In this edition of Augurs on the Town, Derrick Turner steps in for Josh Becerra to talk Google Analytics, GA4 and the death of the third party cookie with Rob Martin, Founder of Inertia Analytics. Learn what GA4 is, how it differs from Universal Analytics and why its important for marketers to begin preparing for this change.
If you’d like more information on GA4 or want to talk analytics – be sure to Contact Us.
Learn More About Rob Martin
Derrick Turner: Hey, everybody. This is Derrick Turner. I’m stepping in for Josh Becerra today for this edition for Augurs on the Town. I am joined today by Rob Martin. He’s the founder of Inertia Analytics. Thanks for being here, Rob.
Rob Martin: Thanks, Derrick. Glad to be here.
Derrick: Well, really excited to have this conversation. I know you’re very much in the weeds and focused on analytics. I had two topics that I would like to talk to you today about. Today, we’re here to talk about Google Analytics 4 and the future for a cookieless world based on the announcement from Google recently.
I’d like to start with Google Analytics 4. This is something that you brought to us very early, just in terms of thinking that was very important. I’m just curious if you can expand upon what it is and why people should care?
Rob: Well, Google Analytics 4, is the successor of Universal Analytics, which is not called Google Analytics 3, but was started the third iteration of Google’s Analytics tool. Both because of Google’s position in the market, and because of the rise of digital marketing as an activity, over the last 10 or 15 years, Google Analytics has been the de facto source of truth for many, many companies when they’re trying to understand their data.
We’re kind of at this moment of handoff between the old guard and the new guard. [inaudible 00:01:32] has really ruled the landscape for almost a decade now and what it replaced, which was called Classic Analytics, which I guess is Google Analytics 2 have the same underlying data model, but just the platform are a little bit different and what came before that was called Urchin Analytics, which was not a Google product.
It’s where it’s the U of UTM parameters that the U is Urchin and it’s sort of still the basis for the underlying data model that Universal Analytics uses. Google Analytics 4 is the successor to that. I think we can have a reasonable expectation that they will also dominate the market over the next decade or so and that’s really how Google is positioning the product.
The main problems it’s trying to solve, number one, is try to bring your website data, and then your native app data into the same place. There’s a solution for that, a Google solution for that already in place called Firebase and there’s a connection between Google Analytics and Firebase, there’s a product called Google Analytics 4 Firebase.
There’s ways for these interfaces to talk to each other but GA4 is built from the ground up to keep those contexts side by side, because ostensibly, if you’re a company that has both a web app is monetized or captures leads and a website that does the same thing, you don’t care so much if a thing is happening over here, over here, you want to get a complete picture of the sales cycle, the customer lifetime cycle, those sorts of things.
It brings those into a single house. It is an event-based data model, as opposed to Google or Universal Analytic’s session-based model. I guess, technically, it’s hit-based, but really, Universal Analytics try to present everything within the context of a session, which is a single person, a single browser, a single cookie over a described lifetime within these rules of windows and session start and stop.
GA4 is much less concerned about that. What it’s concerned about is activities that happen that might be of some value. One of the things GA4 tries to nail out of the box is things, metrics that have to do with timing. How are people actually engaging with your content? How are they moving through your content? When are important actions happening?
Then you get this much more flexible data model with each of these events that happen, you can just add, well, not unlimited, but you can add a lot of arbitrary parameters that meet whatever your business needs are. Some of that stuff comes for free.
There’s some things like the page URL, the page path, these things will all be sort of baked-in parameters but you also can add quite a few of your own, which is very powerful but also puts marketers in a position where they’ll need to think a little bit harder about how they build their data model and how they’re instrumenting their website to capture useful information and then maybe– Oh, sorry, go ahead.
Derrick: I was going to ask, what type of reporting do you think that this is really going to– Like what kind of questions I guess do you think that this is going to be able to answer better?
Rob: The paradigm with Universal Analytics was, first, you collect and then you analyze. Google Analytics 4 is trying to make this shorter cycle. One of the ways that they’re doing that is all of the reports that are baked into GA4, really follow this customer life cycle. Trying to understand more specific behavior patterns, and then how that actually affects revenue.
