Must-Haves in GA4 & GTM: Effective Data Tracking Strategies

Step-by-step guide to configuring custom dimensions, e-commerce funnels, and enterprise tracking pipelines.

Speakers: Zafer Yıldız, Ferit Berkay Kelecek

26 Nov 2024

Video Transcript

⏱ 28 min read• 4,895 words

Read the full speech or switch languages

— Hey everyone, welcome to ZeoTalks. Today, Ferit and I are discussing Google Analytics 4 and GTM. We'll cover best practices and current tracking trends. This is our second ZeoTalks on Analytics. We previously recorded a session available on YouTube and Spotify. Quick intro: I'm Zafer, Web Analytics Manager. Ferit is a Sales Specialist. Welcome, Ferit. Let's dive right into the topics.

— Sounds good.

— Some of these topics are pretty deep. First up: Why shouldn't we track Google Analytics 4 without GTM? Google Analytics 4 has multiple setup methods. You can hard-code it manually. Or use a third-party CMS integration. But those of us in analytics always prefer GTM. Why do you think that is? I'll toss the ball to you. Why do we prefer GTM?

— Sure. Why do we prefer GTM? The reasons are simple. It’s for configuration. GA4’s event structure works. It works for basic integrations. But it isn't detailed enough. You know that as well as I do. Can you do it without GTM? Yes, if the system is small. If you aren't tracking many events. If you only need users, new users, and sessions. You can send custom events directly to GA4. That works too. But why wouldn't you use GTM? Why not simplify your interface? Most third-party tools are GTM-compatible now. So, why wouldn't you use GTM? That’s the real question.

— It really comes down to that, doesn't it? Why wouldn't you use it?

— Exactly. Why wouldn't you? Everything is compatible. You can add so much detail. Why skip it? I want to touch on other tracking tools too. Let me ask you a question. Are tools like Yandex Metrica as capable as Google Analytics 4? Should you use GA4 without GTM? Or should you just use other tracking tools instead? It’s a valid question. Google Analytics 4 is the industry's top measurement tool. We all use Google. I think we’re the ones fueling its power. It’s a tool that constantly updates. It lives in the space where we search.

It draws power from its users. Even if you use Yandex Metrica, you still search on Google. You’re just feeding the same tool all over again. What’s your take on this?

— Like you said, there’s no better measurement tool. Google develops it. It’s the successor to Universal Analytics. It’s a much more powerful tool. I think back to what we did in Universal Analytics. You had to set things up beforehand. There was no event-based tracking. But Google Analytics 4 is event-based. We can manipulate those events however we want. Even without prior setup, we can build funnels or integrations. We can do it all through Google Analytics 4. We don't have "we can't track this" issues anymore. I can’t imagine digital measurement without Google Analytics 4. Sure, there are alternatives. Some tools do things that Google Analytics 4 can't.

For example, if you need a heatmap, Google Analytics 4 won't provide that. Heatmaps are a measurement method, right? They’re often lifesavers. So, what do we do? We turn to tools like Clarity. We use Hotjar or Clarity. We get what we need and move on. Honestly, we don't live in Clarity every single day. That’s just how it is.

— In our industry, these tools feel like supplements to GA4. That’s how I see it.

— Exactly. We have a powerful measurement tool here. The Google team works on it day and night. Sure, there are bugs. We complain about them too. Sometimes we’re frustrated, asking why they haven't fixed things or added certain features. But for the average user, it’s plenty. It still feels complex, though. I notice this often. Brands we support give us that feedback constantly. It’s complex because it’s feature-rich and built for customization. You have to tinker with it. You need to get your hands dirty to get real value.

— Exactly. It starts basic. But as you customize events, it gets powerful. It becomes a highly efficient tool. That’s what makes it so strong. I see it as the hub. Other tools just support it.

— For sure. You have to know how to use it. That’s our job. We’ve built our entire careers around this. But if you own a site or work in digital, data matters. You need to know how to pull data. Even at a basic level. It’s really not that hard. When Google Analytics 4 first dropped, I felt the same way. Why change? Universal Analytics was fine. Remember the early bugs? We were constantly questioning it. But honestly? I wouldn't go back to Universal Analytics now.

— I don't even use it. I caught the tail end of Universal Analytics. Honestly, Universal Analytics feels way more complex now. GA4 is easier to use. The naming conventions are simpler. Plus, it’s a constantly evolving system. They’re improving measurement in the background. It’s not just about visual updates. The tool’s core measurement keeps getting better. Constant updates are a huge plus.

