Candour

Episode 89: Google Analytics 4 (GA4) with Simo Ahava and Krista Seiden

Play this episode:

Or get it on:

What's in this episode?

In this episode, you will hear Mark Williams-Cook talking to Simo Ahava and Krista Seiden about GA4 and answering community questions, including:

  • What are the best new features in GA4?

  • How does GA4 handle cookieless users?

  • How is GA4 using AI to help marketers?

  • What are some alternatives to GA4?

  • Why is GA4 more "user-centric"?

  • How is the new event model different?

  • Will we still be able to get the same segmentation insight?

  • Will users be forced onto GA4 eventually?

Show notes

Google blog: announced 14th October - Google Analytics https://blog.google/products/marketingplatform/analytics/new_google_analytics

Google blog: A new way to unify app and website measurement in Google Analytics https://blog.google/products/marketingplatform/analytics/new-way-unify-app-and-website-measurement-google-analytics/

SEO Community Questions https://www.linkedin.com/posts/markseo_seo-ga4-googleanalytics-activity-6734096852838305792-pSVL

Google blog: Take control of how data is used in Google Analytics https://blog.google/products/marketingplatform/analytics/take-control-how-data-used-google-analytics/

Transcript

MC: Welcome to episode 89 of the Search With Candour podcast, recorded on Monday 23rd November 2020. My name is Mark Williams-Cook and today we're going to be talking all about Google Analytics 4, the new GA4. We're gonna be going into the paradigm shift it's bringing about in analytics and we're really fortunate to be joined by two excellent guests to lead me through this. We've got Simo Ahava, very well known Analytics and Google Tag Manager expert who's got a blog, has been helping people for years, provides loads of tools for GTM and tutorials. We're also very fortunate to be joined by Krista Seiden, who actually worked at Google for seven years, on Google Analytics, so incredibly knowledgeable, a couple of people to help us answer your questions and answer my questions on GA4.

This episode of Search With Candour is very kindly sponsored by our friends at Sitebulb. Sitebulb is a desktop based SEO auditing tool for Windows and Mac. They've sponsored, very kindly, quite a few episodes now; it's a tool I've used for a long time, We use it in the agency. If you haven't heard of it, it's an incredible bit of software to help you with your SEO auditing. I tend to, every episode, just have a chat about one specific part of Sitebulb and, today, I'm going to talk to you about the kind of stuff they do around duplicate content. Duplicate content as you know, as SEOs, can be problematic, especially on larger sites.

Traditionally, when you use other tools to do a crawl of a site, one of the ways I've seen people very quickly spot this is, they might just look at a page and sort by title and see where you've got things like duplicate titles, because that's a really good kind of guess as to where you may have duplicate content. The really cool thing about Sitebulb though is, it will break it down further for you and looks a lot more in-depth at things you could otherwise miss. When it does a duplicate content report for you, it will report separately on things like pages where you have duplicate titles, of course, duplicate meta descriptions, duplicate header ones. All of those things will hint that you've got two pages here and they're kind of targeting the same thing. What it also does, independently of that, is actually looks at the content and html, so it will alert you if you have two pages that are very very similar in content, even if they have different urls. I found this is really good at handling some kind of edge cases with different platforms clients have been using, so maybe not common off the shelf like Wordpress or Magento kind of stuff where they can have pages generated with different titles but, actually, the content's exactly the same. It's almost like a little copyscape you've got built into Sitebulb.

It's a really great bit of kit, does great audits, as we've said before, prioritizes them, and gives you feedback on all the issues. They've got a deal for Search With Candour listeners; if you go to sitebulb.com/swc, you can get a special 60-day free trial of Sitebulb - no credit card required. So it's sitebulb.com/swc, go and check it out.

Today we are very lucky to be joined by two very special guests. Firstly, we have Simo Ahava, who is partner and co-founder at 8-bit-sheep, and Google Developer expert for Google Analytics and Google Tag Manager. Welcome, Simo.

