This was clue that something was not quite right with:

1. their GA4 set-up

2. or Consent Mode

3. or that the Cookie Banner functionality or UX was deterring Users

4. or most Users blocked cookies in browser or used Brave or something else

5. or there was not enough traffic to enable behavioural modelling (which it looked like there was)

… or nothing was wrong with GA4 and 70% of all Users were opting out & Consent Mode V1 was deployed

How do you know how accurate GA4 is, if the main source of truth is GA4?

1. Look at Google Search Console (GSC) as this is Google’s own data on how many clicks your website received from google.com

   1. (*BOOM* went my brain!)

2. (or perhaps) Look at the stats from the CMP (which may be limited depending on the platform and the subscription)

3. Compare Meta Ad traffic landing page clicks to sessions (never has been reliable or useful)

4. Compare Google Ads clicks to sessions (I did not have access to PPC)

After checking the GTM container set-up there were definitely issues with how the triggers and consent mode was configured in conjunction with the CMP (cookie management platform) so I had to audit and explain this.

User Data Collection Acknowledgment (Acknowledged)

But in the meantime we also toggled a setting in GA4, yes that’s right, we accepted/acknowledged the “User Data Collection Acknowledgment” statement hiding at the bottom of the Admin > Data Collection page in the Property Settings.

And as mentioned in the google documentation, 7 days later this happened….

Total Traffic Increased by 160%

But this did not happen evenly across all Channels:

• Paid Social increased by a factor of 10!

  • it was basically hardly tracking any Meta traffic at all and some of it was being assigned to organic social

  • the pixel was also not tracking anything either (for other reasons)

• Organic Traffic doubled (not bad for just pressing a button)

• Paid Search tripled

• Referral and Direct doubled

So it’s definitely worth making sure the basics in GA4 are ticked off the list!

(possible) Implications…

So it was not obvious to me from the documentation what this setting does or means and tbh it’s still not totally transparent.

What I have learned however is that Consent Mode V2 + Data Acknowledgement is key to unlocking almost all user traffic to GA4 regardless of consent.

The “modelling date Vs observed data” does not so much look like any modelling is actually happening but rather it all becomes Observed under the guise of being “modelled”!

Which is interesting!

That’s It!

Fed up with cookiebot scans coming back to say consent is not fully active across your site from things like embedded YouTube videos?

Me too!

“Blocked until accepted by user: No”

“Prior consent fully enabled: no”

1. But I can only block scripts from GTM

2. These scripts are coming from YouTube and I can’t control them from GTM

3. I’m using gtm4wp and cookiebot doesn’t seem to like that

The explanation provided by cookiebot is linked below but the gtmwp plugin on WordPress was causing a headache.

Easy Steps to Follow

1. Install the cookiebot for WP plugin and connect it to your cookiebot account

   1. Modify some of the settings as needed to let this plugin manage placing the GTM snippet so that GTM is not blocked by auto-blocking feature

2. Modify the gtm4wp plugin settings so that it does not insert the GTM snippet again (as the cookiebot plugin is now doing this)

3. Pause the Cookiebot CMP Tag Template in GTM as this is no longer required (as the auto blocking settings of the cookiebot plugin is now doing this)

1. Install and Modify the Settings in the Cookiebot for WP Plugin

We want this plugin to take over from dropping the GTM snippet (which is currently done by GTM4WP).

1.1 Within General Settings, connect the plugin your cookiebot account and make sure cookie-blocking is set to automatic

1.2 Move along the the Google Tag Manager section and paste in your GTM container ID

1.3 In the Google Consent Mode Section, enable the features as they are in the Cookiebot CMP Tag Template that is currently in GTM

2. Modify the gtm4wp plugin settings

I was using some very useful dataLayer integrations provided from GTM4WP which I rely on so removing GTM4WP was not a realistic option. Changing the settings shown below removes the GTM code but retains the dataLayer integration. Once complete I could see that the dataLayer events from GTM Preview mode were being sent a bit later than previously but I was not using any critical ecommerce settings so it was not an issue. (I do not know if this would be an issue for ecom sites or not).

2.1 Go to the GTM4WP plugin settings and turn off the container code (eeek!)

2.2 In the Integration Tab enable the checkbox for Cookiebot auto blocking

This adds the:

data-cookieconsent="ignore" 

data attribute to any scripts coming from GTM4WP. This confused me greatly as the help link goes to GTM related article. What took a while to realise is that it also adds this attribute to any of the other integrations e.g. Contact Form 7 or WooCommerce.

2.3 In the same Integration tab, make sure the Google Consent Mode is not enabled from this setting

3. Pause the Cookiebot CMP Tag Template in GTM

Publish the container and save all the other plugin settings. Eek!

Final Result (from Cookiebot)

Well, in one sense, it is a massive let down!

The next scan results will be sent in an email and it will contain… no information about anything.
But the detailed report confirms all scripts auto-block until consent is given. Phew.

How Does This Look in the HTML?

Something like the screen grabs shown below…

some time lower down in the <head>

Useful Links

Remove Cookiebot Branding from your banner

cookiebot resource: Google Tag Manager and Automatic cookie blocking

cookiebot resource: Automatic cookie blocking

That’s It!

This is a total no brainer for all large Ecommerce Sites. (Almost as big as running an App campaign if you have an app store.)

GA4 uses its own AI to create predictive audiences of Users that fall into certain scenarios. The only restrictions are:

1. There is a mimimum volume theshold needed to make these work so they may not be available

2. They are available but they have not been created or added to campaigns yet (human error or incompetence).

Google’s AI Propensity Models

They should be a huge deal but I rarely hear about them being deployed. They use maching learning of your own Users and website data to predict:

• Likely seven-day purchasers

  • Users who are likely to make a purchase in the next seven days

• Likely first-time seven-day purchasers

  • Users who are likely to make their first purchase in the next seven days

• Likely seven-day churning users

  • Active users who are likely to not visit your property in the next seven days

• Likely seven-day churning purchasers

  • Purchasing users who are likely to not visit your property in the next seven days

• Predicted 28-day top spenders

  • Users who are predicted to generate the most revenue in the next 28 days

Default standard templates are found in Property Settings > Data Display > Audiences > Predictive (tab)

The probability and other options can be built out from these examples and tweaked to the data available in the account. e.g. LTV might be set-up so not a relevant rule to follow.

Available to be used for targeting in Google Ads

As like any audience built in GA4 it automatically becomes available in Google Ads for use in e.g. Remarketing Campaigns or on Observation for Search Campaigns.

The size of the audience in Google Ads will be smaller than the GA4 audience and the specific audience size per Google ad channel can be viewed in Google Ads or within GA4:

Advertising > Tools > Advertising segments

Being able to predict the future and then influence it - It’s like magic!

Other ways GA4 and Google Ads work together

Google’s guide to predictive audiences

That’s It!

Within Google Ads, the Landing Page report is not available for Performance Max campaigns. :(

But they are available in a custom report in GA4. So for any lead gen or shopping feed based PMax campaign it is possible to glean some insights not typically available. Even if Final URL expansion is not enabled, it is worth setting up this report. If Final URL expansion has been enabled then it is essential and this report can be used to setup exclusion URLs that should be blocked.

Surprising URLs to Exclude that you may not know about

For any PMax Shopping campaign I can guarantee at least 10% of revenue from PMax is actually coming from post purchase pages. What? Yes, PMax will find users who have just converted and send them back to e.g. their signed in Shopify profile thank you page!

As standard, I now always check and exclude some URLs for any new campaign launch.
e.g. If the campaign is for new customer acquisition then all pages behind a User login should be excluded. Don’t assume PMax will know to do this.

