Google Analytics - Innaccurate Avg. Session Duration and Pages/Session Data

Sorry if this has already been asked, I could not find a similar enough question in the forums.

Recently I have had a lot of difficulty matching my Google Analytics data in Domo.
For instance I have a campaign where it is important to be able to drill down to show
A certain Source/Medium, a set Campaign Name and traffic by Ad Content.

From these dimensions, I need to show Average Session Duration and Pages/Session.

In one example, I am looking at traffic from July 2020 for users arriving via just one source/medium and via my campaign. When replicating this filter in GA and Domo I get an Average Session Duration of 9 and 30.11 seconds respectively. When I change to Pages / Session, I get 1.26 and 0.52 respectively. Obviously both are way off the GA figures.

In an attempt to improve accuracy, I am only pulling through very few metrics and dimensions:
Avg Session Duration, Bounce Rate, Sessions, Pages / Session
Date, Source/Medium, Ad Content, Campaign

Best Answer

  • user08001
    user08001 Member
    Answer ✓

    I have found a solution and am posting here so that others can learn from my mistake.

    It seems Domo was successfully pulling data from GA, however the fault was in the interpretation. For instance: consider the following:

    DateSessionsPages / Session
    1103
    2202.5
    33001


    When calculating the average Pages / Session, Domo only averages the Pages / Session column. Therefore the average of 3, 2.5 and 1 = 2.17.


    Google Analytics however is calculating the average Page / Session figure, per Session. As the original calculation did not factor in Sessions for any given day, days in which Sessions peaked, did not have enough impact.

    To match GA's calculation in Domo, I added a Beast Mode which reads:
    sum(`Sessions` * `Pages / Session`) / sum(`Sessions`)
    This in the above example equates to 1.15 which finally matches Google Analytics.

Answers

  • Hi @user08001 

     

    With what granularity are you pulling your data in GA and also when importing into Domo? Some metrics will be different if you're using different granularities as GA will dedupe data for certain metrics.

     

    Have you confirmed your Domo API request is the same as how you're pulling the data within Google? Have you tested with the Google Analytics Query Explorer (https://ga-dev-tools.appspot.com/query-explorer/)? Does that return the same information as Domo or GA?

     

    Are you taking into account the same filters or partitions between GA and Domo?

    **Was this post helpful? Click Agree or Like below**
    **Did this solve your problem? Accept it as a solution!**
  • user08001
    user08001 Member
    Answer ✓

    I have found a solution and am posting here so that others can learn from my mistake.

    It seems Domo was successfully pulling data from GA, however the fault was in the interpretation. For instance: consider the following:

    DateSessionsPages / Session
    1103
    2202.5
    33001


    When calculating the average Pages / Session, Domo only averages the Pages / Session column. Therefore the average of 3, 2.5 and 1 = 2.17.


    Google Analytics however is calculating the average Page / Session figure, per Session. As the original calculation did not factor in Sessions for any given day, days in which Sessions peaked, did not have enough impact.

    To match GA's calculation in Domo, I added a Beast Mode which reads:
    sum(`Sessions` * `Pages / Session`) / sum(`Sessions`)
    This in the above example equates to 1.15 which finally matches Google Analytics.

  • Thank you Grant. Yes the granularity was tested in the Query Explorer, which resulted in no errors.

    The specifics of the request in the original post:
    Avg Session Duration, Bounce Rate, Sessions, Pages / Session

    Date, Source/Medium, Ad Content, Campaign