I think folks who either have e-commerce as a part of their sales process, or leads that have understood values, like being able to put dollars basically behind actions are going to see the most immediate benefit. There’s reporting like churn that helps you understand like how audiences are entering and leaving your sales cycle.
That’s all part of the basis, the baked-in reports of GA4. Then there’s this other analysis hub, which is a part of GA4, which allows you to do ad hoc analysis. It’s really similar to Adobe analytics product, where you can just drag and drop dimensions and metrics, you can do custom conversions of data and do analysis on the fly.
You could do custom reports in Universal Analytics, but it was clunky and they’re trying to shorten that analysis cycle to– I think they’re trying to get people to make decisions much quicker based on the data rather than hoarding the data and then figuring out what to do with it later.
Derrick: That’s interesting. When you said that, we’re very used to use it as marketers, just in terms of trying to justify the work and understanding, what can we use next best to drive success for a client from a metric that they care about, which is revenue versus just digital KPIs, conversions, et cetera. What is interesting is, it sounds like if it has a very user-based data model, that this might actually have some implications for companies that are outside of marketing.
When I hear that, I think, if you have a customer base that is logging in persistently, well, you might be able to use GA4 to answer questions about retention, and just other things that could have a relevant impact on your business in a way that, I don’t think GA, Universal Analytics, is really being used that way very often so that’s curious.
Rob: That’s a great point. Throughout most of Universal Analytics history, the concept of a user is basically tied to a cookie that lives on a browser, as long as that cookie can live. The default lifetime for that cookie initially was two years.
If you think back a decade ago, and what our behavior looked like at that point, you can imagine somebody who, maybe doesn’t know about clearing their cookies is less privacy minded, maybe uses the same browser for everything, especially if they’re at work, maybe uses the same device all the time, and is less likely to pick up in one place and continue in another place, like on a tablet, or a smartphone or something like that.
The Universal Analytics model worked okay, from that perspective, because these cookies were persistent and because we were less fragmented in how we moved across our devices for these journeys to answer a single question. Maybe we did some of our question answering, like we do our shopping on one device and we did our work related stuff on another device, for example.
Google Analytics 4 has a tougher nut to crack, because obviously, our behavior isn’t like that anymore. We’re more privacy conscious, we use more devices and there’s a lot more touch points along the way to consider. We’re also entering this world where cookies are having shorter and shorter life cycles, both because of settings in platforms like Google Analytics, but also because of privacy initiatives that have moved forward and we can’t really count on the cookie to tell us who somebody is.
If you are lucky enough to be a company that has a login experience, then you can use a user ID to identify that person in Google Analytics. That feature was eventually released in Universal Analytics, I might have said Universal Analytics there twice, but first one was supposed to be GA4.
There was the concept of a user ID and Universal Analytics that was released somewhat late in the product lifecycle, but it’s one of the core features of GA4, and there’s this catch all system of trying to figure out who somebody is. Cookies are still going to be set, that’s a way.
If you have a user ID, and you can match that across your app and your website, that’ll be the ultimate source of truth. Then there’s this product called Google Signals, which is Google’s way of just figuring out who you are because of their dominant market share.
There are some things there, you need to opt into that as a marketer and then there are some privacy side things too people need to have when they’re logged into Google services but a lot of people are and don’t even realize they are. A friend of mine the other day was telling me that he’s like opted out of Google and I asked him what he used to get his driving directions. He’s like, “Oh yes, Google maps [laughs] obviously.” Or Waze, which is another Google product.
They have their tendrils everywhere. Google Signals is going to be, I think, pretty good at piecing those things together. In the absence of any of those things, they’re pitching GA4 as a machine learning oriented tool. They’re going to use modeling to fill in the blanks.
What that looks like yet, we don’t really know. iOS 14.5 is top of mind because that just came out this week but one of the things that happened in another Google product, Google Ads, is they’ve introduced model conversions. If we don’t know what’s coming out of an app, we’re going to take a guess at what happened. You can still attribute that.
I’m sure modeling like that is going to be an important part of how GA functions in the future. I don’t think you’ll see less data even when data is sparser. I think you’ll see the same amount of data but how much of that is observed data versus how much of that is model data, I think that ratio is going to shift quite a bit over time.
