Publish a postmortem for every public incident, in plain words, within days of the outage.
If you go silent after the fix, customers explain the incident to each other with screenshots, Reddit threads and guesses.
In January 2017 GitLab lost about six hours of production data after several backups failed. The team livestreamed the recovery and published a detailed postmortem.
I wrote a chapter in The Crisis Playbook on how to write yours.