Analytics vs DataWarehouse | Community
Skip to main content
S
Sophie_H
Level 2
April 8, 2025
Solved

Analytics vs DataWarehouse

  • April 8, 2025
  • 2 replies
  • 626 views

Hi,

In the screenshot attached, this is a list of user IDs, there is a metric to count distinct value of userID (Visitor ID unique Count). 

I do not understand how this metric displays a different number (#2 in the picture) as the volume of one id listed down (#1), there are clearly 5361 rows.

 

So when I prepare an extract using DW to get the volume of distinct user ID, in my extract I request the user id this way I do my own deduplication in another reporting tool, and I end up with the same figure as #1. So in the report in another reporting tool I will display 5361 users while in Adobe I show 5428 users, this is confusing for the end users of the report (and for me).

Anyone that could help me to understand the difference and if there is anything I have to fix on my end to display in Adobe.

Many thanks

 

 

Best answer by leocwlau

Hi @sophie_h,

 

Suppose you are using the "Approximate Count Distinct" function in the calculated metrics. As the name suggested, it only provides an approximate count instead of an actual one, which uses a hyperloglog algorithm to quickly get the distinct count with no more than a 5% difference.

 

In a table setup like your screen is better to use "row count" as a function to show the number, which will match with the number of (1), as the rows of table are ID themselves. However, if you are going to have a table with some different dimensions, such as site sections to know how many distinct users accessing each section, where row count doesn't work. You will then only be able to use the approximate distinct function to get a rough count of ID.

2 replies

leocwlau
Community Advisor and Adobe Champion
leocwlauCommunity Advisor and Adobe ChampionAccepted solution
Community Advisor and Adobe Champion
April 9, 2025

Hi @sophie_h,

 

Suppose you are using the "Approximate Count Distinct" function in the calculated metrics. As the name suggested, it only provides an approximate count instead of an actual one, which uses a hyperloglog algorithm to quickly get the distinct count with no more than a 5% difference.

 

In a table setup like your screen is better to use "row count" as a function to show the number, which will match with the number of (1), as the rows of table are ID themselves. However, if you are going to have a table with some different dimensions, such as site sections to know how many distinct users accessing each section, where row count doesn't work. You will then only be able to use the approximate distinct function to get a rough count of ID.

    S
    Sophie_HAuthor
    Level 2
    April 9, 2025

    Thanks @leocwlau for your reply, you are right I am using the approximate count so that explains the difference, thanks for pointing this out!

    D
    Deepika_S_V
    Level 2
    April 9, 2025

    The reason the Visitor ID Unique Count (5428) is higher than the number of IDs listed in the table (5361) is because Adobe counts all unique visitor IDs that match the selected filters and date range  not just the ones shown in the table. The Freeform Table displays only a subset of the data (like the top 5361 IDs), but the metric reflects the total unique IDs across the full dataset.

    Reference:https://experienceleague.adobe.com/en/docs/analytics/components/metrics/unique-visitors

    Terms & ConditionsAccessibility statement

    Loading footer...