Airbnb stopped showing either side's review until both had written theirs

Under Airbnb's original review system, a host's review of a guest was published the moment it was submitted, and the guest could then read it before deciding what to write about the host. That ordering created an obvious incentive: publish something positive first to encourage reciprocity, or hold back honest criticism because the other party could still retaliate with a bad review of their own once they saw yours.

Simply asking users to be more honest could not fix this, because the incentive to hedge or retaliate was structural — it existed as long as either party could see the other's review before finishing their own. Airbnb needed to remove the sequencing itself, not the people's willingness to be candid.

Switched to double-blind release: neither review is shown until both are written, so no one can condition theirs on the other's, collapsing the retaliation incentive.

Sequential reveal let the second reviewer condition on the first, creating a retaliation incentive.

Double-blind release removed the ability to react, so hedging gained no strategic advantage.

With no advantage to lying, the only rational strategy was to report the actual experience.

The fix was structural, not moral, so it worked without policing user behavior.

When feedback is sequential, the second writer retaliates; hide both until both are locked in, and honesty becomes the only rational move.

Users reported writing noticeably more candid reviews right after the change, since neither side could react to what the other had posted. The double-blind, simultaneous-reveal review model became a standard design pattern across two-sided marketplaces, adopted in variations by other platforms seeking honest peer feedback without retaliation risk.