GA Shifts from UA to GA4, What This Means for Audience Targeting

April 11, 2022 | Tune in here!

Google loves changing things. There’s always a name change, product improvement or the most annoying thing - a UI overhaul. I was barely a year into my digital marketing career when Google Ads introduced a new UI and I had to relearn the platform in the next 6 months or so. 2017 was quite the year for learning Google Ads. I remember clients talking to me about it and hating the new UI. I sympathize. I am a creature of habit because I want my time to be spent on things that are additive. I want to be able to form a baseline and build upon that. Nowadays though, Google introduces new features with new UIs so at least that’s something.

A few weeks ago on March 16, Google announced that GA4 would replace Universal Analytics. The world was lit.

Marketers have been complaining that GA4 is so unintuitive and that the user experience is just terrible. Search Engine Journal even wrote a compilation of negative feedback about GA4 across a few categories - difficult to use, horrible, awful, unusable and brings users to tears.

Users were complaining that pertinent data was tucked underneath a sleuth of pages, third party add-ons don’t work and simple questions like ‘how many site visitors did I generate’ could not be answered.

If you search ‘GA4’ on Twitter today, the sentiment hasn’t improved much since it was first launched. There’s still a ton of people complaining and, of course, a ton of agencies and analytics companies offering their services to help brands navigate the forced transition.

It isn’t all bad. It’s often difficult to see the vision of something new if you were not a part of building it. Google is also a company that listens closely to user feedback so over time this should dramatically improve, as is the way with Google Ads. Ok, enough of me being a Google advocate.

HISTORY

Let’s get into the history and context. I wish I could do a whole deep dive into the history of Google Analytics tracing back from its roots to the Urchin Software Corporation but there’s already a lot for the history of GA4, which traces its roots to Firebase.

Google purchased Firebase in 2014 for an undisclosed amount and grouped it under the Google Cloud Platform umbrella. Firebase, to my understanding, is this platform where apps can be built such that you do not need a separate workflow for sending data and storing data. This would become one and make apps act in real-time. I am not a developer so if anyone listening is one, please please reach out and let me know if I got this correct.

This push to GA4 started 5 years after that in July 2019 and was brought about by the whole notion that advertisers couldn’t understand their web engagement and app engagement in an integrated fashion. By engagement, I’m referring to generally all activity that happens on app and web properties.

Even today, it’s quite a cumbersome process to understand how the journey continues once a user switches over to an app. You’d need an SDK like Firebase to build the app and an MMP to track all in-app actions and conversions. While this is possible, it is separate from all other marketing data reported.

What this means is that your strategy to drive users to apps is different than your strategy to drive app usage. If I’ve lost you already, please bear with me.

Driving users to apps is a fairly simple campaign. You’d run your normal campaigns across different channels to encourage people to download your app. You would likely optimize toward app downloads and measure your cost per app install. Google and Apple have pretty robust smart campaigns that do nearly all the work for you other than build the app.

Once someone has downloaded the app and begins to use it, your campaigns lose visibility into any action on the app. When you compare to driving people to your site, this is different because you can still track user action and behavior on your site and pass back that information to your campaigns. So if someone makes a purchase on your site, this conversion action is tracked back to your campaign.

Apps are different because they don’t automatically pass that information back. This is where an MMP comes in. It will allow you to track user action in-app and if you connect this data to like a Google Ads account, you will then be able to share the data.

App + Web, which is the property that would become GA4, was the property created by Google Analytics to bring both of these data sources together to better understand the customer journey. Brands would be able to understand which of their platforms drive the highest user engagement, how many active users they actually have regardless of platform and so on.

This is some very powerful data. It is already hard for brands to understand where to be investing their media dollars. Every brand has to understand the marketing funnel and decide how much of their budget should be allocated to supporting different stages of the funnel.

From there, brands need to navigate all the different channels that can support these different stages. You also have different tactics within these different channels. Take Meta as an example, you can choose between reach, traffic, conversions and many more as campaign objectives. You then need to decide whether to serve a video ad, carousel or static image.

All of these decisions are made by interpreting data. There’s even media mix modeling, attribution, incrementality that brands study to understand the veracity of their marketing strategies. All topics I should do an episode on. And, this is all just to drive users to the site.

Once people are on the site, that’s a whole other set of data and strategies to drive desired user action. This is where UX, CRO and web development teams come in to configure the site to make it as conducive to driving action as possible.