• Site Account pages and Post Purchase

  • /account/

  • /account/login

  • /account/logout

  • /order

  • /thank-you

• Checkout Pages (all or some of)

  • checkouts/etc

• Cart Page

  • cart/

  • would this repopulate with a signed in Users full cart history? If not, why is PMax sending users to an empty cart page!

• Footer type Pages

  • /policies

  • /pages

  • /about

  • /career

Suggested Metrics to Include in The Landing Page Report

• Views (pageviews)

• Views per…

• Revenue

• New Users

• Returning Users

• Sessions

• Key Events

For this website, all of these pages are only available to signed in Users and a Users account details or recent purchases. Some of these Landing Pages also contributed to revenue which it should not have done (due to an issue with the website).

This information can then be used to go back to the PMax campaign settings and Exclude this list of URLs from “expansion”.

Additional use cases for filters and groups for Ecommerce sites could be PDP (Product Description Pages) compared to PLP (Product Listing Pages) compared to Blog pages.

E.g. for Shopify sites:

• /collections

• /products

• /blogs

Perhaps PMax is sending a lot of traffic to a less than relevant blog post with zero revenue (or low engagement) that can now be added to the excluded URL list.

Adding to the Landing Page Filter for PMax only I often also say, Exclude any of the following pages with this Regex :

(.*collections.*)|(.*products.*)|(.*blogs.*)

It quickly reveals some interesting exclusion opportunities!

Percentage Of Performance Max Campaigns Running As Remarketing

Edit any Acquisition report and make sure Active Users, New Users and Returning Users are added as Metrics.
In the example screen grab below Remarketing activity within the Cross-network channel accounts for 36% of Users.

This is interesting but to know but hard to do much with directly in PMax as there is no direct mechanism to control this.

Actionable
It could be used in a test with an actual pure Remarketing campaign to compare side-by-side and trends for cannibalization.

Bonus: Get GA4 Metrics in Google Ad Reports

When GA4 and Google Ads are linked cross-sharing Audiences happens automatically with Auto-tagging and Google Signals enabled.

Within the Google Ads side of the connection you have to go back in and edit the link to also import GA4 metrics which will become avaiable within columns in Google Ads reports. These metrics are:

1. % of Engaged Sessions (the inverse of Bounce Rate)

2. Events/Session

3. Avg. Engagement Duration per session
   

See also a related post I wrote on using GA4 Audiences that use Google AI Propensity models for “Churning” and “Predicted to Purchase Soon” within Google Ads

Steps to create a Filtered Landing Page report from existing

Steps to Create a Cross-Network Comparison in GA4

That’s It!

1. Settings > Integrations

2. Settings > Setup

Select the Google Ads account to link and that is it almost configured….

Missing Sessions From Advertising Dashboard?

The setup documentation does mention having to ensure Google Ads campaign level utm data is attached for this integration to actually work properly but it’s not that obvious what to do about it.

As no advertiser relies on this for google ads tracking between Google Ads and GA4, chances are Google Ads has not been setup for this.

What UTM Parameters Does Clarity Accept From Google Ads?

Microsoft Clarity supports:

• utm_source

• utm_medium

• utm_campaign
  

This can be confirmed (and checked to see if this gets updated in the future) by navigation to the Filter section and looking at “Traffic” section. Referring site is included as a default browser parameter the others are defined from their respective UTM parameter. Hovering over Channel tooltip reveals that “UTM Channel: A friendly name for a grouping of UTM Source and UTM Medium”.

Note that if it does not have a Filter and so it does not accept any other UTM Parameters e.g. utm_content, utm_term, utm_id

For more information on when and where to use utm parameters on non-Google channels is in this linked post.

Update Google Ads Final URL Suffix

Tracking parameters can be added to Final URLs in Google Ads at Account, Campaign, Ad Group and Keyword level. As Clarity only accepts campaign data then we need to add this to Campaign URL tracking options in Campaign Settings > Additional Settings > Campaign URL Options > Final URL suffix (for every Campaign in the Account)

• utm_source=google&utm_medium=cpc&utm_campaign=manually write campaign name here (case sensitive)

Clarity should now be able to associate ad metrics with session data for further analysis.

Not Seeing Performance Max Campaigns in the Advertising Dashboard?

No, me neither. Assume this a “bug” that will be rectified by Microsoft at some point.

Assuming the GA4 integration is enabled then campaign name from Google Ads should be available from this side of the integration but not directly from the Google Ads integration.

That’s It!

1. Create a new Channel Group if you don’t already have one in use

2. Define the new Channel Name and Rules based on domain named like chatgpt.com

3. Re-order the new channel above the Referral channel so that it is prioritised and populated first

capture Gen AI Traffic From Referral Bucket

At the time of writing, Google haven’t updated any rules or created a new category for this traffic. By default it is categorised under Referral. We want to see how much Gen AI site traffic is growing in its own right and to compare against other Organic Traffic.

The screengrab below shows the Primary Channel Group capturing this traffic as Referral and Gen AI is about 60% of all Referral traffic on this site! I want to know about this without having to drill into Secondary Dimensions or create custom reports!

1. Create a new Channel Group

Make a copy of the default group if you haven’t already set this up.

In GA4 Navigate to Settings > Data Display > Channel Groups and then click on the 3 dots to “Create a Copy”

I like to name it “Detailed channel group” and update the description to “Copy of…”

2. Define the new Channel Name and Rules

Edit the new Channel Group by Adding an New Channel which I creatively named “Gen AI” with some Rules to capture this traffic:

Source matches Regex :

chatgpt\.com|perplexity\.ai|gemini\.google\.com|copilot\.microsoft\.com

This rule can be expanded as more Gen AI referrals come in over time.

3. Re-order the new channel above the Referral channel

The Channel’s order in the list is important as this is the order, from top to bottom, that the rules are applied and any traffic that matches the rule is bucketed into this channel. Think of it like a waterfall.

As the new Rule is more specific than the generic “Referral” rule as long as it is moved into a position above this then it will be good.

The View We End Up With in GA4

The result is available immediately in GA4 and it is applied to historical reporting. GA4 will generate the “User” Channel Group from the “Session” Channel group as shown in the screengrab below where I’ve navigated to the User Acquisition report and selected the new User Detailed Channel Grouping drop-down as the Primary Dimension.

In this wider context we can see that Gen AI traffic is only 2% of all traffic, Referral is slightly less and 75% comes from Organic (indicating resource is put mainly into onsite content with no link-building (referral links back to site).

That’s It!

Copy/Paste the code below (or from the link above) and add it to a custom HTML tag in GTM to fire on e.g. “Instantization - All Pages” and then publish the container.

<style type="text/css">
  a#CybotCookiebotDialogPoweredbyCybot,
  div#CybotCookiebotDialogPoweredByText {
    display: none;
  }

  #CookiebotWidget .CookiebotWidget-body .CookiebotWidget-main-logo {
    display: none; 
  }
</style>

Before and After Screenshot

HOw To Maximise Brand Logo

The options and information on how to add your own company logo is also not provided by Usercentrics. After some trial and error the best way to maximise your own branding is to use a very wide and short image. I’ve found that 1200 x 110 px image does a very good job of this. A square would be preferred but the result when saved is very disappointing from a brand visibility and user trust perspective.

• 1200px x 110px works the best to maximise brand logo image size

• avoid a white background to emphasize the logo as much as possible

• avoid 1:1 square image ratio as this is less than optimum and can look a bit silly it renders so small
  
  

THat’s It!