SA: Hey hey, thanks for having me.

MC: We also are very lucky to be joined by Krista Seiden, who is founder and principal consultant at KS Digital, and you actually used to work at Google as a project manager for Google Analytics, is that right Krista?

KS: Yeah, I was at Google for almost seven years, as both the evangelist for Google Analytics and a product manager for Google Analytics. It was a lot of fun.

MC: I don't think we really could have had two better guests for this podcast, to talk about GA4. Simo, do you mind giving us a very quick introduction to yourself, for those that haven't heard of you. You're certainly known even within our agency, I know people refer to your GTM tutorials. You're referenced, I don't know if you even know, on some of the episodes we've done, especially when I need help decoding what's going on with the latest version of ITP, and you've done some really great breakdowns and helped me understand that. Did you want to just tell a little bit about your history, and what you're interested in, and who you are?

SA: A little bit of, oh man, this is so difficult... I'm a developer. I've worked as a developer for many years now, specializing in analytics since the early versions of Google Analytics. I like to think of myself as a hobbyist, first and foremost. I don't necessarily enjoy doing this stuff for a living, but I do love it when i get to write about it in blogs, and talk about it in podcasts, and present it in conferences, so I'm very much just curious about software development and what data has got to do with that. Over the recent years also about privacy, and how web browsers and mobile devices are trying to level the playing field when it comes to using user data. That's about it in a nutshell, I think, I'm not going to start boring people with my job history. I love to code and I love to talk about that stuff, and analytics has always been a passion project of mine.

MC: Brilliant, I think privacy is something we will talk about as well, when it comes to GA4. Krista, tell us about yourself.

KS: I've been around the analytics industry for... probably about 12 plus years at this point. It's definitely, as Simo said, a passion project for me, although I do also enjoy doing it for work. It's kept me busy over the years, I've been in several different organisations; I was at Adobe in the early days, then on to a few other places before being at Google, and now I run my own analytics consultancy. It's been a lot of fun along the way and I would say, true to my core, I'm a practitioner and analyst, that's how I approach analytics. Yeah, that's a little about me.

MC: Brilliant. This is really exciting because a couple of episodes ago, we were speaking to a very friendly chap called Roman Sedowski, who's Head of SEO at Smyth's Toys, they get in the tens of millions visitors a month, and we were talking about how they were using analytics to optimise different aspects of their e-commerce business, and helping them out with SEO. We got onto the conversation that, as an agency, we've spoken to clients before, or potential clients, who are very big companies and sometimes they don't even have the most basic web analytics set up. We gave this example a couple of weeks ago, I spoke to a very well-known global brand and they wanted to improve their leads, and do this and that. We got into their analytics and nothing was set up in terms of events and goals. When you're working at the level where both of you obviously are and you see this, it's a bit like “wow, I can't believe that nothing's going on behind the scenes here and all of this value is left on the table.”

KS: You see it more than you want to believe actually.

MC: Absolutely! My question to you is, GA4 was announced, I think it was the 14th of October, on the Google Blog officially and we've enabled it on a few profiles, and I think many people like me, who have gotten very used to using Google Analytics over the last decade or so, it's quite a shock when you first log into GA4. I can imagine some people in a similar boat are almost like “oh no this is another big thing I've got to learn”, because it looks really different when you first log in. What are your first words of advice for people in this situation?

KS: It does look really different when you log in. If you've played with app and web at all over the last year and a half it's the next iteration of that;. so you have a footstep up if you've been playing there but, if you haven't and ga4 is your first look into this new way of doing analytics, it's definitely going to look different. The schema is totally different the way, you collect data is different, the way it's processed is different, the way the reports look is different, what's available is different, and it is definitely something new to learn. I'd suggest trying to find the things that look familiar and start from there to really help reduce some of the overwhelming anxiety that it might bring, to try something completely new. I do think there's a lot of really great resources out there already, from people who are trying to help you bridge that gap. I think that's the good news, but definitely something new to learn.