In parallel, digital marketers would want to understand what people are doing on the site to get an understanding of the quality of site traffic that is being driven. People who use a calculator, comparison tool, form submission and any other conversion would help hone the targeting of existing campaigns and enrich prospecting audiences through lookalike lists.

The same optimizations need to happen for any app a brand is offering their customers. It is fairly easy to get people to download apps but uncertain to get users to continue using apps. I, for one, just deleted maybe like 25 apps because I needed to free up space on my phone to take more photos of food I make to post on Instagram. Point being, there’s a lot of apps I don’t use anymore, which is revenue that I no longer provide to about 25 different companies. 

App + Web solved for having data in two different places, not talking to each other with brands not able to implement a more comprehensive strategy across their digital marketing activities.

Now, integrating data took this long because tracking activity on a site is inherently different than on an app. Universal Analytics, which is what GA4 is replacing, tracks based on sessions or page views. If someone lands on a site, that’s a pageview and if they stay on the page for a period of time that would be a session until there’s about 30 minutes of inactivity. 

For an app, it’s not really about time on site or anything that is calculated based on a pageview. Apps are created for different things and some apps don’t require long sessions to be meaningful.

This is why Google migrated to an event-driven data model with App + Web. It essentially means that actions would be aggregated into qualified events that are meaningful to brands. So, instead of equally valuing a person who reads a few lines on a page with someone who views a video for 30 seconds as one session each, GA will now count each as a different event. 

REASON

Let’s unpack the reason why this is happening.

I’m quite excited about this. It was 2018 when Google announced that their second party audiences would be available for search campaigns. These are the in-market and affinity audiences. I am actually not quite sure when affinity lists came out. That went through a whole other transformation. Story for another day.

Quickly after though, paid search practitioners would realize that these lists don’t actually do much. At this point in time, keyword targeting was still reigned supreme, and it still does today, but automated bid strategies weren’t as sophisticated and necessary as they are now and keyword match types haven’t changed yet either. I actually have 2 separate blogs that outline my updated views on all these topics so please do check them out.

But back to 2018, what we knew to do was just to add these lists to our existing campaign and bid up on ones that improved our conversion rates. This was alright but I wanted to do a whole lot more than a few bid adjustments here and there.

Enter GA audiences. It was such a pain to do this because you’d have to coordinate with the client, ad tech teams on client side and agency side and potentially others. But I dabbled in looking at behavior on-site as an audience targeting list. So instead of just people who are in-market for let’s say ‘dairy and eggs’, because I recently discovered that is a list that exists in Google Ads today, which like who is searching online to buy dairy and eggs. Ok, enough tangents. I can also look at people who spent a considerable amount of time on the site and/or people who also navigated to other pages.

With GA4, you take this a lot further. I’m getting a little ahead but, as a preview, I think it’s golden to be able to retarget and model off of people who have performed qualified actions on a site that indicate they are of high high intent. Say if someone is browsing shoes and they view a size guide and reviews but don’t immediately purchase, I would want to create an audience off of those 2 events and remarket to them.

This is why GA4 is becoming the standard and was renamed from App + Web. It’s really a shift in philosophy for how to view interactions users have with brands. And, this is regardless of property which means that even if brands did not have an app, it is still valuable to review events rather than sessions. Also, even if you don’t have an app, users can view your site from mobile browsers.

Now that’s one part of it. The other part is a response to the changing privacy landscape. Between GDPR, CCPA, iOS14.5+ and public opinion, cookies are being deprecated and there’s really not going to be one solution that replaces it. There is one word - identity - but it will not be a single solution across the ecosystem.

Digital marketing was ruled by the cookie and the 3P audiences it made available. With the move away from cookies, each platform is now devising its own way to make it privacy-compliant by shifting focus to 1P data. And with GA, Google is removing the option for IP addresses to be anonymized or not in GA4. UA allowed brands to opt into anonymization vs it being the default.

I don’t particularly think this piece has much to do with strategy but it does make GA compliant with GDPR because they view IPs to be PII.

STRATEGY

Now let’s dive into strategy.

Google Analytics has always been the platform for the brand team, web team and SEO team and site visits are really the KPI that they look at, in a universal kind of way. Generally, the more traffic to the site the more downstream actions that users take.

With GA4 using an event-driven data model, I can see this changing. Not all site visits were ever created equal and with more value being placed on the action performed on the site, teams can now make optimizations that would drive more of these actions.

