Endowment effect
Owning something raises what you would sell it for above what you would pay.
People demand more to give up an object than they would pay to acquire the same object. The gap between willingness to accept and willingness to pay appears immediately on ownership and does not require any use or attachment over time. Whether it reflects loss aversion: Losing something feels worse than gaining the same thing feels good., ownership itself, or features of how the experiments are run is actively disputed.
How it shows up in software
Trials, free plans, and customisation all create a sense of ownership before any money changes hands. A configured workspace, a built playlist, a trained model, and a populated profile are all things a user now owns and will price above what they would have paid to get. The same mechanism makes churn: The rate at which customers stop paying or stop using the product. sticky and makes migration off a competitor hard.
Using it well
- Get real user content into the product early, since a workspace with the user's own data is the thing they will not want to give up.
- Let people customise something meaningful in the first session: a view, a workflow, a name.
- On downgrade, preserve what the user built in read-only form rather than deleting it, so the decision is reversible.
- Remember the effect cuts both ways: your users are endowed with their current tool too, so switching costs are your acquisition problem.
Where it turns manipulative
- Giving a user an entitlement then removing it to force an upgrade manufactures a loss you created. Free-tier limit cuts applied retroactively read this way whatever the reasoning.
- Making export deliberately lossy so what the user owns cannot leave is a lock-in pattern, and data portability duties under GDPR and the EU Data Act speak to it.
- Trial flows that build heavy setup and then hold the result hostage at the paywall trade on ownership rather than value.
Where you have seen it
Spotify
Playlists and library remain associated with the account when a subscription lapses to the free tier.
Canva
Designs made with premium elements stay in the account, with the paid elements watermarked until unlocked.
Google Takeout
Users can export their data across Google services into standard file formats.
What the research says
- Kahneman, Knetsch and Thaler, 1990Mixed evidence
Participants randomly given a mug set a median selling price roughly double the median price buyers would pay, and the gap persisted across repeated trials with real money.
- Plott and Zeiler, 2005Contested
When participants were given extensive practice, anonymity, and clearer instructions, the gap between willingness to accept and willingness to pay disappeared, suggesting a large part of the classic result comes from subject misconception.
- List, 2003Mixed evidence
In a real sports card market, inexperienced traders showed the effect while experienced dealers did not, indicating that market experience reduces or removes it.
This is one of the most genuinely contested items in this file. The phenomenon shows up reliably in standard lab procedures and weakens or vanishes under tighter procedures and with experienced traders. Treat it as a real effect on novices in unfamiliar transactions, not as a universal property of ownership.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.