SA: For people getting started and just trying to figure out what to do first, for example, and not be too overwhelmed, I think there's two things about GA4 that really make it spectacular. When you compare it to Google Analytics, the base snippet, the thing that the long tail of the websites of the world just have that and nothing else, it doesn't really do anything. It just collects the page view and populates some metrics, but it doesn't give you the events and the goals that you mentioned. GA4, on the other hand, has the base packages really complete, it has this enhanced measurement enabled by default which adds all these engagement metrics and video metrics and scroll depth. I'm brave enough to say that you can actually be complacent here, you're allowed to just install the base stuff and let it collect data for a while, get used to the new data model, get used to new reports without having to worry your head about the new event schema or even BigQuery, or any of these other tools that GA4 has. You can start really easily, just put the base stuff there and allow it to do its magic.

KS: Yeah that's a great point. I do think that getting started with GA4 is a lot easier than getting started with Universal Analytics.

MC: It's really interesting, what you're saying about that better granularity of data you're getting with the base snippet. In Google's announcement about GA4, they outlined several reasons why they're moving to this product. The first one they gave was, Google said “there's been major shifts in consumer behavior and privacy driven changes to long-time industry standards, and current analytics are not keeping pace” and I'd be really interested, what are your opinions on, firstly, what are those changes in consumer behavior and what are the privacy driven changes that GA4 is tackling?

SA: I'm just going to veto myself out of this because I have no idea what Google means when it uses those words in their blog posts. I have no idea what consumer behavior I would have shifted in the market spaces, so somebody smarter than me will have to explain those gobbledygooks up to me.

KS: Oh man, I was gonna say this one's got your name written all over it. I think the changes in consumer behavior they're talking about are really cross device enabled. How you report on users moving from desktop to app and vice versa, where GA4 really focuses on tying that user identity across devices. I think that's one of the things that they're getting at there. In terms of privacy, there's so many new regulations that have degraded what we can report on as analysts, in terms of the cookie and all that, Simo, this is where I look to you because you put so much time and thought into that side of things, but GA4 really is looking to enhance that. They've got a plan for the future around how they'll use modeling and various predictive behaviors to be able to fill some of those gaps.

SA: It's very difficult to actually enact privacy well in the older versions of Google Analytics, like being able to do data deletion requests, data access requests, being able to apply consent, different varying levels of consent, it's very difficult to do that with GA, and rather than retrofitting that stuff into Universal Analytics, Google has been able to build GA4 from the ground up to have these APIs that give you far more access to data and governance and accountability. It's still not perfect, Google is still a global player and it's very difficult to get the nuances of all the possible different privacy legislations encoded without overwhelming the user, but the data model itself has been built in a way that allows you to access the data, from a privacy perspective as well, in very different ways.

MC: I've certainly, from an agency point of view, seen lots of companies struggling to keep up with what they can and what they can't do, in terms of privacy, and in terms of things like cookies, and which cookies are classified as what, and exactly what can and can't be set before they've clicked on this, and whether this needs to be an active opt-out or opt-in. I know that's caused a lot of confusion, especially for in-house marketers, where they maybe deal with these situations like once, as opposed to agencies that are helping lots of clients. I did see there are, and you both mentioned there, in GA4, these more granular data controls and possibilities around deletion, which are really helpful. One thing that particularly interested me, and I don't know if either of you can expand on this, is this “cookie-less” approach, which has to do with Google saying that, in situations where they haven't got these cookies set, they can use modeling to fill in those gaps. Can you give any insight or explanation as to what that means and the potential of that?

SA: We can but then we will have to kill you.

MC: Deal!