I don’t know yet that this would entirely change an SEO keyword strategy. It will change the prioritization piece of it, I think. This is an over simplistic way of looking at SEO strategy but you typically want to optimize pages that have the highest chance of moving up in rank. If we factor in events to put weight on pages where there’s a lot of engagement, then the pages being optimized first can change. All of this is of course speculative at this point because not every brand has shifted over to GA4. UA’s forced retirement date is July 1, 2023 and more capabilities will be unlocked leading up to that date and especially after more brands are using it.

Over on the paid side, I touched on this a little bit ago but it really is all about audiences. From the ad side, Google shares lists of people who are searching for things that they like or are looking for and on the analytics side, brands know how valuable each person is to them.

Audience targeting strategies up to this point have largely been more of a singular view of things. You’re either leaning on search history or, if it’s 1P data, it would be based on customer value or something else that's static. I think the power of all this data is being able to identify the intersection of both that drives value to the brand. This could be people searching for related products and looking at a comparison tool. The former is something that Google can supply through Google Ads and the latter is something brands would know about their site. So, targeting the audience that does both would drive business impact to the brand.

Whatever the future holds for GA4 and privacy, I think what will hold true for years to come is that brands need to optimize toward user behavior because that is something that can be repeated. You can steer people toward the direction of the most engaging piece on your site and the more they engage with that, the more enticing it would be for them to be a customer. Whereas, if we only rely on someone being a high value customer, that’s not a behavior that can be encouraged other than providing reminders or being present the next time that customer is ready to buy.

ALTERNATIVES

First is, who is the alternative to Google Analytics. There’s actually quite a few but the main contender is Adobe Analytics.

Consumers know Adobe as the company behind PDFs and photoshop but they also have what they’re calling the Adobe Experience Cloud. Under this division is Adobe Analytics, which traces its origins from Adobe’s purchase of Omniture in 2009 for $1.8 billion.

Adobe Analytics is generally used by large e-commerce brands, mainly because the platform is so customizable to anything a brand would need. In my experience, large brands are also uneasy with Google accessing all their data so they like to use a variety of providers to minimize Google’s exposure to their data.

It is a little strange to me that Adobe has fallen off my radar for quite a while. I’d love to hear if this is the case with others. I wouldn’t be surprised if others share the same sentiment. Adobe is just so hard to use for anyone who is not the engineer that set up the platform for the brand. I used to manage paid search for a client just as their analytics person left and it was just a pain to get any digestible data.

I didn’t have much to share on this other than it is a viable alternative but not one that I’ve had a good experience using. So that’s that.

Second is, who could be the alternative. I’ve been noodling on this for a couple weeks but why is Microsoft not in this space? Or, are they?

Microsoft Clarity is a free analytics service that tracks site traffic. However, it’s really more of a user engagement tool than a traffic measurement tool like Google Analytics. Their unique prop is heat maps and session recordings which gives brands an understanding of how users are interacting on the site.

I don’t know what to think about this just yet. They only became widely available in 2020, 2 years after they launched as a beta. It isn’t clear where they are headed yet. If I were at the helm though, I think there’s a lot that could happen here.

Let’s start with the Microsoft Audience Network. This was launched in 2018 and was sold as all of Microsoft’s data coming together and being able to be used on Search targeting. It combines Office 365, MSN, Outlook, Microsoft Advertising, LinkedIn and even Xbox data, probably some others as well. The visual they showed when pitching to agencies and brands showed a constellation of data points that they had. And while I loved my Microsoft partners, this was pointless.

All that data and the best they could do is use it to build native advertising for their properties - MSN.com, Outlook and Edge.

Here’s what I don’t understand. Why is Microsoft Advertising still only Search? They have Bing.com, LinkedIn, PromoteIQ and Xbox. They recently acquired Xandr and Activision Blizzard. There’s also Skype and they run ads on Yahoo! and AOL.

Why aren’t these properties grouped into one advertising platform? Meta does it for all their platforms. In fact, it is the default for campaigns to serve on both Facebook and Instagram.

Imagine a campaign where you input keyword targeting, audience, creatives and this can serve on Bing, LinkedIn, Outlook, MSN, Edge, sure Xbox or any other property where they have access. Now if they use the MSAN graph to power this, that could be something.

They would then have the clout to create and push for something like Microsoft Clarity to be standard in user behavior and traffic measurement for any website.

Well, July 2023 is still over a year away but that’ll come around before you know it. In this case, it’s not too early to configure search campaigns to be more audience focused. You can eventually enhance these strategies with richer audience targeting data. I think the digital marketing industry will probably see much more traction by 2024.