we’re testing a substantial distribution with >20K users, which makes bulk updates really difficult as the service is crashing intermittently and sending users back to the main page. therefore, we thought that a bulk delete might make more sense from a prune on the database side