This will capture any relevant HubSpot form submission data at the point of a successful form submit button click and make it available in Tag Manager. From here this event or user provided data can be sent to any Marketing or Analytics Tags such as Google Ads Tag with Enhanced Conversions or a GA4 generate_lead Event.

1. Create a new custom html Tag that will listen for the form submission and add it to any pages with the form on it (or add it to all Pages)

2. Create some dataLayer variables to store the form data

   1. The listener tag will send the form data to the dataLayer along with a dataLayer event

3. Create any marketing or analytics tags that should be Triggered on this dataLayer event and make use of any of the form data contained in the dataLayer variables

1. Event Listener to be added to as a Custom HTML Tag on page load

This tag listens out for a “message” from the page that the form was submitted successfully so it needs to be fired on page-load so that it is ready and waiting to respond.
The firing trigger can be the specific page(s) the form is on or “All Pages”.
As it is not a marketing or analytics Tag it doesn’t need any consent as it doesn’t send any data (except to the dataLayer)

<script type="text/javascript">
     window.addEventListener("message", function(event) {
        if(event.data.type === 'hsFormCallback' && event.data.eventName === 'onFormSubmitted') {
           window.dataLayer.push({
              'event': 'hubspot-form-success', //New Trigger to fire marketing tags with
              'hs-form-id': event.data.data.formGuid, //Define as a dlv
              'hs-form-transaction-id': event.data.data.conversionId, //Define as a dlv
              'hs-form-field-email': event.data.data.submissionValues.email, //Define as a dlv
              'hs-form-fields': event.data.data.submissionValues //Define as a dlv. All other values from form
         });
        }
     });
</script>

2.1 Define the dataLayer Variables in GTM

The basic ones are exactly as written above e.g.

hs-form-id
hs-form-transaction-id
hs-form-field-email

A HubSpot form can have a lot of other fields so if there are any other being used then they will all be contained within the last dataLayer variable called “hs-form-fields” without the quotes. This will be an Object which will show up in Preview mode during testing.

2.2 Define any additional dataLayer Variables that are required

These will be shown in the hs-form-fields variable. Some common values would have the format e.g.
hs-form-fields.phone
hs-form-fields.firstname
hs-form-fields.lastname

3. Set-Up Any Marketing or Analytics Tags that Need to Use The Data

This could be a Conversion Tag for facebook or Google Ads or a custom Event in GA4. All these Tags would have a Trigger set to fire on the Custom Event ‘hubspot-form-success’

Custom GA4 Event Tag

This would show in GA4 as an Event named hubspot-form-success and would require new custom Dimensions to be defined named hs_form_id, etc for each of the Event Parameters shown

Purchase GA4 Event Tag

To start showing a “revenue” value in GA4 e-commerce reports then it’s possible to also send a GA4 ecom “purchase” Event that includes the required event parameters

copy/paste the code below into a new custom JavaScript variable (which I chose to call GA4 Items Array)

function(){

  var items = [{
  item_id : {{dlv - HubSpot Form - ID}},
  item_name : {{LUT - HubSpot FormID FriendlyName}},
  price : {{LUT - HubSpot FormID Values}}
}]

return items;

}

Google Ads Conversion Tag for a specific Form Submission

Assuming that a specific conversion for a “lead” or “contact” has been created in Google Ads e.g. for a Newsletter submission, then this Tag can be populated with User Provided Data values e.g. email address for Enhanced Conversions and should be set to Fire with a Trigger with an extra rule based on the specific form ID.

The Tag configuration above incorporates an additional condition on which to fire the tag. The values sent in the Tag back to Google Ads include the unique transaction ID (good for de-duping) and the users email address for enhanced conversions which is automatically hashed by Google. Even though we have defined some addititional custom variables that could populate “conversion value” and “currency” I prefer to leave these blank as these can be configured from within Google Ads by the marketing team. Likewise, the conversion count setting can be set in Google Ads, which should be set to “one” as opposed to “every”.

Meta Pixel Tag Set-Up

HubSpot GTM Container Import

Test the features described in this post by importing this custom container that contains all the required Tags, Variables and Triggers

Click on this link to the google file called “Hubspot Conainer Merge on Import to GTM.json” and “download” a copy of the file

1. In your own GTM Container go to Admin > Import Container

2. Import file into your own GTM workspace (using the “merge” import option)

3. Update any “dummy” or “example” values with your own

Useful Links Used While Writing This Post

Hubspot form event documentation

GA4 “purchase” Event format dev guide

Original inspiration of GTM recipe from Julius

HubSpot community solution

That’s it!

The “business intelligence” team tell the marketing department that mobile users who convert are less valuable to the business than Desktop Users.

Marketing goes and tries to up-weight budgets to desktop by reducing mobile ad spend and exposure through bid adjustments and notice:

• It’s not possible to down weight bidding on Search campaigns when using tROAS or tCPA. The only option is to exclude mobile by setting it to -100%

  • It may be possible to select say, -50% but this is ignored by these bid strategies

• There is no option to down weight mobile on Performance Max campaigns at all as this feature is not available

Reluctantly, Google provided a compromised feature to limit ad exposure to mobile through the Conversion Value Rules setting…

Conversion Value Rules in Google Ads

Conversion value rules allow marketing to override the best efforts of Google’s automated Smart bid strategies based on its own business intelligence. The 3 ways this can be done is by:

1. device

2. location

3. audience

These can be adjusted up or down by a specific amount (plus or minus) or by a multiplication factor.

Beyond reducing spend on mobile, this is also a practical way to signal to Google that certain audiences are less worthy and this is essential data to share for lead generation companies

How to Block Mobile from PMax Campaigns

As of Q2 2025 it is now possible to choose which devices PMax is enabled on as a campaign level setting. How to edit this setting is shown below

So, it’s not actually possible to exclude mobile from Performance Max campaigns but the bid strategy can down-weight the influence of users who convert on mobile by multiplying the conversion value of mobile devices by 0.5 which is the same as reducing its value by a half! Over time, the smart bid strategy will learn to prioritse desktop over mobile and that’s the best you can do.

A more drastic approach which would stop all YouTube advertising and reduce mobile is to completely remove all Ad Assets from any Asset Groups (for shopping campaigns). This turns the PMax campaign type into a standard shopping campaign with benefits, as signals do still work and can be used to help steer the campaign in addition to the product feed.

Ad Assets to remove:

1. Images

2. Videos

3. Logos

4. Headlines

5. Descriptions

How to Reduce Mobile Bidding on Google ADs?

Use conversion value rules

• The setting to do this can be found in “Tools & Settings” > Conversions > Value Rules (left hand menu)

• It is available at account or MCC level so it can be set across all campaign types (including Performance Max) across all accounts

• Specifically for Performance Max, it can also be set at campaign level within the campaign level settings for each Performance Max Campaign

-100% Device Bid Adjustment

It is possible to fully exclude Mobile or Tablet devices on the “standard” campaign types e.g. Search or Display but the way to find had been made deliberately convoluted by Google and its not a Campaign Setting…

Google Ads help on Bid Adjustments

Make sure the top-level Campaign Type Filter is filtered on a “standard” Campaign type then navigate:

LHS > Insights and reports > When and where ads showed

Main Screen > 3rd tab along > Devices

See also, this link on How to Block Mobile Apps from Display Ad Campaigns

That’s It!

Configuring this is really easy and when done it allows Organic Shopping Listings to be monitored in GA4 by reporting on metrics such as:

• Impressions

• Clicks

• Conversions

• Revenue

Where is the Auto-Tag Setting for Organic Shopping?