Derrick: Sure. I think we’ve been using that more and more are on the Google Ad side. I know tracking in-store visits is largely modeled but still very useful, very compelling in terms of optimization. I see where you’re going with that. Where is GA4 now? Is it ready for prime time? Or can you speak a little bit about that?
Rob: You’re just lobbing softballs here, Derrick. No, it’s not ready for prime time. The product that came before GA4 was called Google Analytics App + Web and it was in beta for a long time. It was largely the same. It used the Firebase data model and allowed you to have your app and web data in the same place.
That’s been out for over a year and then late in 2020, Google announced that that was coming out of beta. It was now going to be a full fledged product, called Google Analytics 4. Normally, when something comes out of beta, that’s a signal like this is safe to use, like, “This is stable, we’ve tested for bugs.” That is just not the case. It’s not ready for prime time for a lot of reasons.
Some of those reasons are, it’s not a mature product, it just hasn’t gone through it’s paces. Think about how many installations of Universal Analytics exists in the world. Pretty much anything that can be discovered about that has been discovered, which just means it’s very stable and if you’re a business, maybe you’re just used to seeing things a certain way.
Even if Google Analytics 4 is better and you can explain to somebody why it’s better, sometimes you just want to look at the thing that’s familiar. That’s going to be an impediment, not necessarily on the technology side but on the change management side and the kinds of conversations that marketers and digital analysts need to have with the ultimate consumers of this data.
Beyond that, it’s not feature-complete. Even within the Google ecosystem, there’s a great integration between Google search console and Universal Analytics, which I try to turn on for a 100% of the implementations I touch, it doesn’t exist in GA4 yet.
A couple of weeks ago, if you wanted to filter out your internal IP traffic, if you wanted to set up referral exclusions, these are fundamental settings in Universal Analytics, things that you look at every single time when you set up a new implementation that exists or only existed for some customers because they were rolling out piecemeal.
Then beyond that, just sort of in the measurement community, there’s lots of observations of things that just don’t compute, just bugs basically. Things that are unexplained. Maybe one more thing to say about that is, again, a change management question, which is like, when are we ready to consider things differently?
For example, I like to dunk on bounce rate. I think bounce rate is just a pox on marketers analysts everywhere because it’s front and center and all these Universal Analytics reports and it’s almost universally unactionable or the opposite of useful. It causes empty work to be done. Bounce rate is gone in G4. I think that’s amazing. That could take some explaining to people. Another thing which was core to Universal Analytics we talked about was this concept of a session. The concept of a session does exist in GA4 but it’s different.
There’s not going to be metric parity, even if you have completely identical implementations across UA and GA4, it’s not going to look the same because of how that metric is calculated in GA4 and what it actually represents. I think these are all the sorts of things that there’s no shortcut. It’s just going to take time.
The recommendation, though, is if you’re committed to Google Analytics products, if you’re committed to the Google ecosystem, install GA4 right now. It’s quite simple to do, in most cases, it’ll give you some data, which you can’t go back and collect later, and I think maybe most importantly, it’s really the best possible I think we have for that change management piece.
We know that Google Analytics 4 will be the de facto analytic solution again, or at least we have every reason to believe that it will be, because of Google’s dominant position [unintelligible 00:16:10]. Why not give yourself a leg up on everybody else and start getting used to what the future is going to look like?
Derrick: It makes sense. That was my next question. If I’m a company, what should I do next? Sounds like get started, implement, and start collecting data and let the product mature. Thanks, Rob. It’s going to be fascinating, I think, just so with those, basically, concepts being removed from what most people are going to be looking at, how that changes just how we talk about digital marketing, to a degree, bounce rate, for better or for worse, is talk about stuff quite a bit.
That’ll be interesting. Thanks for that. Anything else or other high-level notes? I’d like to talk about the future of cookies, or lack of them, rather.
Rob: Yes. I think there were just two other really high-level bullet points of why GA4. One standout feature is a free native integration with BigQuery. Right now, if you want to dump your Google Analytics data into BigQuery, you need to buy the 360 version of Google Analytics, which costs six figures a year. That’s a pretty big deal. It says a lot about how Google is trying to evolve its data ecosystem.
