Profile cover photo
Profile photo
Lindsey Buckle
172 followers -
Love my kids, love my husband, love my family and friends, my job (digital marketing analysis), my home and the place I live. Love my life.
Love my kids, love my husband, love my family and friends, my job (digital marketing analysis), my home and the place I live. Love my life.

172 followers
About
Lindsey's posts

Post has attachment
I have a best practice question.

We are currently in the process of implementing a new (replacement) website. The website CMS will be Sitecore with ADX forms embedded in iframes (sigh). What is the best method for ensuring accurate tracking of the forms? I've been reading a fair bit about this and I'm just trying to get it straight in my head. AFAIK the forms will be on the same domain as the pages.

We'll need to track form submissions, but also (probably) engagement, abandonment and possibly timing; the usual suspects.

We'll also need to ensure that we don't inflate the pageviews (naturally).

So should I put the tracking code on both the parent and the child frame, with an exception on the pageview tag (as per Simo's article: https://www.simoahava.com/gtm-tips/prevent-tag-from-firing-in-iframe/). And then just pick up all of the other event tracking/form submission tracking as normal?

I read about the postMessage method in this article (http://www.lunametrics.com/blog/2015/10/21/google-analytics-iframes-form-submissions/) but I'm thinking that I probably don't need it as I should be able to push straight to GTM from the iframe. (Assuming we can access code in the iFrame; I have no experience of ADX and I'm waiting for developers to confirm.)




Does anyone know why Tag Assistant might give me different results when I'm logged in as different Chrome users?

Same public website. Looks like my Tag Assistant settings are the same. In one browser, both the GA and GTM tag have happy faces. In the other browser the GA tag is showing a "non-standard implementation" error.

Is it possible another extension is interfering with Tag Assistant?

Post has attachment
Can anyone solve this mystery for me?

When I segment my data by sessions containing a specific event I get 31,313 users. When I add a second session, segmenting by users that have sessions with the same event, the Users metric for the first segment adjusts, as do my All Users metrics. See the screenshots. The time period is a year so could it be a sampling issue? And if so, how can we trust data when applying mulitple segments? (The two segments are exactly the same but one is session-based and the other is user-based.)

Is there any other way to analyse this data over this time period? I have PowerBI but still learning so I'm wondering if that is a better option than GA UI?
Photo
Photo
9/9/16
2 Photos - View album

Quick question.

I have a client with cross-domain tracking set up to a booking engine. Using tag assistant I can see that the booking engine is sending data to my client's GA. What is confusing is that the booking engine pages are showing up as landing pages (source/medium = direct/none). I wondered if it could be from sessions timing out but there is a high number of New Users.

About 20% of all sessions are starting from a booking engine page and together these have an average of 81% New Users (higher than site average of 74%).

Any ideas why this could be happening or how I might investigate further? I can't help but feel like I'm missing something really obvious here.

Quick question re. Data Import.

Does anyone know whether I have to upload my data file again following a change of views in the data set?

I created my data set to apply only to my test views then uploaded the data. I'm happy that it's working and have now extended the data set to all relevant views. Just wondering whether I now need to upload the data again. I will anyway, for expediency, but it would be good to know whether this is actually necessary for future reference.

TIA

I have a cross domain tracking question.

tldr: How can I remove a certain querystring parameter from one sub-domain but not another in a cross-domain view, preferably using view filters?

So I have two websites (e.g. www.myco.com and www.portal.myco.com), each with their own property and also tracking to an aggregate (cross) property. One of the sites is a customer portal that users can log into. Sometimes an account number appears in the url, either embedded (e.g. /content/1234567/subcontent) or as a parameter (e.g. content?id=1234567). This results in duplicate rows in the All Pages report and complicates aggregate data. So, in GTM I grab the id from either the url or the parameter and stick it in a custom dimension, then remove the embedded id. In GA I have added a filter to remove query parameters. I also add the full page path (prior to clean up and param removal) to another custom dimension, just in case.

So far so good. Except, when I do this in my cross property, it will also remove the id param from the main site, where id is often used to identify content (e.g. news?id=26). So ideally, in these views I only want to remove the id param from the portal hostname, not the other one.

Can I do this with filters?

My other options are to
a) stick id into custom dimension on main site too. I'd rather not do this as they are genuinely different pages. But I will if I have to.
b) rename the portal id param in GTM when sending the data through. I am already rewriting the page path anyway but due to the backend system I'm not entirely sure this will work for querystring parameters.

Any other simpler ideas that might be quick and easy to implement?

TIA, clever people.

Post has attachment
What's the best GTM plugin for Wordpress? I was going to use DuracellTomi's but then I saw this post. What gives? What does everyone else use?

Why are Unique Pageviews less than Users?

Tl;dr:
I created a filter ga:previousPagePath=~pagesA;ga:pagePath=~pageB (where pagesA is actually a group of pages). The number of Unique Pageviews returned (68) was less than than the number of users (113). How?

Background and Detail:
I want to return (via API) the number of sessions where any page within a set of pages (pagesA) is followed (at any point) by page B.

Ultimately what I think I need is a segment which is set up as follows:
Advanced>Sequences
Filter: Include: Sessions Sequence Start: Any user interaction
STEP 1: Page matches regex pagesA
Is followed by
STEP 2: Page matches regex pageB

However, in trying to get my head around segments versus filters; and sessions/pages/unique pageviews etc, I ran a test to see what would happen if I compared the results of a similar segment (but with “immediately followed by”) with a filter that looks at previous page. I figured this was pretty much the same thing, but clearly not.

The results were mostly what I expected. Both methods returned the same number of users (113). The segment returned 114 sessions whereas the filter returned 0 (because none of the pages I’m looking at were landing pages). However, they differed vastly with the unique pageviews. The segment returned 541, which makes sense to me as it would count ALL pages visited in sessions that also included group A followed by B.

What didn’t make sense to me is why the filter method returned only 68 pageviews, which is less than the number of users. Can anyone explain this to me?

And whilst you’re at it can you also confirm that it is safe for me to use a segment that counts sessions in this situation, even though we’re using hit-level dimensions (pages) to filter them?

TIA.


Does anyone have any call tracking data on how many people call after visiting a website on their mobile device, versus number of people calling after visiting on a desktop?

I realise this will change industry to industry and scenario to scenario, but I'm just after a rough idea. We can implement click-to-call for mobile and that may work for some desktop users but we're unable to change the phone number and track usage that way. So, I have to assume that there will be a number of desktop users who will call but will remain untracked. I can observe the mobile trend and apply it to desktop but I reckon that a smaller percentage of desktop users will pick up the phone because it's not as convenient. I could guess this number would be around 50% (i.e. if 50% of mobile users click-to-call then only 25% of desktop users might bother) but it would be great to base this on real figures if possible.

Post has attachment
Strange blank events showing in my Events report.

The website in question has old GA async tracking code on the page and there may be some legacy event tracking embedded around the place too. More recently GTM has been employed for event tracking. I have noticed a really strange event showing up in the reports. It has no Category or Action and disappears from the report when I add a secondary dimension of either Label or Page. It shows a reasonable number of unique events but 0 total events.

1) How is this possible?
2) What could be causing it? Some kind of spam hack injecting events into our GA? (We use a -2 property ID, which should reduce the chances of spam and I haven't noticed spammy URLs like I have with our other site)
3) How can I fix it? I'm thinking of adding a Custom filter to Exclude Event Category (leaving Filter Pattern blank) - would this work?
Photo
Photo
5/19/16
2 Photos - View album
Wait while more posts are being loaded