Navigate to your Google Merchant Centre > Cog Settings > Conversion Settings

Note, it is not where it may be more obvious which is in the “Linked Accounts” section!

GA4 Dimensions for Shopping Free Listings

The Source / Medium will not change as this will still only show “google / organic” but GA4’s Default Channel Groupings has been configured to understand the auto-tagged tracking ID and it allocates this traffic to “Organic Shopping”.

• Source Platform = “Shopping Free Listings”

• Default Channel Grouping = “Organic Shopping”

• Campaign = “Shopping Free Listings”

Monestisation Reports

Within the e-commerce purchase reports in GA4 a filter can then be applied to only show the specific items viewed, purchased, etc from the “Organic Shopping” value of the “first user default Channel Grouping” Dimension.

Or different channels can be compared such as “Cross-network” (Performance Max) with “Organic Shopping”

That’s It!

utm tracking parameters are essential to track, measure and compare non-Google channels e.g. Facebook or LinkedIn in Google Analytics. The benefits of consistently naming and applying UTMs with standard values is often massively overlooked by organisations, marketing departments and agencies. If utm tracking parameters are not added to every URL used in marketing campaigns then the true source of that traffic can be lost to Direct or Referral and in-platform metrics like cost, impressions and clicks cannot be imported.

GA4 comes with cross-channel attribution as standard, so the ability to compare channels side-by-side and end-to-end, from impression to conversion, is not possible without an agreed cross-team plan on how to build campaign URLs consistently by appending these parameters in advance of launching any non-google marketing activity.

TL;DR UTM Builder Sheet Template for Proper URL Tracking

“File > Make a copy” and use this “utm builder” google sheet that incorporates the advice of this post!

The ease of which is it possible to import non-Google in-platform metrics into GA is also dependent on building UTMs consistently. The best way to achieve this is to consistently name the source and medium while taking the campaign value directly from the channel e.g. from Facebook. Exporting this data from Facebook, segmented by day, along with the metrics: spend, clicks, impressions means it can be uploaded to Google Analytics without any further data manipulation.

• Google Analytics can then also generate from these metrics: ROAS, CPA, CTR, CPC

Cross Channel Advertising performance in GA4 reports on Ad Cost, Conversions, Cost per Conversion, Total Revenue and ROAS which can be segmented by the Dimensions of:

• Default Channel Group

• Source Platform

• Source

• Medium

• Campaign

Why Does Google Analytics Need UTM Parameters

Google Analytics relies on utm tracking parameters in order to identify marketing campaign traffic as this is not something it can understand without our help. By default, the only inbound sources that Google Analytics can “see” is referral url’s which get passed to it as standard by the browser. This allows Google Analytics to categorise the following “medium” / “source” value pairs

1. organic (with e.g. “google” as the source)

2. referral (with the “source” being the website domain that the user clicked on before landing on your site)

3. (none) / (direct) is the classification used when the browser doesn’t contain any information as to how the user landed on your site

   1. It may mean the user typed your exact domain into a browser or already had it bookmarked from a previous visit but that is an assumption

GA4 Default Channel Grouping

GA4 is much better about using the referral information and applying some additional rules in order to create meaningful buckets e.g. a referral from facebook.com, linkedin.com, instagram.com will the grouped under “Organic Social”

Overriding the Default Rules

To go any deeper about performance of a specific posting, placement, audience, paid link or full blown marketing campaign then the landing page URL used in the “paid” activity needs to include utm parameters. These override and overwrite the default referral data. The utm parameters are written as key-value pairs where the “value” is any text you like so being consistent is critical. utm_source and utm_medium are required and necessary as the minimum number of “keys”.

• utm_parameters must include at least source and medium (no longer a requirement in GA4)

• Do not add any parameters to any links on the website Google Analytics is installed on.

  • utms are only so you know how a user first found your website and are only relevant internally to the business for reporting

  • do not dictate utms from any links on your website to a different website

    • These should be provided by that website owner as it relates only to their google analytics reports

Auto-Tagging In Google Ads and Microsoft Ads

Google ads does not require any of this to be set up manually as there is an account level setting to enable “auto-tagging”. In addition, make sure the Google Ads and Google Analytics platforms are linked. Auto-tagging does a lot more than forgo the need to define utms manually including passing of data via gclids, defining GA4 reporting for “session google ads ad group name”. Note that there is no utm for “ad group” which is a massive anomaly.

Microsoft (Bing) Ads also has an account level option to enable auto-tagging but this literally adds all utm info it can to the ends of URLs and that is as much as this can do.

Available UTM Parameters

1. utm_source_platform e.g. SA360, DV360, Manual

2. utm_source * e.g. “facebook”

3. utm_medium * e.g. “paid-social”

4. utm_id ** (campaign ID) e.g. 123456

5. utm_campaign ** (name of the campaign)

6. utm_term (keyword)

   1. GA4 Dimension is “Manual term”

7. utm_content (something about the ad worth labelling)

   1. GA4 Dimension is “Manual ad content”

8. utm_creative_format *** e.g. "video”

9. utm_marketing_tactic *** e.g. “remarketing”

* Must be included (required) in the URL for Universal Analytics (There are no requirements to include any specific UTMs in GA4)

** utm_id and campaign are no longer required as part of the schema for importing Cost data to GA4 (but I would highly recommend including utm_campaign)

*** Can be included but not reported in GA4 (at time of writing)

*required for Importing Cost Data to GA4 and “use without quotes”

Valid Landing Page URL Format

1. Start with the URL as usual

2. If it already has a ? and some query parameters that the CMS needs then add on new utms with an &

3. If there is no ? then add a ? to denote the start of the utms

An example of this if I had created an Ad on LinkedIn the landing page URL could be: https://playhouse.digital?utm_source=linkedin&utm_medium=paid-social&utm_campaign=sme&utm_content=market-sizing-post

Build URL Parameters For Facebook Ads

There are two ways to do this:

1. Do nothing and Facebook will automatically add its own values for utm_source, medium, etc

   1. I’d advise against this as the values it has chosen by default are unlikely to align with the rest of the GA4 naming conventions unless of course doing nothing becomes the default way to manage this feature!

2. Follow a naming convention to send meaningful and user-friendly names for these values which can be easily updated and implemented with a UTM Builder sheet

1. Leave Blank for Facebook to Populate :(

It will auto-populate the following:

• utm_source=Facebook (or Instagram or the value of the dynamic insertion of “site source names” i.e. fb, ig, etc)

• utm_medium=paid (this would throw off all my other grouping for paid media activity)

• utm_campaign={{campaign.id}} (e.g. 123456)

• utm_content={{adset.id}} (e.g. 123456789)

• utm_term={{ad.id}} (e.g. abcd123)

It does not appear to populate utm_id with anything even though it uses the value of this for utm_campaign i.e. {{campaign.id}}

2. Create Dynamic UTM Codes that will auto-populate with more useful names :)

• {{site_source_name}}

  • this is e.g. facebook (fb), instragram (ig), messenger (msg), audience_network (an)

• {{placement}}

  • this is the ad placement/inventory e.g. it might be Instagram_Stories, Facebook_Right_Column or Facebook_Desktop_Feed

  • Facebook_Desktop_Feed

    Facebook_Instant_Articles

    Facebook_Instream_Video

    Facebook_Marketplace

    Facebook_Mobile_Feed

    Facebook_Right_Column

    Facebook_Stories

    Instagram_Explore

    Instagram_Explore_Grid_Home

    Instagram_Feed

    Instagram_Profile_Feed

    Instagram_Reels

    Instagram_Shop_Tab

    Instagram_Stories

    Messenger_Inbox

    Messenger_Stories

• {{campaign.name}}

  • actual name of the Campaign in FB

• {{adset.name}} (note: there is no direct equivalent of this as a UTM)

  • actual name of the Ad Set in FB

• {{ad.name}}

  • actual Ad Name

• {{campaign.id}} should be mapped to utm_id in the “Add parameter” section

When using the builder, multiple dynamic values can be stringed together with delimiters, for example a - or _

The value can also be filled in with any static text instead of in addition to the dynamic values.