— Yeah. I don't want to bash Universal Analytics too much. We might be coming off a bit harsh. We have total respect for it. But we’re talking about a much more powerful tool. Think of it like a new phone or car. You get used to the upgrades eventually. You know our industry. If you can't adapt, you get left behind. I can say that with confidence. If we can't do something well, that’s on us. Obviously, nothing is perfect. Google Analytics 4 isn't perfect either. But I always look at myself first when I struggle. That mindset is key. Let’s pivot to GTM. You mentioned why we should or shouldn't use it with Google Analytics 4. That’s a valid point.

It reminds me of something. Remember that brand where we couldn't get Google Analytics 4 to fire? Google Analytics 4 just wasn't working. Everything looked right. GTM was integrated. We had the History Change setup. It was a unique site structure, sure. Remember that weird issue? Users from different cities weren't showing up. There was a glitch. We did the opposite of what we planned. We ditched GTM. We integrated Google Analytics 4 directly. We hard-coded it manually. We had to try it. Our daily work is just trial and error and debugging. It started working. We got data again. The brand had been dealing with this since the start of the year.

We don't just blindly rely on GTM. That's important to note. There's a nuance here. Google Analytics 4 actually recommends manual installation now. It's not just us saying it. Google Analytics suggests it too.

— Exactly. They want it added directly.

— Why is that? I couldn't find a clear answer in the docs. But clearly, some projects have complex source code. They run into tracking issues like we discussed. That's a big deal.

— Let me mention something we talked about recently. Clarity. Installing Clarity via GTM. I'm still having the same problem. I don't know if your brand has it. But my Clarity sessions via GTM are still undercounting. Adding GTM can cause issues with some third-party tools or Analytics. We shouldn't just dump everything into GTM. GTM is powerful, sure. But some things need to be hard-coded. That’s often the healthier way to go. Don't rely on GTM for absolutely everything.

— Exactly. I’ve had that Clarity issue too. Since GTM runs asynchronously, it can cause conflicts. It’s all JavaScript-based. Sometimes it blocks things during page loads. You might not see console errors, but issues still happen. That’s why debugging is critical. When we set things up in GTM, we use preview mode or extensions. We test forms and track events. We check page views, scrolling, and everything else. We monitor live data for anomalies. Sometimes it looks fine, we publish, and then it breaks. Then we’re back to troubleshooting the integration. I always prioritize GTM for third-party setups. But I don't trust it blindly for the reasons mentioned. That’s exactly why we’re here.

We check accounts to ensure everything works and tracks correctly. GTM is fine. But always double-check. Check everything multiple times.

— Let's expand on that. GA4 is a must. Whatever you build or launch, GA4 should be there. GTM should be there too. But I see GTM as a tool that needs more oversight.

— Exactly. GTM is my favorite tool. It even comes before Google Analytics 4. I can do anything with it. I can tweak the source code. I can inject custom JavaScript. Especially in this AI era, where our podcasts are thriving. You don't even need to be a JavaScript guru.

— It starts.

— Exactly. We can do anything. We used to write JavaScript and hunt for fixes on Stack Overflow. Now we have the power of AI. I just drop the code into Claude, find the error. And paste it back into custom HTML. Sometimes I'm shocked. I think, is that it? It gives me a day's worth of code in seconds. It usually works perfectly on the second try. That's why GTM, supercharged with AI, remains my favorite tool. If I need to set something up, I look at GTM first. Then I ask developers for help. Maybe a data layer. Or pushing something custom. GTM naturally captures what happens on the frontend.

GTM solves most of our problems without needing developers. Anything else you want to add?

— Nothing extra right now. We summarized how GTM and GA4 work. How to audit them. How to keep them clean. We need both. We covered how to run them. Yeah, we pretty much covered it.

— Exactly. Once the setup is done, our Analytics is running. But there are nuances. Like Google Consent Mode. Getting user data consent. These rules are getting stricter. You know Europe is strict with GDPR. We have to watch out for KVKK here too.

— We’ll be fully GDPR compliant by 2025. That’s the gist of it.

— Exactly. It’s the last quarter of 2025, right?

— Yes, the last quarter.

— Right. Don’t wait until the last minute. Brands always leave things to the end. Get ahead of it now. Be able to say your data tracking is solid. You must respect cookie consent. Yet, you still need those key insights. There’s a common myth. People think they’ll lose all their data. They think tracking will become impossible. That’s the big fear, right? It’s partially true, but only to an extent. That’s where Google Consent Mode comes in. We’ve mentioned it before. It’s a hot topic in the industry. Google has a real solution with Google Consent Mode.

