I'd also recommend confirming the control and treatment experiences on all target wikis, and checking for translations (although I'm not sure what we could do to speed up translations if they're missing - maybe we just need to get messages into the queue earlier for future experiments)

I did some basic checks [1]

• We are seeing roughly sized hits for page_visit from the different groups (assigned)
• We are seeing a small amount of hide_badge events in the treatment groups and none in the control group
• We are seeing a small amount of heart_badge events showing minimal clicks but evidence that people are clicking it
• I wasn't expecting exposure events for the control but it seems like the code should be firing exposure events for all users in control. Doesn't seem like a big deal.
• The exposure events seemed low. These do not fire if the badge was disabled via hide. These should be equivalent to page_visit - hide_badge. So either the 3 hide_badge events account for half of the page_visit events.... when I look at distinct subject IDs [2] all was good in the world.

I can also confirm we are seeing events from both dewiki and enwiki [3]

So this looks like an experiment that is working as expected!

[1]

SELECT
 action, action_context, experiment.assigned, count(*)
FROM
  wmf_experiments.experiment_event_v1
WHERE timestamp >= from_iso8601_timestamp('2026-06-18T14:30:00')
  AND experiment.enrolled = 'donor-delight-badge'
  group by action, action_context, experiment.assigned

[2]

SELECT
 action, action_context, experiment.assigned, count(distinct experiment.subject_id)
FROM
  wmf_experiments.experiment_event_v1
WHERE timestamp >= from_iso8601_timestamp('2026-06-18T14:30:00')
  AND experiment.enrolled = 'donor-delight-badge'
  group by action, action_context, experiment.assigned

[3]

SELECT
 mediawiki.database, count(*)
FROM
  wmf_experiments.experiment_event_v1
WHERE timestamp >= from_iso8601_timestamp('2026-06-18T14:30:00')
  AND experiment.enrolled = 'donor-delight-badge'
  group by mediawiki.database