Some Tips for Choosing Which utm codes to use with Facebook’s Dynamic Tracking Values

The best tip is to be consistent. Not just within a Campaign but across the whole Ad Account and for every Campaign activated year after year regardless of which person creates the campaign be it across multiple marketing agencies or internal marketing team members

For utm_source *, I tend to use a static value. If Facebook Ads is a new traffic source “meta” would be nice! If the precedence has already been set then stick with whatever this is (it’s probably “facebook”). Avoid using {{site_source_name}} as the dynamic insertion does not always work.

For utm_medium *, I tend to use “paid-social” but this could be “cpc” or something else, as long it is done consistently and makes sense alongside all the other acquisition channels. Avoid using {{site_source_name}} or {{placement}} as the dynamic insertion does not always work.

For utm_campaign *, I tend to string together the dynamic values from {{campaign.name}}_{{adset.name}} (even though dynamic values do not always work)

For utm_content, I tend to use at least the dynamic values from {{placement}}

For utm_id *, use {{campaign.id}}

Add utm_id to all Paid Channel Landing Page URLs

Importing Cost Data into GA4 requires utm_id, utm_campaign, utm_source and utm_medium so for Facebook this is collected via the {{campaign.id}} dynamic value

Once saved, this pop-up updates these fields in the correct format and appends them to the destination URL of the website. But there’s no need to stop here and only use these parameters.

• ?utm_source=meta&utm_medium=paid-social&utm_campaign={{campaign.name}}_{{adset.name}}&utm_content={{placement}}-{{ad.name}}

Unfortunately the Audience Network tracking does not always populate, even though there are values for this and if that happens all dynamic values associated with the Ad Placement also do not populate. The only way to ensure this is avoided is by not using the Audience Network. At Ad Set Level targeting, Advantage+ placements should be avoided.

UTM Tracking for Organic Social

Building URL parameters can and should also be factored into every organic social media post (that contains a link to your website). Although GA4s ability to more astutely separate out the different organic sources, it cannot know anything else than it came from LinkedIn. If you want to monitor on individual organic post performance in driving traffic and beyond then add some UTMs. The principle is exactly the same.

Campaign Naming Convention (from Platform to GA4 Report)

utm_campaign is the main key-value pair that can add instant value to all acquisition reports in Google Analytics. The easiest way to deliver this goldmine of data is to factor in the whole utm journey from “facebook campaign name” > appears in the web address of the browser > appears in GA4 reports

Facebook (or any other marketing channel) Campaign Name

The marketer decides this when they create a campaign in Facebook but what this is called will now become “stuck“ forever in an analytics report! Choose carefully and set it once. This is especially true for facebook as once a dynamic parameter “name” goes from Draft to Published the campaign “name” sticks even if it is subsequently changed within Business Manager. So if the “name” of a published campaign needs to change (and it hasn’t gone live yet) then make a copy first to re-draft it.

How Does the URL (Query) Look in the Browser

I used to avoid using spaces in campaign names as the browser would need to encode this query part of the URL by substituting the space for a + sign or %20. Facebook or the browser will do this automatically. (Technically, if there was a space in the /path/ part of the URL this would be substituted for a %20 sign and if it was in the query part it should use a + sign).

Facebook has an odd anomaly in that if, for example a campaign contains spaces and the ad placement is right hand column, it uses %20 in the query parameters but + for desktop feed ads

• RHC Ad

  • ?utm_source=meta_fb&utm_medium=paid-social&utm_content=Facebook_Right_Column&utm_campaign=New%20Traffic%20Campaign_New%20Traffic%20Ad%20Set

• Desktop Newsfeed Ad

  • ?utm_source=meta_fb&utm_medium=paid-social&utm_content=Facebook_Desktop_Feed&utm_campaign=New+Traffic+Campaign_New+Traffic+Ad+Set

Both the browser and GA don’t seem to mind/care too much about this and it rarely “breaks” these days but still, I’d rather be safe than sorry (if I can be bothered)

How Does It Look in Google Analytics

URLs and GA are both case sensitive. Dimensions such as Source / Medium in GA will reflect exactly how the utm values were written into the landing page URL. If a Microsoft Ads Campaign name is edited from “Proper Case” to “lower case” then this would create two rows of data and appear to be two different campaigns. This could be avoided by using absolute campaign IDs but it then makes the GA reports impossible to read in-situ. The campaign ID could be appended to the utm_campaign value but it would still be quite unsightly to read in a GA4 report.

A useful Universal Analytics filter to avoid this situation was to create a rule to “lowercase” all incoming UTMs but this is not a feature of GA4 (at the time of writing)

Character limits of UTMs

I’ve found contradictory advice about this from multiple sources e.g. there are different character limits per parameter which I can’t verify is the case. Some say:

• 100 characters (custom event parameter character count in GA4)

• 200 characters

• 255 characters for campaign names within campaign manager 360 (but this is something different)

• 500 characters

What seems most likely is that, generally speaking, max length of URLs should be less than 2000 characters (which includes query parameters).

There is also this google resource stating that GA4 event parameter value character limit is 100 characters. This does not apply to utm parameter values however as they are an exception to this limit.

Having tested this out with various utm_campaign parameter lengths I’ve found that 500 is not the upper limit so the main constraint appears to be the total URL length is less that 2000 characters.

Useful Links used while writing this Post

Facebook Specifications for URL dynamic parameters

Facebook automatically added URL parameters

[GA4] Traffic-source dimensions, manual tagging, and auto-tagging

[GA4] Import cost data

[GA4] URL builders: Collect campaign data with custom URLs

Google Analytics Campaign URL Builder

OG MeasureSchool Guide

character limits of GA4 event parameter values

That’s It

Thankfully I can still override what Google suggests is best practice with what I know to be better practice, without being overtly punished for it with a low Optimisation Score. If any recommendations are dismissed they will only stay quiet for about 4 weeks at which point they will have to be dismissed again (and again and again) to maintain a high Optimisation Score.

I’ve already covered some of the more common recommendations in the account level auto-apply recommendation article as there is some overlap but there are unique recommendations related to Optimisation Score that can’t be automated (which I have an opinion on too, so read on)!

If these recommendations are followed they only lead you down a path to more and more automation which assumes the inevitable outcome is improved performance

“Upload Customer Match lists”

“Show impactful and personalised ads to your existing customers by using the contact information that they share with you”

Being able to dismiss such recommendations in certain specific circumstances and it not having a negative impact on the account Optimisation Score is essential.

“Add broad match keywords”

“Get more conversions at a similar or better ROI by adding broad match versions of your existing keywords”

“Recommended because using broad match type keywords in these Smart bidding campaigns can help you effectively grow conversions within your performance targets “

A Longer explanation in the small print:

“Your Smart Bidding campaigns use a host of auction time signals to automatically deliver as many conversions as possible at or below the target or budget that you set. In combination with Smart Bidding, adding broad match versions of your existing keywords allows your campaigns to reach more customers who are likely to convert while ensuring that your performance targets are met.”