With the right setup, you can track users in Google Analytics 4. Even without consent, you can send anonymous pings. How?

— It's like matching users. You're basically modeling the users who didn't opt in.

— Exactly. It uses machine learning to model data. We set it up and see data loss. Okay. We lost a month of data. We see this in cases. Google Analytics 4 says 28 days, right? You set up Consent Mode. Your parameters work in the console. Good. But you still see data loss. Your traffic takes a hit. Okay. But how long does that last? Google has a 28-day learning period. Exactly. But here is the catch. You need decent traffic volume.

— There are specific requirements. You have to meet those conditions. You need users to accept and decline. Your traffic must be high enough. I don't know if you've seen this, but I've seen modeling take even longer.

— Exactly. We saw it with a brand recently. It was a high-traffic brand. They work in insurance. They saw traffic loss for about a month and a half. Then, the modeling kicked in. It shows up in the reports anyway. It uses estimates to fill in missing data. It’s actually surprisingly accurate. We verified this with a brand. We matched their monthly purchase revenue. The variance was tiny. It looked like there was no consent management at all. Even though a perfectly working system was in place. We confirmed the data matches. So, what do we usually recommend?

It’s the brand’s call, of course. We don’t force it. But we provide the info. You need a cookie management tool. Google has partners for this. We work closely with Affiliate in Turkey. Buy an integration package from these platforms. You can get it running quickly. It’s not a blood, sweat, and tears integration. Some brands try to do it themselves. I recommend using these cookie management platforms. Google has recommended partners. They tell you to buy the service from them. They work correctly. Just set them up.

— They save you one or two weeks of effort. You can solve it in two or three days. Personally, I don’t think building it yourself is the right move. The implementation often misses the mark. Their expertise might be off. They assign a developer for this. They have to read guides. They have to handle integrations. It can go wrong. I don't want to take that risk. I usually steer them toward cookie management platforms. You backed that up too. Data is being lost. We want machine learning to recover it. If done wrong, you lose a huge chunk of data for months. That’s why using cookie management platforms is key.

But Zafer. I think having an Analytics expert who knows these integrations is even more important.

— Exactly. What do we see? There’s a banner. But does that banner actually work?

— Yeah.

— Right? A pop-up appears. Yes, no, manage preferences. I click yes. The same thing runs. I click no. The same thing runs. We see this even with huge brands. Just showing the pop-up isn't enough. It’s not enough for consent management. Use auto-block or manual block. It doesn't matter. Manage it via GTM or hardcode it. You cannot miss those cookies. Otherwise, there's no point in using it. It just sits there for show. Big brands rely on this digital user data. Data is vital for everyone. Budget should be a secondary concern here.

We see the budgets spent elsewhere. You must do this to respect user consent. It also minimizes your data loss.

— Brands must meet their legal obligations. It is the law. We cannot force-process data from users who opt out. That is simply wrong. Follow the rules and Google's recommendations. Then you will have no problems. We have tested this on many brands. When done right, the process runs very smoothly.

— Exactly. Consider that our public service announcement. Let's warn those who don't comply. Let me pivot to this. If they don't want to do it, there are cookieless tracking tools. This isn't just about tracking, by the way. We deal with measurement daily. Remarketing, ads, or any third-party tool. Everything needs consent. Because cookies are involved. But there are alternatives to Google Analytics. These tools make a promise. They claim to be a Google Analytics alternative. They track without using cookies.

— But how much of an alternative are they really?

— Exactly. That’s my point. Are they truly alternatives? In some ways, yes. We often suggest using them if possible. Why? Because you will face data loss. Think about Consent Mode and model learning times. If you ask about your October traffic, you won't know. That data just evaporates during the learning phase. But with these tools, you get better visibility. I won't name-drop. Just search for "cookieless tracking tools." You'll find 3 or 4 solid options. They usually have very basic dashboards. Unlike Google Analytics 4, they are quite simple. They track only the essentials. You won't see complex events or funnels like in Google Analytics. They mostly just count 1, 2, 3, 4, 5, 6.

But I find it useful. Forget about consent management for a second. Using another tool alongside Analytics is a lifesaver. Sometimes you lose data. Things break. Things change. You need a second measurement tool to fall back on. We see this all the time. Someone breaks something. They remove it from the source code. We can't track it daily. You end up losing three days of data. Something gets removed from the GTM source code. It’s just a snippet running there. A 3-4 line code block. Remove it, and the whole system crashes.

— It saves effort. Since we set it up via GTM.