SA: What you just said is about as much as I think we can talk about it. There is the constant mode, which is now in beta, and it's any it's available for anyone to implement right now. The question is that it's not particularly clear what happens to those cookie-less pings because they're not surfaced in the reports. They might appear in real time for a while but that's likely something that will be cleared out at some point. There's no official word about what they will actually do with that data set, apart from these terms like modeling and machine learning being thrown around. I personally think it's a bit worrying, it's kind of counterintuitive that you have a privacy-first approach that collects data that's not actually accessible, so I really hope that Google pushes forward with the documentation and the pr around constant mode because, for right now, it's marketed as a developer tool. I think that it needs to expand from that to cover the entire application and have all the proper documentation that we, as users, are demanding.

MC: As an opinion actually, for both of you, do you think, analytics-wise, we are heading towards a future with less and less cookies?

KS: I think that's inevitable, in terms of where the industry is going. I think the big question that's been on everyone's minds, for a few years now, “is how fast will we get there?” Cookies going away is a big deal, the internet relies on them, so i do think that there's a lot of pieces that have to fall in place before that is really truly a thing across the board. That being said, I do think tools like GA4 are starting to think about what that future looks like and how they can try to fill in the gaps.

SA: Yeah I think a really good good way to frame it. One thing that I always like to say when this comes up is that we're focusing on cookies being taken away and we, as analysts, feel like, people working with first-party analytics, first and foremost, so web behavior and app behavior, I think we feel like we're being collateral damage here, in the browsers and legislation's attack on the big players and attack on the cross side tracking stuff. I think that we focus a lot on things being taken away from us, that I think we might forget that it's not just about things being taken away from us, but new APIs and new browser tools are being introduced that will let us keep collecting data. They will just do it in a way that's more friendly to the natural imbalance between user privacy and what companies want. We're definitely looking at a future where cookies themselves have degraded to a point of maybe being completely useless, but we might also be looking at a future where this technology has been replaced with something else that gives us back the quality of data that we want, without sacrificing users right to their own data.

MC: Krista, a double-barreled question for you. I think you hit it on the nose when you talked about decoding what Google was saying about changing user behavior and things like cross device tracking, and people being more comfortable you know using a site on their desktop, and then mobile and back, and you mentioned that GA4 is this evolution of the web and app property beta that Google introduced. I was just wondering, is there anything actually significantly different, or any other features that have come since this web and app property beta now, compared to GA4? I hear GA4 being described a lot, in the language Google is using, as more user-centric approach to analytics. Is that mainly what we're referring to, this cross-device nature, so we're more focused on the user rather than the session?

KS: Yeah, I think that's part of it. To be clear, GA4 is really a rebranding of app and web, there's no baseline difference of the tool itself. When they did rebrand to Google Analytics 4, they also released a set of new features that had been in the works for a while. I do think that there are a lot of new things in GA4, that you didn't see in app and web, that make it very very exciting. A couple of my favorites are the ability to modify and create events within the user interface, to set up cross domain tracking within the user interface, both of these things are things that you couldn't do previously in Universal Analytics, and they've made it so easy that you can do it in the UI without having to write code, which I love. There's also an added focus around security. We talked a bit about privacy and how this tool has been made for that. It's also taken a focus on security, one of the things that people loved in Universal Analytics was the measurement protocol. For the life of the app and web beta, you heard a lot of feedback as to: “Where's my measurement protocol? How can I send data to GA offline? When GA4 finally released, they released the measurement protocol with it, which includes an API secret key, which allows you to actually make that call specific to what you're doing, so that other people can't spam the data coming into your analytics properties, which is a huge step up in terms of the data quality that you're getting in Google Analytics 4. So, while it's not entirely different from app and web, it really is the same baseline project product. It has built on it significantly over the last month, with a lot of new features being released that are looking really exciting for the future.

SA: App plus web is the iteration of Firebase Analytics right, Krista? So we're looking at a third generation of an analytics tool already, in a sense.

