user measure google analytics

[jcabeza]

I have wrong data from Google Analytics to PowerBI, they are different than data in Google Analytics in "user" measure.

[EHas]

I'm running into the same problem. I think it may have to do with how GA counts users, it involves their client ID but as far as I can see that's not available when importing to Power BI.

[Anonymous]

Facing the same issue and understanding how GA works, I can provide an answer but not a solution.

The fact is that a user is not a summable metric against the days dimension as "users" are actually "unique users". That's because Users on day 1 + Users on day 2 will NOT be equal to Users on day1+2, since there are some users that have been visiting our site on day 1 AND day 2.
So imagine that on day 1 you have users

A, B, C

and on day 2 you have users

C, D, E

GA tracks users with cookies so it's able to recognize that user C on day 2 was the same on day 1.
Unique Users Day 1 = 3 (A, B, C)

Unique Users Day 2 = 3 (C, D, E)

but Unique Users Day 1 and 2 = 5, not 6 (A, B, C, D, E)

When you run the report for a certain time period in GA, it will count only uniques on the time period, But when you connect PowerBI to the GA and you get the users metric against and the time dimension, it will report you 3 users on day 1 and 3 users on day 2, hence you sum them and makes 6 not 5.

What could be the solution? Keep in mind that I haven't tested yet:  downloading the user metric but choosing only the week or month dimension instead of the day, so (probably) GA provides the uniques for the week or month, which you can show in your report when the period is week or day. But you really can't count unique users like you do on GA on any time range UNLESS you download each and every occurence of a users with it's "client ID" (GA unique ID), which I don't think it's possible.

[stiggrr87]

I thing this is correct. When you use different time periods from the GA then you will see the "unique" users for the each selected period. Eq. when monthly view then it matches with monthly view in GA page. 

In my opinion, the easiest way to get correct numbers for different periods is to extract separate tables. Daily unique then take the stats by days. But you cannot sum those days, because monthly stats is different then. Its all about how to explain the result then. 

[PaulDBrown]

This article by @michaelbilling on how the data is treated in GA is well worth a read:

https://www.scitylana.com/2017/09/07/why-google-analytics-connectors-are-useless/?utm_source=pbic&ut... 

here is another informative thread:

https://community.powerbi.com/t5/Desktop/Data-from-Google-Analytics-Doesn-t-Match/m-p/133680#M57135 

[Ram_DCT]

Hi Team,

Please help me out here, even I am facing the same issue with count of users not matching with actual GA users. 

Thanks,

Ram

[RolandKramer]

Hi there,

Same issue for me. This is what happens:

When I load 'Pageviews', 'Session', 'Bounce rate', etc. it looks all good. The values are identical to what is measured in Google Analytics.

But when I combine these data with the metric 'Users' or 'Number of sessions per user' all data is wrong. 

This tells me that all data is up-to-date, except the 'Users' metrics, which affects other metrics when loaded in the same sheet. 

Hope this extra info helps. 

[Ram_DCT]

Thank you Roland! Let's see if we get any attention from BI Dev team.

-Ram

[GilbertQ]

You have to remember that the Google Analytics is a cube. So when you are using measures across different dimensions this will appear to make the numbers incorrect.

But if you get them individually they will be correct. So one way around this is to bring in the common ones together. And then merge them in Power BI Desktop in the Query editor.

[RolandKramer]

Thanks for your answer.

I'm aware of combining metrics can cause numbers to appear different than in GA. However, combining Users with Pageviews, never caused any problem before and still doesn't for other domains I am reporting for.

It seems to me now that the particular problem I'm facing lies within GA. When I report Users over a longer timespan this affects the number of Users within GA as well. This problem appears at some point in the past.

So, as a result of that, a follow up question: is there a way to import data from GA from a certain moment in time? Now it imports all lifetime data, but I don't need all of this, and in this case this causes numbers to appear wrong. 

[BriannewtonO]

I'm encountering this issue as well. When comparing a week's User count in GA and Power BI, they don't match. However, when comparing them day by day, they match.

What I suspect is going on is in how GA calculates Unique Users. I'm pulling daily unique users from GA. However, if you just aggregate those numbers, you're not doing the same process GA is. When you pull a week's data of unique users from GA, it's actually going through a week's list of users and doing a distinct count based on a userID.

So essentially if a single user came to to the site 2 times on Monday, 3 times on Tuesday and 5 times on Thursday and you were doing a Monday through Sunday report where you're pulling daily data in Power BI, your Power BI would show 1 unique user on Monday, 1 on Tuesday and 1 on Thursday for a total of 3 for the week (Each day they were a unique user). However GA sould be able to recognize that was the same unique user across all three days, and would just count it as 1.

I don't see a way around this aside from pulling down a separate query for Weekly users (not daily), or running queries ad hoc based on the time frame.

[GilbertQ]

I would suggest making sure that you are using the identical measure from Google Analytics when getting the data. I am going to assume that the metric you have chosen is slightly different from the one shown in Google Analytics

[jcabeza]

Is the same measure in the same time period

[GilbertQ]

Interesting the only thing I could think of is there a delay between what is on the Google Analytics site and their cube?

Do you get matching numbers on a previous day or week?

[jcabeza]

The period to evaluate is May 2017. The delay is not a problem.

if I use the "session" measure is the same in Google Analytics than Power Bi. The problems is in "User" Measure

[GilbertQ]

I would suggest contacting Google with regards to this, because Power BI is just consuming what it receives from the API