It appears that out of the box Quicksight doesn’t support any type of distinct count level aware calculations.
Does anyone know of any work arounds?
I have a dataset like this:
| customer_id | event_type | event | first_period | cohort_month | vertical |
|---|---|---|---|---|---|
| cus_EmwpgeOVwihaBQ | feature_usage | login | 1 | 2026-02-01 | investor |
| cus_EmwpgeOVwihaBQ | feature_usage | Viewed Profile | 1 | 2026-02-01 | investor |
| cus_EmwpgeOVwihaBQ | feature_usage | Printed Report | 1 | 2026-02-01 | investor |
| cus_Eoppet1VxDPwh3 | feature_usage | Viewed Profile | 1 | 2026-01-01 | commercial |
| cus_Eoppet1VxDPwh3 | feature_usage | login | 1 | 2026-01-01 | commercial |
| cus_ErToGpAnXzw4OZ | purchase_transaction | Criteria Bot | 1 | 2025-11-01 | commercial |
| cus_ErToGpAnXzw4OZ | feature_usage | Viewed Profile | 1 | 2025-11-01 | commercial |
| cus_ErToGpAnXzw4OZ | feature_usage | Printed Report | 1 | 2025-11-01 | commercial |
| cus_ErToGpAnXzw4OZ | feature_usage | login | 1 | 2025-11-01 | commercial |
| cus_EsDTFS9eV7YHdX | feature_usage | login | 1 | 2026-02-01 | commercial |
| cus_Eu60j7sq0Qzgnw | feature_usage | Viewed Profile | 1 | 2026-01-01 | residential |
| cus_Eu60j7sq0Qzgnw | feature_usage | login | 1 | 2026-01-01 | residential |
| cus_Eu60j7sq0Qzgnw | purchase_transaction | Criteria Bot | 2 | 2026-01-01 | residential |
| cus_Eu60j7sq0Qzgnw | feature_usage | Created List | 2 | 2026-01-01 | residential |
| cus_Eu60j7sq0Qzgnw | feature_usage | Added to List | 3 | 2026-01-01 | residential |
What I want to accomplish is to utilize a barchart visual and have it display the percentage of customers that have completed an action AND be able to filter by vertical.
I can probably create a new dataset specifically for this purpose, but I’d love to be able to actually just use quicksight directly rather than having to make new datasets to fill in all the gaps quicksight has in its feature set.