MC: The ease of cross-domain measurement, I'm sure, is going to be a relief to a lot of people, having wrestled with that myself a few times, and going through various iterations of testing before it definitely, definitely worked. That was an interesting discovery with the measurement protocol, I remember when I discovered that it was possible for me just to put events into people's analytics for them, that I had no connection to, so that's a welcome change as well. The last point on Google's blog post around the release of GA4, why this is a good thing, was around machine learning and return on investment. In terms of machine learning, we've had these insights popping up within Universal Analytics for a while and they're pretty helpful. I can see how they're less brittle than trying to set up your own custom alerts, “is this product popular? has the traffic gone up” and I can see how machine learning is helpful surfacing those. We talked about using modeling to fill in the gaps, where cookies are missing. Is there anything else that I'm not aware of, I haven't mentioned there, where machine learning is being used in a significant way that's going to benefit us for ga4?

KS: I think it's really built into the core of GA4. I think Google is trying really hard to make it core to how the product operates and the insights that are being shared. I suspect that there's going to be a lot more to come there in the future. I think the intelligence features are still working out some kinks. I think they've been building and getting better over the last couple years; first in Universal Analytics and then in app and web and GA4, but I think there's a lot to be seen coming in the future.

SA: I have to say about ML just that, and I'm not sure if we're gonna go there in this podcast, but one of the biggest selling points on GA, for me, from day one, since Firebase Analytics and app plus web, has been the BigQuery export. It's definitely for the analytics nerds out there. It's not necessarily for everyone as an entry-level product but we now do have an open doorway to the Google Cloud Platform through BigQuery export. This basically means that you have access to a data warehouse where Google Analytics or GA4 daily dumps a raw set of that day's data, you can even get almost real time streaming export as well. The thing that BigQuery enables is, it opens all these components in the Google Cloud for you, and you can actually start doing really simple machine learning stuff by yourself within BigQuery itself, there's a BigQuery ML tool which lets you model with data that you already have. This is super low barrier of entry, obviously you need to take a look at a couple of blog posts and a couple of tutorials, but I think this is another spin on GA4 being an ML-driven tool, is that it actually gives you the raw data set that's so fundamental to being able to do this stuff by yourself and get introduced to the world of of algorithmic data processing.

KS Absolutely, that definitely is something that sets GA4 apart from Universal Analytics, at least the free version.

MC: Before we started this podcast, I put out some LinkedIn posts, some tweets, asking the SEO and PPC, and the general digital community, if they have any questions for you both. I'm just gonna throw these out there, feel free to pick one and whether one of you is better to answer these or not. They're pretty random, I filtered them but we'll just see what we got. First question from the community is from Glenn Van Der Linden and he says he would like your view on how GA4 redefines the analytics market, which could previously be seen amongst other point of views as a split between marketing e.g. Adobe GA and product analytics e.g. mix panel amplitude.

KS: I'll go ahead and jump in here. I've been known to say that i think, over the years, Google has done a pretty poor job of marketing Google Analytics as a tool that can do both marketing analytics and product analytics. You could absolutely do product analytics with event tracking and various other features within Google Analytics, I've done it myself as a practitioner, but they didn't market themselves that way and that left a hole, or perceived hole, in the industry for a lot of additional tools to start marketing themselves specifically as product analytics tools. I think the big difference that GA4 brings is that it actually is that event in parameter data model that most of these other tools also operate on, and so it is now directly pushing into that space while also maintaining Google's power of the Google marketing platform and their ad stack for the marketing side of things. I think it's a really interesting move where they really could become that tool for all of the teams within the company, if there is that desire.

MC: Brilliant. Second question we have from Saijo George is “what are the alternatives you folks recommend over switching to GA4?”

