Monthly New users

[ryzen_]

Hi to everyone,

my client asked me to find the number of new users for each months in the last three months.

While at first, this problem seemed pretty easy, it turned out to be quite a mess.

I thought about approaching the problem using a Venn Diagram in order to visualize both the returning and new users, using three segment for each months defined in this way:

But it turns out that all the segment return a wrong number of users (almost 5M higher than expected).

Do you have any suggestions?

Thank you in advance!

[Jennifer_Dungan]

I did a Rock Star session on Non-Overlapping New and Repeat User Segment at Summit, (then did an encore for Skill Exchange) that might help you:

https://experienceleague.adobe.com/docs/events/the-skill-exchange-recordings/analytics/may2023/rocks...

[ryzen_]

Hi Jennifer,

thank you for your detailed video! I followed it and I tried to calculate new users in August 2023. I guess I did something wrong since I got 0 new users (I also activated "Make date range components relative to panel calendar" option)

This is the segment I created:

Could you help me? Thank you again!

[Jennifer_Dungan]

Hi... looks like a simple fix... in your first container you are using AND instead of OR

Days Since Last Visit is greater than 30

AND

Visit Number equals 1

---

should be

---

Days Since Last Visit is greater than 30

OR

Visit Number equals 1

---

That should fix the issue

[ryzen_]

It was just in front of my eyes ;( Thank you! I've tried different ways to calculate the new users, including your segment, but I still can't get my head around what's the most suitable for my client request:

Since the beginning of this year, for each months, give me the % of new users and % of returning users. 

• Create two segments (as suggested in this blog) using visits metric:

• Create two segments using visit number dimension (I saw you mentioned it in your video but I didn't understand what you mean for "all time dimension")
• Cohort table (as suggested in this blog), it doesn't completely respond to the question but it can gives an idea of the magnitude of returned/new visitor:

I put all this segments in a freeform table :

Thank you again for your help!

[Jennifer_Dungan · Accepted solution]

You're welcome

Visitor Segments can be the hardest to wrap one's head around.

It should also be noted that when I wrote those tips, the Visitor segments worked a bit differently (they pulled all-time visitor data, instead of limited to the panel date range... though no one will admit that change occurred, leading me to believe this was an uncaught bug since the beginning of the segment builder), those "time blocks" were needed to compensate for that "all time" processing.

So if you drop the date range, you shouldn't even need to use dynamic date ranges.

Here are versions of updated segments (there is still a slight variance when basing this on Months - because we are still looking at the "average" 30 day return window, instead of matching directly to that month, 31 days or 28 days for example - but as long as you say its based on 30 days in your description you should be fine)

New Visitors:

Returning Visitors:

Now that segments are taking date info from the Workspace, this means you don't have to force a date into the segment, and allows you to look at this for the current year.

[ryzen_]

Probably I need some time to realize how Visitor segment works (and Adobe Analytics generally!). Sounds scary there could be uncaught bugs ahah.

I have tried those segment on a new site that was launched online in June 23 so I expected that, for that month, there would be almost all new visitors. To my surprise, they are still way less than returned users

Am I missing something or it could be related to the segment that applied in order to isolate the visits on the new site? Thank you!

[Jennifer_Dungan]

Keep in mind, in my segments, if a new user comes to the site on June 28 and then returns on June 29, the user is treated as "Return" since within the 30 days they returned to the site...

If you want to defer to "new" then you need to reverse the exclusion....