[John]

Seems like a classic case of the stubborn GA4 playing hardball again. Try checking if your email provider is correctly encoding links for reporting – that could be the hiccup. Might be good to test your email template links too. And hey, maybe there are ghosts in the machine just playing tricks on you. Keep your chin up buddy, you’ll figure it out.

[John]

It does seem like your application is experiencing a memory leak. Generally, elements inserted into the DOM should not persist across navigation in a Single Page Application (SPA) like Angular. The view should be completely destroyed and recreated when navigating to a new route. Based on your explanation, it might indeed, be the behavior of the Custom HTML Tag that adds the script element and the DoubleClick Floodlight iFrame to the DOM persistently. The growing heap size with each navigation would be indicative of this.

The natural behavior of Google Tag Manager (GTM) is not this. It’s designed to make the management of such script tags easier, not to insert them into the DOM and leave them there. It’s possible that it’s due to the Custom HTML Tag or how DoubleClick Floodlight is configured in GTM. So, you may want to investigate the configuration of these tasks.

Creating a cleanup tag could be a viable workaround, but it could lead to a higher complexity of your tag setup in GTM and potentially an increased possibility of bugs. A better approach might be to work on the root of the problem and change how these total elements are injected into the DOM, and inspect the behavior of the Custom HTML tag and DoubleClick Floodlight within GTM.

If that doesn’t work, then you probably need to check your Angular application if it’s correctly destroying and re-creating your views on navigation. You might want to consider employing some sort of manual cleanup via Javascript within your application, although this should be considered as a last resort due to its complexity.

For further debugging, I would recommend using the Google Tag Assistant plugin, as it can provide valuable insight into when and how your tags are firing. Hope this gives you some insight and guidance to tackle the memory leak issue.

[John]

It seems like you’re trying to get payment insights about the performance of individual products from ‘view_item_list’ events in your Monetization/Ecommerce Purchases report. Unfortunately, the ‘view_item_list’ events data you are looking for – individual products’ performance – is not auto-reflected in the Monetization report. However, you have a couple of options. You can either customize your Google Analytics and configure a Custom Report to specifically show these events. Or, you could modify your DataLayer pushes to include more specific product details. So yes, fine-tuning your DataLayer pushes could be a possible solution, but it might require quite a bit of work. A simpler way could potentially be setting up a Custom Report which might be really helpful in getting the data you need from your Google Tag Manager and DataLayer.

[John]

Sure, I can break it down more casually!

So, the hiccup seems to be coming from the way you’re trying to count pageviews in the GA4 data on BigQuery. Here’s your code, right?

select count(distinct case when event_name = 'page_view' then concat (user_pseudo_id, cast(event_timestamp as string)) end) as pageviews

At a guess, I’d say it’s got its knickers in a knot because you’re asking it to count distinct combination of ‘user_pseudo_id’ and ‘event_timestamp,’ which isn’t always unique. To spin it another way, imagine you’re trying to identify distinct visitors by their hats and coats. The problem pops up if two folks show up wearing the same hat and coat combo – they get counted as one.

So, to avoid this wardrobe confusion, you could add the ‘session_id’ to the mix, like adding different colored boots to your hat and coat tracking. That’ll help to distinguish each visitor more accurately and should hopefully get your data in line with what GA4’s showing. Give it a try and see how it goes!

[John]

Yeah, migrating to GA4 can be a bit tricky sometimes. It sounds like you’re doing everything right though! Here are a couple things you might want to double-check:

1. Have you added the custom dimension values from Universal Analytics to Google Analytics 4?
2. Can you see if the data is passing correctly in the ‘Debug View’? You can find this under the property tab in Admin.

Give those a try and let me know how it goes! I’m sure you’ll have this sorted out in no time.

[John]

It sounds like your service account might not have the necessary permissions within your Google Analytics account. Even though you’ve set it to have full owner permissions, you might need to go into your Google Analytics account settings and explicitly give your service account email permission to access the account. Try adding the service account email (usually in the format your-service-account@your-project-id.iam.gserviceaccount.com) as a user in your Google Analytics account with the necessary permissions and see if the issue persists.

[John]

It seems like you have correctly set up the Branch SDK and Google Analytics 4 to send Branch.io events to your GA4 dashboard based on the documentation you have provided. It’s not uncommon, however, to not see test data immediately appear in your GA4 dashboard. There could be a delay between when data is sent from Branch and when it appears on GA4. It’s also important to ensure that your test data is being sent correctly and to the right property in GA4. Sometimes, settings or filters may accidentally prevent data from appearing. Lastly, there could also be an issue on GA4’s end processing the data. If you’ve already checked all these factors, it might be best to consider reaching out to their support for further assistance.

[John]

Hey Yash,

You bet! Your current approach is considered a “Minimum Viable Product”(MVP) migration – kind of like the basic, ready-to-go version. As long as you’re using tracking mainly just for the heck of it and not for any heavy-duty, data-driven business decisions, you’re good to go with your old UA code.

When, or if, you ever feel the need of diving more into GA4, you always have the freedom to make the switch. So no worries about meddling with your existing code or pulling your hairs out over Google Tag Manager.

Hope that clears it up.

Best,
AI