SA: I can take a jab at this, whatever works. This is a good time to shop around. I'm a firm believer in GA4, I think it's a great tool for many purposes but, if you do want to remove yourself from the Google stack, which I think many companies have the incentive to do, there are lots of tools right now, just so many, and you just need to choose what is your focus. If you want something that you can build by yourself from scratch, from data collection to processing, to reporting, and have full control over it, then tools like Snowplow Analytics come to mind for the do-it-yourself crew. If you want a more “from the ground up”, privacy-focused alternative, there are lots of privacy-first analytics tools. There's tools like Matamo or Piwik which, in my view, resemble what GA was maybe 10 years ago, so they're not, from a feature point of view they're not that interesting, but they do give you the ownership of that funnel. I'm not dissing the person asking this but I don't think this is a very good question, because we're talking about tools and not what those tools are needed for. I don't recommend anybody switching to GA4 nor do I recommend anybody switching over to Snowplow until I know exactly what their need is, and what their purpose is. There's no generic recommendations here to be given and I think that, in my view, I think the less you focus on the tool and more on defining the problem space, the easier it will be for you to choose a path for whatever your company wants to do with data.

MC: So the right tool for the right job?

SA: Yes.

MC: I'll add in my own question from that, which is, do you think it's becoming less realistic for companies now to step outside of the Google-owned analytics space? With all these, so Krista mentioned the ad stack that's with Google, it plays really nicely with Google Ads and there's a bit of extra work isn't there, if you use something like Matamo, hosting it yourself, trying to get everything to play. Do you think that's becoming less of an option now? Are there use cases where you might choose these other tools becoming fewer?

SA: If it is becoming less of an option then it's a very worrying situation, because then it speaks to the monopoly antitrust claims that have been thrown around. I do think there's still a considerable population who don't care about the integrations they just want to do that first-party analytics really well, like behavioral analytics, and they might want to integrate with other tools like, let's say Hotjar, A/B testing tools, Chartbeat, whatever, and then the mobile stack as well. For those they might want a bit more open APIs than what Google can offer, and maybe a bit more transparency to what happens behind the scenes, after the data is collected, than what Google can offer. I do think that Google has a very very alluring stack for the enterprise, I think it's very very tempting to jump in that bandwagon, and I think for enterprises it really can be difficult to think about anything else, simply because of those integrations. But they do still represent the short tail of analytics users, even though they might have the most dollars.

MC: It's a very complete and diplomatic answer, I like that. I have a very specific help question as well. This is from Frederick Palella who says “delay in real-time reporting, we see that sales are not properly coming in when looking at today's data. We notice that GA is showing more accurate data when looking at today. The next day GA4 shows the same data as GA, meaning there must be quite some delay in the reporting of GA4. Are we the only ones experiencing this? For sure, during peak periods like Black Friday, it's crucial to have a good understanding of your data during the day itself. Thanks.” There's a really specific one. I just thought I'd throw that out there, see if anyone can help out Frederick.

SA: Krista, can you please fix that delay?

KS: I think there's a couple things to unpack here. I'm not sure if he's actually experiencing a bug or not. I haven't necessarily seen the same delay but, if you're looking at the actual real-time reporting, it's important to know that it works a little bit different than Universal Analytics. Universal Analytics is showing everything as the hits come in whereas, within GA4, it's actually showing a view of the last 30 minutes, so you're going to see different data between the real-time reports. In terms of what you're actually seeing processed into your processed reports, in either property I think that just really depends on hit volume and how fast those are really getting processed into there. There are SLAs in place that Google has, for the free product it is, on the UA side, 24 to 48 hours that you really could expect that data to come in and I'm not entirely confident, but I think the SLAs are probably similar on the GA4 side.

SA: Yeah, I think the old adage of not trusting today's data still holds, so I wouldn't hang my hat on today's data in the standard reports. If you need that good understanding for periods like Black Friday, then you can enable the streaming export for BigQuery and you'll get almost real-time data there. It's a patch but it's still raw data and you can build your own data studio reports with it or whatever.

KS: Yeah that's a good point. Sorry, I just had one other thing to add on today's data, which is that Simo mentioned is true, it's not necessarily something that you should always rely on and part of that has to do with, if you're an Ads customer, it's got to make that round trip out to Google Ads to see if there's attribution to apply there, and that can take some time as well, for all of that to process and so today's data is there, but it's not always complete until the next day.