Then the other thing, of course, again, and maybe this is a perfect segue to cookies, is Google is trying to be more privacy forward. There will be different controls that are available in GA4. It’s supposed to be the solution to the cookieless future. There’ll also be quality of life things, like if you’re a company that needs to handle data deletion requests, it’s much easier to do that in GA4 than it was in Universal Analytics.
Derrick: Oh, terrific. That’s great. Thanks for calling that out. Well, let’s move on then to talk about the future in terms of cookies. In the simplest terms possible, can you describe what’s about to happen with cookies?
Rob: It’s not just one thing. It’s more of an attitude and a movement toward a more privacy-forward online experience. I feel like in the marketing world, there’s lots of jargon that’s turned around like cookie get in, the cookieless future, the death of the cookie.
That doesn’t really represent a single event or a fixed point in time but this transition period that we’re in. It represents a lot of individual changes in a lot of different companies, governments, products.
That’s been the harbinger of what might be coming in the future. Just more recently, the state of California passed this CCPA, which, unfortunately, is different in content than the GDPR. If you are a company that has to comply with both of those things, you have two different scenarios that you need to figure out.
I feel like these are the headlines that people are seeing that are related to privacy. Beyond that, last year, Google announced that in two years time, the cookie was going away. I think that was really the moment that, for a lot of marketers, crystallized like the sky is falling because marketing has relied on these third-party pixels for so long to drive the whole process forward.
Cookies are not going away. I think third-party cookies will be blocked by browsers, cookie lifetimes will be shortened, default settings, across the most popular browsers, will definitely become more privacy-forward over time. Just recently, for example, there is a value in a request header called refer, which tells the page receiving a request where the traffic is coming from.
For example, if you’re on myblog/recipes and you click on a link on that recipes page to go to a site where you can buy those ingredients, the header from that GET request, you would see that the refer was /recipes and that’s great, you know like, “Oh, interesting, this traffic is coming from there.” One of the recent things that has happened is now, the referrer is being truncated in many cases, at least by default, to the hosting. You’re losing the granularity of that page path, that ultimate path.
That’s a piece of it as well, it’s a death by a thousand cuts kind of change, rather than one cataclysm moment. We talked about iOS 14.5, that doesn’t have anything to do with cookies really, but it doesn’t have to do with privacy. I think it’s one of those- Just another item that adds to this, “cookieless future”, what that really represents is that the way we capture data, the amount of data and the quality of that data is going to change and that’s the fact.
Google Analytics 4, again, is trying to account for that. It’s assuming a data-sparse world, and it’s planning on filling in the gaps with first-party Google data, and machine learning, modeling. Yes, maybe that answered that question. I don’t know. I feel like I can just keep talking but I’ll pause.
Derrick: No, there’s a lot going on. I think that’s clear. From our perspective, I think, what a lot of marketers and clients are worried about is just there are levers that they have with their advertising potentially being degraded, as a result of those cookies going away. I know a lot of DSPs, programmatic offers, they do rely on those cookies to some extent to help them figure out when and who to target with an ad.
I guess just generally, we do a lot of Google Ads, Microsoft Ads, LinkedIn, Facebook, those major publishers, what type of impact do you think it’s going to have I guess, just with the specific update for Google and the browser, what type of impact do you think that’s actually going to have on those big publishers?
Rob: I don’t think this spells the death of these publishers. Nobody knows for sure what’s going to happen, because a standard hasn’t been agreed upon. Google created this thing called privacy sandbox, which is just this area, where they will try out some of their cookie solutions that the public can play around with in the sandbox and test out to figure out what the next standard will be.
Google is trying to set the standard. The thing you have to realize is that Google, which is hugely profitable, I think their first-quarter earnings were like $45 billion. They’re doing fine, they’re deeply invested in making sure that this source of revenue for them is going to not go away. I think these publishers are going to try hard to make the impact as minimal as possible.
I don’t think we know quite sure from a technical implementation standpoint, what that’s going to look like. The days of the third-party cookie are numbered, I think that’s a foregone [inaudible 00:24:17] but that doesn’t mean that it can’t be replaced by something equally intrusive and awful.