There are a number of caveats to consider before replacing all your keywords with broad match versions so doing this by “applying” this from the recommendations section would not be recommended! Google’s own recommended (internal) way to implement this is not by pressing “Apply”. Some pre-requisites include:

1. The campaign is using a Smart bid strategy of either target ROAS or target CPA. Max conversions or Max Conversion value is not a Smart Bidding strategy in this context

2. The campaign is not limited by budget and will never become so

3. Monthly conversion volume is 2x - 5x times the recommended minimum

4. The campaign generally meets its bid strategy target month after month

   1. implied by the wording “ensuring your performance targets are met” that the campaign is already hitting its targets before enabling

Bid more efficiently with Maximise conversion value using a target ROAS or Target CPA

“Get more conversion value at a similar ROAS with a value-based bidding strategy”

“Recommended because our simulations show your campaigns are likely to benefit from Maximise conversion value using a target ROAS, which can use data from all of your campaigns, including the conversion values that you’re reporting to optimise performance “

Context and the precise wording of the recommendation is important here:

“Set a target” recommendation is often seen on Campaigns using a “Maximise” Conversion or “Maximise” Conversion Value bid strategy. And at least the CPA recommendation specifically calls out the need to also “remain unconstrained by budget”. The same requirement is also true for the tROAS bid strategy but there’s no mention of staying unconstrained by budget. Following this recommendation can lead to a continually under performing campaign as Google has no recommendation to go back from a “Target” focused bid strategy to a “Maximise” bid strategy. In chronological order, the following will happen:

1. The Campaign currently uses a Maximise Conversion Value bid strategy

2. Google recommends using a tROAS bid strategy so you follow the recommendation

3. Google recommends increasing budget as the tROAS bid strategy is now limited by budget

   1. If you have more budget then you can do this and this is all good

4. If you have no additional budget you will get stuck in a sub-optimimal situation

5. One option is to increase tROAS which may result in the campaign becoming unconstrained by budget and achieving a higher ROAS

   1. The implication is that you’ve just sacrificed growth for efficiency

6. There is no other google recommended option to follow or even alert you that this has happened

I’ve outlined on the “Smart Bidding” resource page a better methodology on how to manage, transition and balance growth with efficiency using Smart Bid Strategies.

“Improve your responsive search ads”

“Get more clicks on your responsive search ads by improving your headlines and descriptions”

“Recommended because some of your responsive search ads have an ad strength below “Good””

The implication here is that an “average” and “good” ad strength is deemed “not good enough” and google will punish this with higher CPCs and / or lower Quality Score.
Getting an ad strength of “excellent” is based on :

1. Headline relevance to the keywords in the ad group +

2. fully populating all available ad real estate (so all 15 headlines and all 4 Description lines) +

3. Minimal but probably no “pinning” +

4. Would be helped by having one “Keyword Insertion” headline

All of which might help the ad be more relevant and generate more clicks but it will also:

1. Not be performing any business related or Smart Bidding objective like driving sales and improving conversion rates

2. Look totally bland and exactly like all the other Ads having no discernible creative differentiation

3. Be totally stuffed full of repetitive headlines all looking very similar

   1. It is very difficult to populate all 15 headlines (unspoken rule) in a well themed ad group with anything else but repitition

Enable automatically created assets

“Get more clicks on your responsive search ads by enabling automatically created assets for your Search campaigns”

Recommended because you have Search campaigns that have not enabled automatically created assets

More circuitous reasoning from Google; “we recommend doing this because you haven’t done it yet and we recommend that you do”.

This seems to combine the above Campaign level recommendation to “improve responsive search ads” with the auto-apply recommendation to “add responsive search ads”. Either way the reason not to do this is the same as already stated.

“Create a Performance Max campaign”

“Maximise performance for online sales, lead generation, shop goals or multiple objectives by unlocking all of Google's ad inventory”

“Recommended because this campaign type can help you maximise results for your goals while finding more converting customers across Google's available ad inventory

The Performance Max campaign uses your conversion goals and the best of Google Ads automation to serve ads on all Google properties, ensuring that you reach your customers at the right time with the right ad. You can set up campaigns to meet your online, in-store and lead generation goals.”

In the small print when hovering over the “?” it says; “You're seeing this recommendation because Performance Max is the easiest way to serve across all Google properties and maximise your performance with Google Ads.”

Interesting that Google are rolling out this recommendation for non feed based campaign types now. There was a very good reason why Google previously did not recommend using PMax for Lead Gen campaigns so things have changed and moved along as they have a tendency to do. Even so, I don’t think the recommendation of it being easy to serve across Google inventory is actually a very good reason, its just re-stating the main feature of Performance Max.

I have nothing against PMax and it can certainly be tested as a campaign “experiment” but there is quite a lot of work to do in setting this up including ensuring all creative assets are made deliberately by the Brand e.g. do not go live with PMax without a video asset.

That’s It!

Standard and Explore Reports in GA4

Exploration Reports are needed to do any sort of analysis that is not provided by the Standard Reports but Exploration Reporting data retention lookback window is restricted to the last 14 months worth of Event data. I explain how to increase the data retention period in this other post. There is no such time length restriction for Standard Reports. So in order to do any advanced reporting and look further back than 14 months one option is to make use of Big Query where any and all sorts of queries can be written. Although the ability to connect to Big Query is now possible the cloud configuration and use of bandwidth used to query it, is not.

CONFIGURE THE STANDARD REPORTING WITH “New” REPORTS AND STAY UNCONSTRAINED BY TIME-FRAMES

The “trick” is that all Standard Reports can also be edited and new reports can be added to the main reporting section within GA4. So these new reports can be customised to show the custom events required and they are not restricted to 14 months. It’s not a perfect solution as only fairly simple report templates are available such as:

1. Trendlines (by day but cannot be grouped by week or month or year)

2. bar charts and bubble charts

3. Tables, with secondary dimensions and multiple column metrics

Any of these reports can be exported, as normal, to a csv file from the “share this report” icon in the top right hand side of the screen. This isn’t perfect either as the “table” data is limited to the first 5,000 rows worth but the “trend line” downloaded data is also provided (for the single metric chosen).

CONFIGURE THE LEFT-HAND MENU REPORTING NAVIGATION FROM THE “LIBRARY” FOLDER

Admins can see a “Library” folder icon at the bottom of the left-hand navigation in GA4. Clicking on this “lifts the curtain” on how all the Standard reports have been built. New menu items can be added as needed and customised to specific requirements of the business. Perhaps there is a Marketing Collection, a C-Suite Collection a Dev-Ops Collection, etc. The Structure follows a logical folder hierarchy of:

• “Collection” is the drop-down collapsible menu item

  • Each sub-menu under this is contained within its “Topic”

    • Each Topic contains the custom reports

      • There can be a “special” type of report for Overview which provides a snapshot of the underlying reports or standalone report cards

Sidenote on the “Search Console” Collection

These reports are not available as standard in GA4. In order to see them in the Navigation do this:

1. Link Google Search Console to the GA4 Property (from the Admin Cog)

2. Click into the Library folder and “publish” the Search Console collection

   1. By default this Collection has already been created automatically but it is not published automatically

New Collection Reports and Insights

I’ve had a few GA4 Properties running since mid-2020 when it was called “Web+App” so I can demonstrate this works with some Event data that could be available as a Standard report but not quite in the form I needed. If you have had GA4 for less than 14 months then this proves that it can be done!

I updated the “Performance Marketing” Collection from the Library with a group of reports listed under the “Enriched Lead Gen Data” Topic. These reports show GA4 ecommerce data I’ve applied to Form Submissions via Google Tag Manager. There is also a report as shown above, where I monitor which “Call to Action” buttons are clicked on and what the click text was. Useful for testing CTA variations and their impact on engagement. Data for these Events started flowing in Nov 2020 and I can look back all the way back i.e. 30 months. I can then download this data as a trend, in aggregate and by some other secondary dimensions if needed.