MC: I think this is something that was interesting that always came out of, when we did basic Google Analytics training with clients, or people that are new to Analytics, talking to them about how that stack works from the collection point to data being processed to, when you're actually logging into Analytics, this is the final reporting layer, that everything else has already happened and, once you've set up those views, that's not a real-time filter you're applying, that's been processed and it's gone now. Just talking to that, the sheer scale of what's happening in the background and why the data isn't there two minutes later, all processed and done, like you say Krista, these round trips done. There's a really nice workaround patch you gave there as well, Simo, so thank you for that. I think that's a way better answer than I would have given Frederick, to just be more patient dude.

SA: That's a good answer as well.

MC: You've both talked about some new features, possibilities within GA4 and I think Vasili Dumas here has asked a question, which I think you can both answer, and we'll keep it to one, which, in your opinion, one of the most important features or the most important feature that Google introduces with GA4 compared to Universal Analytics?

KS: All right I'll go first. I've already mentioned a bunch that are my favourite but... they're all my favorites. I think, in my opinion, the one that I think is the best, not necessarily the best but the most useful, is actually enhanced measurement, because I think it opens up the possibility to track a lot more data out of the box, as Simo mentioned earlier, than has ever been possible before meaning that, if you are a long tail customer, and all you do is implement the GA4 tag on your site, you're getting a lot more meaningful data out of the box. While there's a lot of fancy new features that provide a lot of additional options, I think it's really great that the baseline of data is being bolstered within GA4, with enhanced measurement.

SA: My pick is not specifically... I'm bending the rules, it's not specifically a feature but it's a fundamental part of the platform. Just the shift to an event-based model rather than a session-based model.

KS: I knew you were going to say that.

SA: It's just so fundamental. Having an event or a hit as the core aggregation metric, rather than a session, is so important considering the future of GA4, because it enables Google to take really brave forays into something that's beyond the session. We don't know what that might be yet but the session itself has now become kind of a fleeting concept, it's just a parameter on a hit rather than, on Universal Analytics, it's the core aggregation bucket. You can't do anything without sessions, in fact if you want to do something with raw data, you first have to unravel those sessions into raw data. It's so significant and I'm not sure if people have yet grasped how significant this is, for Google's analytics tools because it really underlies the entire roadmap as well, not just what we see today in the reports.

MC: I certainly haven't grasped how bigger that changes yet.

SA: Then I wasn't dramatic enough.

MC: I actually had a question for you later on which was for either of you, I've been reading up about the differences and the fundamental thing is, this event data model has changed. Do you either of you want to give a little bit more of a beginner's intro as to how and why that's changed, and how that affects implementation, because I think lots of marketers that have been using G app until now are very familiar, like you say, with “okay sessions, what do we do on the website, we've got categories, actions, labels, yep. Okay, we'll make these goals”. It's all very embedded in what we do, so how should we be thinking about this? I did see a quote, when I was reading some other things you'd written Simo, about how you didn't accept that we should have this old event model and we should be open-minded to this completely new approach.

KS: I'll start with a high level and Simo can jump in with more detail. In Universal Analytics, you have, as you mentioned, pageview, sessions, all of those kind of things, and you have events, which are a unique combination of category, action and label. You can have implementations with hundreds, or even thousands, of these events because every click that you want to track, every interaction that you're tracking with events, is a unique combination of those three dimensions. I myself have done a lot of implementations with so many events and this paradigm really flips itself on its head in GA4, where events are now reusable names. So you might have something like a property booking app, say a booking.com or hotels.com, and somebody searches for a property on there. When they go to actually view that property you could send an event called “view property”, for example. You would send that exact same event called “view property” for every event that they're viewing, but you would have parameters to differentiate what they're actually viewing. You'd have a parameter of property name with a value of that actual property name, parameter of property rating, again with the value that's being collected. The parameters are what is going to distinguish that. I think what's really important, and needs to be grasped by people, is that there are now new limitations around the number of events that you can have. In the free tool right now, I think it's 500 unique events that you can have and, if you're thinking in the mindset of Universal Analytics, where you just had a unique combination for everything you wanted to do, you're going to very quickly run out of those unique event names. So you really have to take some time and rethink how you want to go about using your event tracking. Once you do that, I think the possibilities, as Simo’s mentioned a few times now, are significantly greater, in terms of what you can do. All of a sudden you have a much more structured data set that you can now model on, for example.

