Make them do a little of the work

Everything else on this site pushes one way: take work away from the user. Carry their state so they never have to hold it. Cut the step. Make one button obvious and let the rest go dim. Respect the four slots. This piece is the exception, and it’s a large one. There is a kind of effort that makes people value a product more when you ask it of them, and value it less when you take it away. Not despite the work. Because of it.

The cleanest demonstration is almost silly. In a set of studies now filed under the “IKEA effect,” people assembled their own plain furniture, or folded their own origami, or built their own Lego, and were then asked what they’d pay to keep it. They consistently priced their own handiwork far above the identical thing built by a stranger. In the origami version they valued their own lopsided creations roughly as highly as an expert’s, while neutral onlookers saw them for what they were: worse. The labor did not make the object better. It made the object theirs, and ownership is worth money the object never earned on its merits.

The older cousin of this finding is stranger. In 1959, Aronson and Mills made people go through an embarrassing screening to join a discussion group, then sat them through the same deliberately dull conversation everyone else heard. The ones who’d endured the awkward initiation rated the group as more interesting. Having paid a price to get in, some part of the mind refused to conclude that the thing behind the door was boring — because that would mean the price bought nothing. Effort, once spent, quietly recruits your judgment to justify it.

Now hold that against the four-slot gospel and you seem to have a flat contradiction. One essay says strip the load. This one says add the effort. The resolution is that load and effort are not the same axis, and mistaking one for the other is where products go wrong. Cognitive load is the tax of confusion — a code held across screens, a hunt for the primary action, re-deriving where you are. It is always a cost, and it never buys ownership. Nobody feels proud of having decoded a bad menu. Contributory effort is a different thing entirely: labor whose result is visibly yours. A choice that shaped the output. Something you made that now sits in front of you with your fingerprints on it. That kind of effort can be an asset for the very reason load is a liability — it leaves a trace of the person inside the product.

Most teams can’t tell the two apart, so they make one of two mistakes. They strip everything, load and contribution together, and ship something frictionless and forgettable — the app you opened once and felt nothing toward. Or they bolt on busywork, call it “engagement,” and read the user’s irritation as investment.

So the shape isn’t a slope, it’s a hump. At zero effort there’s nothing of the user in the thing; it stays a stranger’s product they happen to be renting. Push the effort too high and they leave before ownership ever forms, because the payoff lands after the moment they’d have abandoned it. The amount you want is the amount that’s enough to feel authored and not enough to feel like a chore, and that number is specific to your product. There is no universal setting for it.

Here’s what it looks like in a build. The setup where a user names their workspace, picks their columns, drags their first three items into place — that’s not onboarding friction to be minimized toward zero. It’s the moment they start signing their name to the thing, and a product that “helpfully” auto-configures all of it can quietly rob them of that. The goal you set in a fitness app, the interests you choose, the board you arrange: each is a small deposit of self that raises the cost of leaving — not because you trapped anyone, but because they built something they’d now have to walk away from. The discipline is to make the effort shape a real output they keep, never to invent hoops. A profile field that changes nothing downstream is busywork. The same field, when it visibly tailors what they see next, is contribution. Same click; opposite psychology.

And here is where I have to turn on my own argument, because the failure modes are severe and this idea lives a short walk from a dark pattern.

Isn’t this just a rationalization for lazy, effortful UX? It can be. The line is whether the effort builds toward something the user wanted or merely toward your numbers. Effort justification will make people rate a dull group more interesting — which means it can prop up a mediocre product for a while on nothing but sunk cost. That isn’t a moat. It’s a debt, and it comes due the moment a genuinely better, frictionless competitor appears and the user realizes the wobbly thing they built was never worth defending. Manufactured effort with no real payoff is a con, and people forgive that exactly once.

There are three places the effect doesn’t bend but reverses. Transactional, high-intent flows — checkout, payment, anything where the person has already decided and just wants out the door — punish every added step; ownership psychology has no grip on someone trying to leave with their purchase. Effort that builds no competence and no authorship — a captcha, a pointless form — breeds resentment, not endowment, because nothing of the user survives it. And the whole thing inverts if what they build turns out bad: the IKEA effect holds only when the assembly succeeds. Hand someone a kit that snaps in their hands and it’s their own failure they’ll remember, and you they’ll blame. Ownership of something broken is worse than no ownership at all.

I’d also be overselling if I made the science sound firmer than it is. The IKEA effect replicates, but with modest, boundary-conditioned effect sizes. The Aronson–Mills initiation study is past sixty-five years old and people still argue about what it really showed. Take all of it as directional: a real force to design with, not a setting you can dial to a known value.

Which means the audit most teams run is the wrong one. “Where can we remove a step” finds load and contribution alike and cuts both, and the product comes out smoother and less loved in the same pass. The harder question — the one worth the meeting — is which steps are the tax of confusion, and which are the user quietly putting their name on the thing. In a funnel report those two are indistinguishable. Both are friction; both are places people drop. That’s exactly why the frictionless, forgettable product is the easiest one to build by accident.