So although it is good that this is possible and will provide some users with some reassurance, it’s still a very basic data export lacking any of the power that the Explorer Reports wield. Ultimately, a Big Query connection will be needed for most medium to large sized organisations.

That’s It!

My Auto-Apply Recommendations & Reasoning

How to Use “Bid More Efficiently….” Options in the “Grow Your Business” Section

The “Bidding” sub-section is a confusing area of this the auto-apply feature and I’ve had it explained to mean different things from different people within Google so it seems some work needs to be done to re-design the messaging around this. My take on it is this.

If all the campaigns in an account use, for example, “up to Max Conv Value” bid strategy then all the “lower” bid strategy boxes can be ticked. What is not clear is how this recommendation is then applied at campaign level and how it is made clear if a “lower” bid strategy is being over ruled by this setting/feature or not. I’ve tested this by enabling the boxes as shown below and then editing a campaign bid strategy to eCPC. The campaign allowed eCPC to be set and there was no messaging anywhere within Google Ads to say otherwise.

Setting to “Adjust your CPA or ROAS targets”

Hugely intriguing setting and not one as yet touted to any great extent by Google! This is an additional but separate bid strategy not available anywhere else in Google Ads i.e. it is not an option to select this as a campaign or portfolio setting. By way of additional explanation google says; “Get more conversion value (revenue) by adjusting your ROAS targets”.

What it isn’t

This is different from “Max Conv Value” which should be only used on campaigns that are limited by budget where the bid strategy will try to deliver as much revenue as possible for the available budget

It is also different from tROAS which is used on campaign not limited by budget AND where the business wants to control the limiting factor by having a minimum level of performance (by which ROAS is a proxy to achieve this).

What It Is

It is a flexible bid strategy that goes beyond the manually set tROAS bid strategy and automatically adjusts tROAS to bring in more revenue. It’s like “max conv value” for campaigns not limited by budget. It also implies that whatever the budget is then Google will try very hard to spend it to get one more dollar. It seems like this will always favour adjusting ROAS downward to drive more revenue.

How is this different from Max Conv Value?

What happens when this is enabled, what will it do and how is this communicated and reported?

I have no idea at this moment in time as there is no consistent advice or best practice for this setting coming out of Google that satisfactorily explains it!

That’s It!

1. It’s relatively straight forward

2. The App is owned and supported by Google so any issues and updates to the set-up are covered by Google and Shopify support

3. It automatically creates a number of conversion touch-points to monitor the full conversion funnel

4. For non Shopify plus stores, it generates conversion actions for steps along the checkout funnel which is not accessible with any other method including GTM

Pre-requisites

I’ve assumed that a fair amount of admin and housekeeping has already been done as this would be a very long post otherwise. You’ve already:

1. Signed up and created a Google Ads account and a Google Merchant Center account (GMC)

2. Installed the Google Sales Channel from Shopify. Do this from within your Store.

3. Linked your GMC and verified it with Google

Things to watch out for

1. Within Google Ads, the email address must be listed in the User section. Agency linked Manager accounts is not sufficient.

2. Make sure whoever is doing this, is doing so has edit access to all required platforms: Shopify, GMC and Google Ads. If you don’t see your Google Ad account in the drop-down section then review the email addresses and access levels. Do not “Create New” ad account if one already exists somewhere.

Ready to Connect

Press the green button to Connect the accounts. Google will send a connection request for Shopify to connect to your Store which you need to accept. This allows the Shopify API to create the Conversion Actions. Shopify will also send a Link request on your behalf from the GMC (which is already set-up) to the Google Ads which you can accept within Google Ads “Notifications”

User Journey & Conversion Funnel

This is great! “Out of the box” we can measure and monitor all website visitors and categorise them through the main stages of the marketing funnel. It’s worth pointing out though, that out of all these points, the main, sole, primary, must-have measurement action is (#1) “Google Shopping App Purchase”.

Edit the default Conversion Settings

Each Conversion Action can be clicked on and then edited. The default settings may be a bit too generous and require reigning in a bit. I’m far more “happy” than I used to be to leave these settings on a longer time-scale than limiting them, as ultimately what matters most is more data to power the bid strategies. Do I think a view-through conversion window of 7 days stands-up as a legitimate use-case? No. But I’m no longer measuring success by the attribution within Google Ads, and reporting back this credit is down to my prowess. Pragmatically, I want to provide as much good and relevant data as possible with the understanding it’s not to claim conversions for any given channel but to allow the channel to optimise as efficiently as possible.

Behind the Scenes

1. Linking the Shopify Google Sales Channel App to a Google Ads account allows one of Shopify’s “ Merchant Marketing Accounts” to create the 7 conversion actions

2. Shopify knows the standard templates used for all Stores and adds the conversion tracking code across all pages of your store

3. It also knows which pages are “special” points in the journey and can map them to the conversion funnel including the checkout funnel

4. Shopify can inject this code onto your Store without needing to mess about with your Liquid template files via Trekkie

5. If generating the conversion actions within the Google Ads account is enabled through the Manager link request is 20% of the solution then the Trekkie code is the other 80% that generates the variables, values and logic as when to send this data to Google Ads.

6. Trekkie is also used to perform similar tasks for other Sales Channels e.g. Facebook & Instagram

Extra Steps, Extra Insight

I also find it useful to add in at least one other conversion tracking method even if this is literally only for back-up. Once one or both of the back up options below are added, remember to go into the Google Ads Conversion section of Google Ads, and make sure their settings are updated from a “Primary” action to a “Secondary” action. This will ensure the conversions are not double-counted and used in any bid strategy. All 3 conversion actions can also be compared to see how they deviate from each other.

1. Link Google Analytics (GA4) to Google Ads and import the main ecommerce Transaction

2. Add a new Google Ads Conversion Tag to the thank_you page either directly on the page or via GTM

The place to add the code for the second option is on the Shopify Settings (Cog in bottom left hand nav) > Checkout > Order Status Page (scroll down) > “Add tracking scripts and other customizations” (paste code into this section).

If there is a Shopify or Trekkie outage of the “Google Shopping App Purchase” conversion then the Google Analytics transaction can be set to “Primary” until the issue is resolved. Read of this post if an outage has resulted in lost conversions so that this period in time is excluded from being factored into the bid strategy for any active campaigns.

Compare Google Ads Conversions to Google Analytics Transactions

Once Google Analytics is Linked to Google Ads and “Transaction” is imported as a Secondary conversion, any attribution differences between the measurement techniques can be quantified directly from Google Ads. To uncover these metrics

1. the “All” Conversions and “All” Conversion Value type metrics need to be added as new columns

2. Segment the main column by “Conversions” > “Conversion Action”

With the “All conv.” and “All conv. value” columns included and then segmented by “Conversion Action” we can see the additional conversions that are occurring for the other Shopping actions in the account. As they are “Secondary” they do not appear in the “Conversions” column. In the “Conversions” column we can see all the “Primary” conversion actions (the red boxes) that are being included and are feeding the bid strategy. These include:

• The Google Ads “Trekkie” Purchase called “Google Shopping App Purchase”

• Android App Purchases (sales from the Play Store App that have occurred (not on the Shopify Store))

• Android app Installs (although these have no monetary value). This Conversion Action could be used as part of an App Download Campaign

We can also see in the last row the Google Analytics Transactions that have been recorded during the same time period and that at 1,276 they are 522 less than the 1,798 the Google Ads Integration has collected. This will be as a result of attribution model and inference differences (google ads associates a conversion with when the click occurred (in the past) , GA with when the sale occurred (in the present) and possibly measurement blind spots due to cookie consent.

