0.57. Real, moderate, not a budget for friction. Joshi and Vogel found the same ceiling in software. Longer prompts raised how much people felt an AI story was theirs. Pushing them to write even more stopped helping.
Ship instead: one build, one artifact, then get out of the way.
5. Templates are not cheating
What teams ship: a blank canvas, so the work feels earned.
Why it fails: the boxes in the original study were standard. No choices, no customising. The effect showed up anyway, and the meta-analysis found it held whether people could customise or not. Blank pages do not buy you ownership. They buy you a bounce.
Ship instead: fill in most of it, leave the last part to them, and make finishing it feel like an event.
| What you ask for | What exists at the end | Ask this instead |
|---|---|---|
| Complete your profile | A row in your database | Name your first project |
| Take the product tour | Nothing | Build one real thing |
| Invite your team | A sent email | Share the thing you just made |
| Choose your preferences | Settings you forget | Save your first view |
| Start from blank | A blank page | Finish a half-built template |
The five minute audit
Open your own onboarding and time the first five minutes. Then answer one question. At the end of it, is there something on screen that the user made, that carries their words, and that will still be there tomorrow?
If the honest answer is a progress ring, you are collecting the effort and keeping the artifact. That is the wrong way round.
Get the PDF