Analytics
3.3K members online now
Understand information in your reports and troubleshoot reporting issues such as self-referrals, (not set) data, and inaccurate information
 
Guide Me
star_border
Reply

The best way to determine how many times users visit a site?

Visitor ✭ ✭ ✭
# 1
Visitor ✭ ✭ ✭

Hi, I am wondering what is the best way to determine how many  times users visit a site.

 

Sessions / Users over a week gives me 1.74 - So I am tempted to say users do an average of 1.74 session during the week.

 

But...

 

If I do Sessions / Users over a day, I get 1.2 - I have difficulties to get my mind around it: if users do an average of 1.2 sessions during a day (and each day of the week), I should have a lot more that 1.74 sessions for a week, shouldn't I?

1 Expert replyverified_user

The best way to determine how many times users visit a site?

Top Contributor
# 2
Top Contributor

Hi,

 

I'd suggest you to check the User Explorer report.

 

https://support.google.com/analytics/answer/6339208?hl=en

 

Let me know if this helps.


Arnold Majlath, Google Analytics Top Contributor
Was my response helpful? If yes, please accept it as solution.
Circle Me On Google+ | Visit me @ Redfly Digital

The best way to determine how many times users visit a site?

Explorer ✭ ☆ ☆
# 3
Explorer ✭ ☆ ☆

Hi Benjamin,

 

I think the issue is you're comparing rates instead of real numbers. Averaging daily Sessions / Users doesn't take into account the weight of each day (variance in users and sessions day to day). 

 

Create a custom report with Date as the primary dimension and Number of Sessions Per User, Sessions, and Users as metrics. Set the date range to the week you saw 1.74 Sessions / Users and it should be pretty self explanatory from there. 

 

Best,

Mike

The best way to determine how many times users visit a site?

Visitor ✭ ✭ ✭
# 4
Visitor ✭ ✭ ✭

Thank you for your answer. I have done the custom report. It gives me the average per day and 1.26 for the week. I have the feeling that GA does an average of daily averages instead of calculating the average amount of sessions/user for the week. I am beginning to understand that it probably can't work from averages :-\

The best way to determine how many times users visit a site?

Visitor ✭ ✭ ✭
# 5
Visitor ✭ ✭ ✭

Thank you for your answer. Indeed, after reading the doc and doing some tests, it seems to make more sense. I have an average of 3 sessions per day and 9 for the week. Of course, it is difficult to get those numbers as it seems that I hit a row limit but that's a beginning.

The best way to determine how many times users visit a site?

Explorer ✭ ☆ ☆
# 6
Explorer ✭ ☆ ☆

Hi Benjamin,

 

The difference between the 1.74 and 1.26 is because of how the "users" part of the equation is calculated - the custom report adds daily users together while Audience Overview is deduped. If your goal is to figure out average sessions per user from that week, I would go with the 1.74 figure assuming that came from the Sessions and Users totals in the Audience Overview report. 

 

The reason you're getting 1.26 from the custom report is because GA added up all of the users from each day. Summing users across days inflates the metric as users returning on different days are counted multiple times when in reality they are just one user. Inflating users increases the denominator of Sessions / Users, which is why it decreased from 1.74 to 1.26 as soon as you added the daily component. 

 

tldr: To get the most accurate Sessions per User metric, set date range and divide total Sessions by total Users. 

 

Best,

Mike

The best way to determine how many times users visit a site?

Visitor ✭ ✭ ✭
# 7
Visitor ✭ ✭ ✭

Thank you for this detailed explanation. I understand the calculations but the result surprises me: if I have an average of 1.x per day, how can I have 1.74 for the week? Shouldn't it be greater? Ot is it telling me that users are simply not coming back in the following days (ie. coming only one day)?

Marked as Best Answer.
Solution
Accepted by topic author Benjamin L
November 2016

The best way to determine how many times users visit a site?

Explorer ✭ ☆ ☆
# 8
Explorer ✭ ☆ ☆

That's correct - it's not all the same users. You can use the cohort analysis to see what percentage of new visitors return to your site over a certain period. For example, 100 visitors saw your site for the first time on 11/20/16. Cohort analysis can tell you what % of them came back on 11/21/16, 11/22/16, etc. 

 

User related metrics such as Sessions / Users will always be higher when you increase the date range so long as you have some returning visitors. 

 

Here's 2 extreme scenarios to demonstrate:

 

Scenario 1:

Your website gets 100 users and 120 sessions a day for a week. The 100 daily users are the exact same users for each day. Sessions / User a week would be (120 * 7) / 100, giving you a figure of 8.4 sessions per user for the week. 

 

Scenario 2: 

Your website gets 100 users and 120 sessions a day for a week. The 100 daily users are new each day - no returning users = no returning visits. Sessions / User for the week would be (120 * 7) / (100 * 7), giving you the same as if you looked at the metric daily. 

 

Your case of 1.74 weekly vs 1.26 daily is somewhere in the middle (closer to scenario 2). 

The best way to determine how many times users visit a site?

Visitor ✭ ✭ ✭
# 9
Visitor ✭ ✭ ✭

Thank you very for this explanation :-)

The best way to determine how many times users visit a site?

Visitor ✭ ✭ ✭
# 10
Visitor ✭ ✭ ✭

Hey Mike, thank you for the answer above. Super clear, but is there a way to see how many unique visitors are returning? Is this just the returning visitor metric? Is it really that reliable as our stats are showing a 90% returning user stat and i am not sure that is very accurate.