— Exactly. Exactly. That’s why adding a second code snippet gives you parallel measurement. You can check it to verify if there’s an error. We need that sometimes. We ask, why is the traffic low? Maybe the traffic really did drop.

— But you can't verify that, right?

— Of course you can't. A brand says their traffic dropped. There are many reasons for this. Analytics never tells you why. We can't answer the "why" question. Maybe you cut your ad budget. Maybe SEO isn't going well. Maybe you were running ads on Telegram for direct traffic. Maybe they pulled them. Maybe your brand awareness dropped over time. There are a thousand reasons in digital. Traffic just drops. So to verify it, you look and see it dropped here too. Your traffic really did drop. Sorry, but it's not Analytics' fault. Don't blame Analytics. Your traffic is down. That's why, alongside Google Analytics 4, you need a cookieless tool.

— A basic measurement tool.

— Exactly. It makes sense to use a paid tool. It won't break the bank. It's a smart move to keep going with one. Yeah, it makes total sense. Anything else we missed on this topic?

— I don't think that's relevant.

— I want to dive deeper into Analytics. We're nearing the end. The first half-hour was about integration and setup. Now, let's look inside Analytics. Regardless of the sector, what do I usually see? I'm talking about essential custom events. Not the default Analytics setup. That works out of the box. I could put it on my personal blog. It would just sit there. But when you open a property, say for e-commerce. What broken features surprise you most? Maybe add-to-cart issues? Or something more critical like revenue tracking? What do you fix immediately in Analytics?

— When you take over an e-commerce account, check Purchase first. Some people say, "Forget other events. Just track Purchase." If you're starting from scratch with a broken account. Purchase must work. I'm talking about e-commerce here. After Purchase, I need other events to fire. I won't name just one. Track every checkout step. Are you having issues there? Where are you losing users? If you're getting traffic and users are adding to cart, you're fine. But if you don't track checkout steps, you're missing out. Purchase events should come first. Checkout events should be second.

You'll run into issues otherwise. If you can't spot the bottleneck, you're headed for chaos. That's why I'd add checkout events second. Then we can add all the enhanced e-commerce events. We can track everything like add-to-carts and product views. I have a question for you. Aside from enhanced e-commerce, what are the must-track areas for any site? This applies to every website.

— Are you asking about custom events? The ones we set up later? It depends on the industry. But for e-commerce, your Purchase event must work. You need to know what's selling. You need to know which categories drive revenue. It all ties back to Purchase. If you can't set up enhanced e-commerce, just set up Purchase. Even without a consultant, a solid Purchase integration gives you so much. Track revenue by category. See which sources drive sales. Check organic revenue or transaction trends. Look at your average order value. Go even further. Export this data to BigQuery. Generate new insights from there.

That’s why Purchase is the first thing on my mind. Data is everything in e-commerce. Purchase is critical. You know it’s something you can’t turn off in Google Analytics 4.

— Something you can't turn off.

— A key event. It’s not just for me. It’s the most important one for Google Analytics 4 too. But here’s the issue. Some people don't track it. Others track way too much. I’ve seen it inside Analytics properties. Add to Cart and Remove from Cart are all marked as key events. People are ignoring the macro versus micro conversion concept. We talked about Purchase. That’s our macro conversion. It’s a true key event. It’s a vital organ for the website. But it’s not the only thing that matters. I get it. Google Analytics 4 gives you a limit for key events. People still mark everything as one. I’ve even seen PageView marked as a key event.

You mark PageView as a key event.

— Just PageViews.

— Exactly. Everything runs as key events. Instead, focus on your KPIs. What’s the goal? Purchase, sure. Form submit, lead gen. Okay. But you also need micro-conversions. Like a newsletter signup. You’re selling products. But you’re also building a subscriber list. Exactly. That’s a conversion too.

— Or maybe you’re pushing for signups. Getting users to register.

— Exactly. A signup. Even a login in some sectors. It’s huge for user loyalty. If they keep coming back to shop without issues, that’s great. Tracking a login event as a key event is a smart move. That’s a good example. But like you said, session start and page view just clutter things up. It’s pointless after a while. Conversion is different. Only focus on what’s truly a key event for you. Don't clutter Analytics with useless data. Keep your reports clean with key events. Think about it. Someone else will look at those reports later. You’re just creating unnecessary confusion.

Sure, it doesn't hurt to have it in Analytics. It can stay. But it’s data pollution. It’s a disadvantage for BigQuery too. If you have high daily traffic, you’ll hit your event limit. Having too many events is the issue. Since there’s an event limit, BigQuery might just stop working.