Shopping Funnel Volumes and Values

We can see how the Google Sales Channel App from Shopify exposes the journey to conversion within Google Ads reporting interface as “Conversions”. The purple boxes in the “All conv. value” is zero for general browsing behaviour such as a product page view. The numbers in the “All conv.” column show how as the path to conversion approaches the volumes decrease, as expected. For example, item views (#6) happen hundreds of thousands of times > Add to Cart (#4) happen 5.8k times > Begin Checkout (#3) happen 3k times and purchases (#1), 1.8k times.

Make use of additional columns & include “Results” & “Results Value”

These columns lift the hood in a more elegant way than using the “All Conv.” column segmented by “conversion action”, in order too see which “Primary” conversion actions are being used at an account and campaign level. “Secondary” conversion actions are excluded for better clarity. It breaks down which conversion actions are at play where:

• “Results” shows the number of “conversions”

• “Results value” shows the "revenue”
  

Monitoring checkout funnel is unique to this method (for non-Shopify plus Stores)

The checkout funnel is not normally accessible to standard Shopify Stores as this part of the process is managed by Shopify and does not happen on the owners domain. This includes: Information (Customer details) > Shipping > Payment > {{Purchase}}

• None of these steps can be measured outside of Shopify as there is no way to add tracking code (even Google Analytics) to measure them with the exception of the {{Purchase}} stage which can be tracked by adding custom code to the checkout settings

Unfortunately, the Conversion Actions that the link creates does not map directly to this Shopify checkout funnel:

• #3 “Google Shopping App Begin Checkout” directly correlates with the Information page of the funnel e.g. domain.myshopify.com/checkouts/c/aiohwh3874/information

  • but we could use a less reliable equivalent of this if it wasn’t available by tracking a click on the “checkout” button on the Cart page

• The Shopify checkout has a step for “Shipping” which does not have any conversion action equivalent in Google Ads

• #2 “Google Shopping App Add Payment Info” directly correlates with the Information page of the funnel e.g. domain.myshopify.com/checkouts/c/aiohwh3874/payment but (in this example at least) the data is inaccurate as the value is much smaller than actual purchases!

• #1 “Google Shopping App Purchase” would be available to other tracking solutions and marketing channels but only by adding “additional scripts” to the “Order Status page” via Checkout Settings which has its own gotcha’s to consider!

One More Step: Add Audience Segments in Google Ads

These full funnel Conversion Actions are used to drive the bid strategy based on your business objectives. This is going to be driven by the “purchase” conversion. If the set-up is new or sales are few and far between then a different lower funnel conversion action (e.g. Add to Cart) may need to be added into the mix to get the marketing campaigns off the ground and benefit from Google’s machine learning bidding intelligence.

These Conversion Actions are not Audience Segments and the Audience conversion funnel still needs to be defined within the Shared Library. Once created, these segments can be:

1. combined with other segments to create custom audiences

2. used to identify an existing customer so that Google can determine who new customers are

   1. which can then be used as part of the bid strategy

3. added to campaigns as observational audiences which would then be another signal for bidding!

It is highly likely that there will be previous “remarketing” audiences in your Google Ads account either from a Google Ads Tag or from Google Analytics and some of these may have been created automatically.

I’d highly recommend creating new audiences, based on “website visitors of a page with specific Tag Type” i.e. the conversion actions from Shopify.

That’s It

Cool, all done!

For context, the scenario

The website checkout and conversion page are still working. Customers can still buy things online. But for whatever reason, the Google Ads Conversion Tag temporarily stopped recording this data. Perhaps it took a few days or longer to fix and Campaigns remained on as they could still drive qualified traffic to your website but conversion data was lost. The issue has now been resolved.

Impact on Smart Bidding

The longer this has gone on, once resolved, the bigger the impact on campaign performance after the fix as :

1. The bid strategy will think it’s targeting signals need adjusting

2. It might incorrectly adjust Max CPC bids

resulting in increased costs and reduced revenues during the recovery period.

Tell the Bid Strategies to Ignore Bad Conversion Data

Use the “Data Exclusions” feature in Google Ads to let it know what days it should ignore so that it doesn’t act on bad data

Where to Find It

Tools & Settings (Spanner icon) > Shared Library (mega menu) > Bid Strategies > Advanced Controls (left hand menu) > Data Exclusion tab (Main window next to “Seasonality Adjustments”)

Option 1 -

Exclude all Mobile Apps Via Category Level Negatives for all campaigns at Account Level

1. Tools & Settings (spanner icon) > “Setup” (Mega menu) > Content Suitability

2. Then under “Advanced Settings” drop down, expand the “Excluded Placements” and “App categories” drop down

3. Tick every category (and sub category) and “save”

Up to 2018, all mobile app placements on the Display Network could easily be excluded by adding “appsformobileapps.com” as an excluded domain. Sometime after google removed this option, as they often do, it reappeared in its current slightly more unhelpful format. (See Option 3 below and comments for another take on this.)

Re-enabling some relevant App Categories is worth testing after perhaps a month and then test for another month with these categories enabled. Individual Apps can also be excluded based on performance metrics. I’m a great believer in automation and letting Google’s machine learning do its job but from experience it does a very bad job at learning enough and having “good” data to make good decisions. Lacking good data it also seems to be quite happy to carry on serving impressions and dodgy clicks across all sorts of terrible apps of which there is too many to block manually and individually.

Alternatively : Exclude App Categories Manually at Campaign Level (and optionally apply to Account)

First of all, on the left hand menu at campaign level, make sure you are viewing “All Campaigns” or Display / Video campaign types. As this is not applicable to Search and Performance Max they do not have this option in the secondary vertical menu.

1. All Campaigns (Primary left hand side vertical Menu)

2. Audiences, keywords and content (Secondary left hand side vertical Menu), (drop down) > Content

3. Placement Exclusions (main window)

4. Edit with the Pencil icon (main window)

This can be selected for all 3 mobile App Stores, Android Play, iphone Apple App Store (and Windows Phone Apps appears at campaign level but not at account level).

Option 2 -

Exclude all Mobile Apps for all or Some campaigns with Google Ads Editor

Watch the video below to see how to do this.

Related to this: How to exclude mobile from google ad campaigns

Option 3 -

Exclude All Apps Using ID for “All Apps” mobileappcategory::69500

This only seems to work at campaign level. When I try this method at Account level, there is an error message when clicking “save”.
Rather than using the “Browse” window, go to the adjacent “Enter” window and type mobileappcategory::69500

“Add 1 placement” & “Save” as you do this it will update the ID from mobileappcategory::69500 and show this as “App categories > All Apps”

That’s It!

Filter functionality in GA4 is limited but it can be done based on IP Address and GTM Debug Mode which is called Developer Traffic

All Filters should be “Tested” for a period before being made “Active” to ensure the Filter is working as expected. While it is in the “Testing” state

Filter Modes

A filter has one of three modes:

• Testing: Analytics is evaluating the filter but not applying permanent changes, and matching data is identified with the following dimension name and dimension value:

  • Dimension name: Test data filter name

  • Dimension value: <the name of the data filter>

• Active: Analytics is evaluating the filter and applying permanent changes.

• Inactive: Analytics is not evaluating the filter.

Where to Find It

Cog (Admin) > Data Settings > Data Filters > Create Filter (button)