Which is in my opinion, at least, what Google’s FLoC standard represents. It’s the Federated Learning of Cohorts. That’s an ID that’s set at the browser level. Then basically, based on your browsing behavior, which is public, by the way, anybody can inspect that attribute in the browser.
It gets put into one of the lists that they maintain and then as a marketer, you can say like, “Oh, I’d like to market to the young tennis enthusiast list, or whatever.” Well, part of what said is they work to make the list so specific that you could figure out who individuals are, but there are major concerns, privacy concerns about what this looks like. Pretty much to a person or to a browser, all the other major players have come out and said, “We are absolutely never adopting FLoC.”
It’ll be interesting to see if Google is able to jam that down the throats of everybody else, because of their dominant market share and just say, “We’re doing it.” Or if they really do take a good faith, approach of getting feedback and finding something that they’ll have other partners take on.
To get back to the original question, I just don’t think there’s going to be a cookie cliff per se, where one day Microsoft Ads works and then the next day it doesn’t. I think it is more like a death by a thousand cuts. The mechanism for how our marketing activities happens weaker, but there is such a deeply vested interest in not losing this revenue for these platforms.
I don’t think our ability, I don’t think they will necessarily, be so unenticing to brands that people will move away from them. At least not wholesale anytime in the near future.
Derrick: Sure. I tend to agree. They leverage a lot of data that they, in fact, own as well. It’s not like they’re relying on third party cookies to generate a lot of the levers that we have for targeting in general. I’m curious, and you might not have an answer to this, but say, Google wins in terms of setting the standard. What are the advantages to that?
Rob: Well, are you saying Google winning or Google’s proposed solution?
Derrick: Google’s proposed solution.
Rob: Well, the advantages are, it really would mean the end of third-party cookies. All of the baggage that comes along with that, will have the moral victory. We did it guys, we got rid of third-party cookies. The FLoC solution, one of its features is that it’s constantly recycling things. You might be in the young tennis enthusiast lists for these couple of weeks, and then next week, you’re in the tech savvy entrepreneur bucket.
Things can evolve quickly. There’s less of a memory, I guess, there’s less of a fear of just gobs and gobs of data being captured by boogeyman out there. Critics of FLoC would say that, bad actors will still abuse this. There’s some scenarios of, it’s not exactly fingerprinting, but basically reverse engineering FLoC to just figure out who people are anyway.
Concerns are raised specifically around groups of people that could be identified because they might be a relatively small group of people. Google is trying to allay these fears, but the reality is, you can’t have it both ways. If we really want a privacy anonymous for first internet, then we’re going to have to find a different way to monetize it.
Take iOS 14.5, for example, who are the losers? In my opinion, the biggest losers are app developers who would monitor, but then selling real estate within their apps. That’s like a common paradigm for monetizing your app is. Maybe it’s free to download and then you gather data about your customers and that’s a product that you can sell to advertisers, and you can also sell real estate within your app to advertisers.
Now, with the consent screen, which by the way, developers don’t have much ability to affect what that looks like, or like this is all happening on the apple server side of things. The default option, the one that’s easiest to click is like, “Don’t track me.”
Early indications pointed about an 80% opt out rate. At least in the short-term, it sounds like folks who rely on monetization through ads in their platform, those developers are going to be the big losers. If you want to put on your tinfoil hat for a second, Apple plays the care about price when nobody else does card. They’re also primarily hardware manufacturers, so they have less skin in the game, but their cut from the App Store is 30%.
If you’re putting on your tinfoil hat, you could say that what Apple is doing is encouraging developers who are sending their money elsewhere through the ad platforms to actually monetize their apps through the App Store, maybe via charging, or if you have a subscription product, that gets charged to the App Store as well, and Apple gets a 30% bite of that, every single time. I don’t actually remember how we got on that topic.
Derrick: [unintelligible 00:30:45]
Rob: There’s my 14.5-download for you.
Derrick: Just kind of a side, you mentioned the dynamicism of being in specific audiences. Is there any way to be able to deconstruct just how that operation works or how people are moving in between audiences?
