Computing Retention with Kusto

Question

question

massimo-1 asked Feb 28, '23

Computing Retention with Kusto

Hello,

My end goal is to compute the retention for a given segment of players. In order to do so, I need to understand how to compute retention like you by using Kusto.

As a starting point, I'm trying to compute the cohort size. The correct value would be the one displayed in the given "Thirty Day Retention Report"

I tried the following query: - of all events - keep only the log-in ones - filter the timestamp in order to keep only logins in "2023-02-26" - count the people by counting the unique ids available

 ['events.all']
 | where FullName_Name == "player_logged_in"
 | where Timestamp  between (datetime(2023-02-26) .. datetime(2023-02-27))
 | project EntityLineage_master_player_account
 | distinct EntityLineage_master_player_account
 | count

However, the result is 2484 instead of 2712. Can you spot the issue?

analytics

screenshot-2023-02-28-alle-121556.png (189.4 KiB)

1

10 |1200

Attachments: Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Xiao Zha commented · Mar 01, 2023 at 09:52 AM

Since all times in Reports are in UTC, you may check if the Query time in Date Explorer is set to UTC. If the issue still exists, please feel free to let me know.

0 ·

capture.png (8.1 KiB)

Answer 1 · 2023-03-01T10:24:03Z

massimo-1 answered Mar 1, '23

@Xiao Zha Thank you for your answer! Actually, the query time is also in UTC. Do you have any other idea?

1

10 |1200

Attachments: Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Xiao Zha commented · Mar 03, 2023 at 09:26 AM

An incident: PlayFab Status - Data and telemetry ingestion is delayed that occurred around 2023/2/25 may cause this problem. You can check whether the non-normal data occurs around 2023/2/25. And you can also check if there is still a problem in the data in recent days.

0 ·

question