Hi All,
I'm trying to build a segment that will allow me to look at repeat visitors to a specific page, let's call it "Page A". I'm trying to see repeat users for July 2024 and after. I want to have the first visit be on or after July 1st. Then I want to be able to look within a rolling 90 day window to see if a user repeated or not within 90 days. If it's like July 2 (example within 90 days), I don't want to look past July 1st. Hard cap it there. The issue here too is that I cannot look at month "greater than or equal to" Jul 2024. I have to select a distinct value. Also, how can I make sure to look at repeat visitors within 90 full days? I need to also look at this on a monthly level in workspace. Get the repeat visitors within 90 full days for July, August, September 2024, etc. Is this possible?
This is my attempt at it, but I don't think it's right. I also don't think it's right when I break out in a freeform table by the month dimension:
Thanks!
Solved! Go to Solution.
Views
Replies
Total Likes
Have you considered using a cohort table, then create a segment from there and modifying it?
Basically, if you create a simple cohort, something like this:
Using Page Views and a segment for your page.
Make sure you have a date range selected that will show some cohort columns...
Then you can right click the column to "create segment from cell", which will build out the nested containers... you should then be able to modify it to fit your 90 day time frame.
Have you considered using a cohort table, then create a segment from there and modifying it?
Basically, if you create a simple cohort, something like this:
Using Page Views and a segment for your page.
Make sure you have a date range selected that will show some cohort columns...
Then you can right click the column to "create segment from cell", which will build out the nested containers... you should then be able to modify it to fit your 90 day time frame.
Hi @Jennifer_Dungan,
This makes sense to me and I got this far. So my follow-up question is it sounds like I have to manually segment this out for each month that rolls by? In other words, I need to basically create segments for each 90 day look back window to say a user came in july or august or september and then a user came again in july august or september and I have to repeat this for each new monthly going forward? If so, once I have these segments, do I still need to filter down to the page level in my workspace and/or panel to be able to get the hits for a given month based on those metrics?
Views
Replies
Total Likes
You might be able to make a rolling date range?
This will also require setting the panel to "relative dates" and you will probably have to choose a single day (like "today" or "yesterday") at the panel level to get the correct 90 day "lookback"?
I'm not saying this will be easy....
Or we can try going back to your original solutions, looking for a sequence and within 90 days... and try to work through that logic together.
Hi @Jennifer_Dungan,
Originally I wanted to try "Page A in Visit" THEN WITHIN 90 DAYS "Page A in Visit". This is my segment logic for September to "Page A" for rolling 90 day (rolling 3 month and/or rolling back to July 2024).
I then use this segment to put over UVs to get a UV count, but then I need an overall just Hit segment to "Page A", then I break down in freeform by month to try to get the repeat users for that month based on these segments. I apply the same logic when brining in the other months on the dimensions for each segment to each month.
Let me know if this make sense and is the optimal way and best logic. It seems doable, just tricky and super manual every month I'll need to report this (I'll need to update segments for new month historical rolling 90 days/3months).
Views
Replies
Total Likes
So if I do something like this:
And I pair it with a time frame (let's say we are looking at the Last 12 Full Months)... this will look for a Visitor that that hit the home page, then hit the home page again within 90 days of that original hit... but that visitor will be returned in every month for that 12 month period (if they came to the site), regardless of if they hit the home page that month...
Example:
January
February
March
April
August
This visitor met the criteria in March>April, but they came to the site outside that range, and so, they will be counted as a UV in Jan, Feb, Mar, Apr and Aug; since the visitor as whole, has met the criteria... even if they didn't meet the criteria each month...
This also means that past months will change over time... yesterday you might see UVs as 158, and today you see 163, and tomorrow it might be 179.... because as new users start meeting the criteria (and they were active site users), they will be included in all those reports... and there's no easy way to only "include sequence"... you can include ALL users, you can include "Before" Sequence (which includes the sequence) and you can include "After" sequence (same as Before, this will include the sequence)....
So you could apply "only After sequence" to prevent past numbers from changing... but if the user hits the page in the first half of the year, do you still want them included after they stopped coming to the page??
This is some of the complex logic that I find just doesn't work well in the Segment builder... which is probably why the cohort table locks the dates in place...
Now, if you are only going to be looking at 90 Days at a time... maybe this is less of an issue....
Hi @Jennifer_Dungan,
I think that's why I have to also add another hit segment to make sure that the user actually saw the page when broken down by month. I think I just have to manually create a new segment as each new month approaches and just adjust the 3 months to be the previous 3 month (90 days).
Views
Replies
Total Likes