Hoping to get some direction.
I'm trying to figure out how many people come to a set of pages on my site from our partner sites. So, we have pages on our site related to an issue and we know that several partners' websites have links to these pages. We want to know what traffic their pages generate on our site.
Perhaps making things more difficult, we all have the same domain name, but we each have different sub-domain names (a.test.com, b.test.com, c.test.com, etc.).
I have tried a few different ways of assembling a report, but none of them provided what I was looking for. I have started with building a top pages report and then tried to filter by referrer to no avail and then went the other way, creating a list of referrers and tried to filter the pages they viewed, again to no avail.
I want to create a report that lists the referrers filtered by:
- referrers are: a.test.com, b.test.com, c.test.com
- page on our site they viewed was: a.html, b.html, c.html, etc.
Incidentally, when building a filter for an initial referrer, there's a dropdown list with the label, "-- Select an entry referrer --" It doesn't help my confusion when "entry" and "initial" are being used in the same context. All of the entries in the dropdown list are domain names, not page addresses, so I get confused also on what a referrer is - is it a referring domain or a referring page?" and the first item is "
To further muddy things, I don't quite get the difference between an entry referrer and an initial referrer, so any explanation as to the difference between them would be greatly appreciated.
Appreciate any time anyone can afford to help answer some of my questions.
Initial referrer is the referrer of the first visit of a visitor, carried forward onto each subsequent visit - eg:
Visit 1 Referrer: Google
Visit 1 Initial Referrer: Google
Visit 2 Referrer: None
Visit 2 Initial Referrer: Google
Visit 3 Referrer: facebook.com
Visit 3 Initial Referrer: Google
(and so on). It was a simple way to do a first-visit attribution method, but it only works on referrer and is generally quite pointless. Due to the confusion it causes I typically remove the report (unless the client really wants it)
Do you have all the sites (a,b,c etc) in a single profile or one profile per site?
Thanks for the explanation regarding initial/entry referrers, Bob. I wonder, if I'm not using cookies, initial referrer wouldn't be too accurate anyway (not that I can find any practical use for it to begin with).
No, I don't have these sites in my profile, since they aren't part of the sites I'm responsible for. The sites from which I'm trying to track traffic, are government sites and all have the same domain name structure (e.g. gc.ca), but each department has their own subdomain (dept_acronym.gc.ca).
So, when trying to track referrers from each department, I need to be able to track more than just the gc.ca, but the actual subdomain (e.g. abc123.gc.ca). Just to be clear, when I do an entry referrer report, one of the referrers is gc.ca, so my problem is trying to get one level deeper, filtering them by individual subdomain names. And then, the second problem is trying to figure out how to apply this action only to specific pages on my site, as opposed to my entire site.
Incidentally, I can see how this problem could also be extended to other subdomains. For example, in Canada, there are provincial subdomains of the .CA country code top level domain. For example, Quebec's would be qc.ca, and any organization that registers under that subdomain would have orgname.qc.ca.
Notwithstanding the referrer issue related to subdomains, I might be having difficulty in my logic as to the approach. I was originally creating referrer reports, trying to filter them to the specific referrers I was interested in, and then planned to filter by the specific pages on my site that I wanted measured. However, perhaps it makes more sense to start with a top page report, filter by the specific entry pages I want to measure, and then add the filtered entry referrers (if only I could filter them by subdomain). I don't know why one approach would be better than the other, but my dyslexia might have something to do with that. ;)
If you have a list of known referrers I might have a look under options -> marketing -> referrer groups and define a named referrer group for each of the subdomains that you care about - if you keep a consistent prefix/suffix then it would be trivial to create a filter to match it.
You can then use either the entry referrer breakdown dimension to get at the actual page on their site or create a new parameter to extract the referrer information again.
Some combination of the following:
Entry Referrer (filtered as per above)
Entry Referrer Breakdown (filtered with the above referrer filter would be good)
Entry page (where they land on your site
should do the trick.
I am a big fan of creating a full subdomain referrer parameter, as Bob suggested... a very useful one:
Associate with: View
Parse from: all pages
You may choose to exclude your own domains, either by applying a search and replace rule to drop unwanted values, or by using a negative lookahead in the regex for parsing. Or, of course, you can filter those out at the reporting level later.
Another option, if there are only a handful or dozens, may be to create custom metrics. For example, views filtered by view referrer, or visits filtered by referrer.