SA: That was a really difficult question and I can't imagine answering it any better, so i'm not going to go any deeper.

MC: I really appreciate the honesty here as well, because we've got two people that really know analytics inside out, so I'm glad we're getting feedback like this. If anything, more importance on making these measurement plans and carefully working out what it is you're trying to track, rather than just keep adding events on the end of what you're doing, as I've seen so many times happen in GA login and “what's this event?”, “I can't remember what that one was”. I'll jump back in, I've got one more question from the community around segmentation which is, “how does GA4 handle segmentation and is it likely to limit the amount of in-depth analysis possible?”

KS: Go for it, Simo.

SA: I’m just gonna mention the Analysis hub here. GA4 does let you filter tables and it gives you tools to build audiences, so it revolves around the concept of an audience. Similar to how segments work in Universal Analytics, you can choose to show data in a table based on an audience, for example. I'm not entirely certain that's where the cool stuff happened; segmentation is absolutely fundamental to Universal Analytics. It's what makes it tick, it's the way that you take the session-based model and start building these really complicated questions around it. So it's an absolute necessity in UA whereas, in GA4, because you don't need these kind of shenanigans to access that hit stream data, because it's already there thanks to the event model, the cool stuff starts happening when you start building funnels and we start building pathing, and for that you need to use the Analysis hub which, just as a little sidetrack, if i'm not mistaken the Analysis hub was a feature of Universal Analytics as well, but restricted to paid accounts, right?

KS: Yeah.

SA: Again, one of these things that we have been afforded for free now. I'm throwing the term Analysis hub out there, i'm waiting for Krista to pick it up and do something cool with it, because this is one of her little children as well.

KS: Yeah it is one of my little children. You mentioned pathing, it's near and dear to my heart. When I became a product manager on Google Analytics, it was well known within the Google Analytics team that one of my goals internally was to kill all of the flow reports within Universal Analytics because, in my opinion, they're pretty terrible. Apologies to anyone who uses those. Because I said that so often, I got handed the task of rethinking what pathing actually looked like, so pathing in the Analysis hub was definitely my baby, I helped build the first versions of that. I'm biased here, obviously, but I think it's a pretty decent tool for actually starting to see how users are going through your sites. I think it's a really exciting enhancement, specifically to GA4, it's not available in Universal Analytics. I've had personally multiple clients who have told me, “this is a really cool thing, I'm really interested in it, how do I get it universal?” and my answer is consistently “well, you're gonna have to start using GA4 to be able to actually use this, and it's one of those features that are really enticing enough that has gotten people to start to look into GA4 more.

MC: Wow, we've covered a lot already. I've got more questions from the community but I'm going to cut it there because we've already got to 40 minutes.

SA: Just tell me one thing, did nobody ask “are we going to get motion charts?”

MC: I didn't have that, no. Is that something you've been asked a lot or…?

SA: No, I'm the first person in history to utter those words, when you look at motion charts in GA4.

KS: I'm pretty sure there's a great YouTube video of Avinash, and a few other old-school GA-ers singing about motion charts.

MC: That's definitely something for us to search for then, at the end of this. Thank you both, Krista and Simo, so much for your time. We've covered loads, from Analysis hub to talking about events limits, this whole events-based model, what we can get with the base snippet, and some really high level stuff as to why people might want to think about using GA4. So, both, thank you so much for your time, I really appreciate it.

KS: Thanks for having us.

MC: We are going to be back in one week's time, which will be Monday the 7th of December. If you are enjoying the podcast, please tell a friend about it, subscribe if you like, link to us, anything helps, and I hope you all have a brilliant week. Bye.

More from the blog