I know, for instance, Google Ads has a bunch of audiences that you can use to target. One thing that we don’t know is like, well, who’s in the audience, for how long, and are they in other audiences? How do I understand the status? Curious if you have any insight on that.
Rob: Well, maybe one thing I’ll talk about is just Google Analytics 4 is solution to that. One of the pitches for Google Analytics 4 is a tighter integration across its ecosystem, which is pretty significant because Google is, by far, the most dominant search engine in the world, more than 80% market share, and the number two search engine in the world is YouTube.
One of the selling points of GA4 is you get YouTube data in your Google Analytics 4 reports now, like viewed data, engaged viewer data. That’s a tighter integration. That can be used as a part of how you segment your audiences. Remember, GA4 is a customer lifecycle-centric model that then gets used in Google Ads.
The term that gets thrown around here is walled garden, being the Facebooks of the world, the Googles of the world that already have massive, massive troves of first-party data, are going to be affected less by all these changes because they can just keep you in their ecosystem. You’re already logged in. They already have your data.
From an audience targeting perspective, on the Google side, I feel like FLoC maybe isn’t the biggest piece of that. It’s more like the walled garden side of it, which is Google’s going to try harder than ever to keep you in their ecosystem. The tools that they’re building are really designed to not lose that ability to segment and target.
Derrick: That sounds to reason. With the amount of no-click searches we’re seeing, that’s the start of it. I love how you were able to tie that back to GA4, Rob. In terms of parting thoughts, any other things that you’d really want to get out there, as far as just the changes that are coming and what people need to know or do about it?
Rob: I think the first thought is just, in my opinion, the sky is not falling. Things are going to change, but things have always changed, and they’ll continue to change. We are at a transition period. Those always feel strange, especially at this moment in time when I feel like, as a species, it’s been a transition period.
I understand the anxiety around that, but marketers aren’t going to lose their ability to market. Brands aren’t going to lose their ability to gather analytics on what their customers and what their properties are doing. I think it’s actually pretty reasonable that the net of all of this is that people get smarter with their analytics because the tools are getting better. Even if the data is getting sparser, that the tools are getting better.
From a Google landscape perspective, maybe this is one point to call attention to here, is they have never had a clear vision for what their data pipeline looks like. What that looks like is that Google Tag Manager is this product, which, by the way, they just released a version that is server-side that is a serious heavy hitter. It’s not geared toward marketers at all.
Google Tag Manager was initially geared towards marketers, but server-side implementation is geared really towards developers. You couldn’t set it up without somebody who knew what they were doing on the backend. It still has all the features that marketers loved about the ease of use to manage tags.
They’re using that as their mechanism to ingest the data. Google Analytics 4 is the initial data store and pre-processor of the data. Then you’ve got two exit points from there. One is up at BigQuery where you marry that with all of the rest of your data and create the full picture.
Then the other outlet is Google Data Studio, where you do all of your visualizations and that’s the front end. When you’re ready to put a line in the sand of this is the stuff we care about, you take it out of the analysis platform and you put it into this visualization platform, Google Data Studio.
They’re removing barriers every step along the way in terms of who might want these products and what they can do. It shows that more than a decade has passed since we’ve started doing this work in a more mature vision for how it all fits together has arisen to meet our current moment in time.
Derrick: Great. I like how you phrase that. Reason to be hopeful. This is great. We covered a ton of content here, Rob. [inaudible 00:36:16] in some areas, I think these are important topics. Very much appreciate your time and just your thoughts. It’s clear that you think a lot about these things.
Rob: [laughs] It’s my pleasure. I don’t know why I get so excited about it. I just think it’s interesting. I feel like this little corner of the marketing world is a little microcosm for issues that are way bigger than that. Societal issues of who we are, how we think about ourselves, what we want [inaudible 00:36:49].
How do we function as an ethical democracy, if such a thing exists? You can wax philosophical about all of this and I think that’s the appeal.
Derrick: Awesome. Well, thanks so much for your time today, Rob. We’ll sign off from here, but excited to keep the conversations going.
Rob: Thanks, Derrick. Appreciate it.