— Also, consider the changes you mentioned. These affect your metrics. You can set goals for bounce rate or engagement rate. If you mark a page view or session start as a conversion. You’re messing with those goal metrics directly. Then you realize it’s wrong and remove it. What happens? All your goal metrics break. You have to reset everything. You end up disrupting the brand's entire workflow. That’s why picking macro and micro conversions is vital. Just like you said.

— Exactly. You need to define what matters most. Or what’s slightly less important. You can see the rest anyway. No need to redefine them. You have all that data as events already. That’s where it beats Universal Analytics. You can access data without defining it beforehand. I also wanted to mention GTM. Remember? We talked about this before. How should we NOT use GTM? We see so many bad examples. Everyone just does their own thing. No workspaces. The same mess I mentioned for Google Analytics 4 happens in GTM too.

Pages of tags. Endless triggers. Tons of unused variables. Folders everywhere. The naming conventions are terrible. When I see that, I know an amateur set it up. Not that being a beginner is a crime. Brands getting Analytics consulting should expect perfect GTM setups. You can’t manage what you don’t understand. But when we inherit GTM accounts from others, it’s a disaster. It looks like a war zone in there. That’s why I’m careful when I set up a tag. When I publish a version, I use proper naming. I always add the necessary descriptions. You might need to check a tag from months ago. You need to know when it was set up.

Why was it paused? Why was it published? You need clear naming to find these answers. With good naming. I can look at a five-year-old event and know exactly why it exists. Exactly. So, name your tags right. Publish them the right way. Use version names. Write long descriptions if needed. Sometimes I fix things quickly and publish. But I always use a good name. I don't always write paragraphs for descriptions. I’ll critique myself there. I don't always sit down and write long notes.

But I rarely publish a version without a name. Maybe only on my test sites while playing around. You need to make this a habit. That’s how you use GTM the right way. It’s a critical tool. It’s still not valued enough. Seriously. Many brands have been online for years. We’re the first ones to set up their GTM. Something is wrong there. If you’re doing digital business, you need client-side and server-side tracking. You should manage all your other setups through it. Brands just don't give it the importance it deserves.

— I want to highlight something here. You mentioned naming conventions. I’d add workspace management to that. I avoid the default workspace whenever possible. Clients often don't know better. They just jump into the default workspace to make changes. They push whatever changes they make directly from there. If you were working in the default workspace, your work gets overwritten. Any tests you were running are gone. My advice is to create a specific workspace for every task. I usually name them something like "Zero Workspace" or "Project X Workspace." Always do your work inside those.

You might have multiple developments going on. Always check the workspace overview before you push. You could accidentally push someone else's changes. I have workspaces with 70 or 80 pending changes. I check them periodically to see if anyone pushed anything. It’s that critical. Don’t work in the default workspace. Create your own. It’s just as important as naming conventions.

— Exactly. GTM gives you three workspaces. Use them. If you’re setting up Clarity, open a "Clarity Setup" workspace. Debug it there. Let third-party developers handle their own JavaScript in their own space. Let them run their tests. You run yours. Keep going until you're ready. You're separating things there. Workspaces are exactly what they sound like. You separate them. We don't push everything live instantly. I finish a task. I wait to push it live. I might publish that workspace a month later. Others can still work and publish. I’m not hogging the default workspace.

They can just open their own workspace and publish. Naming workspaces correctly is key. Even for tiny events, open a new workspace. Make this a habit. Every Analytics consultant and brand employee should do this. It’s a top priority for me. Tag naming is the same. We have a template. For every event, identify the tool. For example: GA4 - slider click. You know exactly what it is. You see they set up a slider click event. You’ll understand it even in 10 years. If I just write "click" or "image click," which image? Is it a slider image?

Is it a menu image? A product image? You can't tell. Same goes for buttons. Like a call to action click. Is it an apply button? Or something else? Keep it short and descriptive.

— You need to add detail.

— Exactly. Detail it so when you open GTM, it looks like poetry. You should know exactly what’s what. That’s why folder structure matters. We often skip folders, honestly. We just open a workspace and go. But use folders. Who built this? The Zeo Analytics Team. Events built by the Zeo Analytics Team stay in that folder. You don't even need version history. The folder name tells you everything. You know those tags belong to the Zeo team. You create a new folder for your own tags. This keeps GTM organized. Anyone can jump in and understand the setup quickly.

It’s not hard to do. Otherwise, you end up with a total battlefield. Everyone just creating random stuff.

— We can't figure out what's going on there.

— Exactly. It’s running. There are paused items. Why are they paused? Can we delete them? There are no answers. They’re just paused. They’re sitting there. Not deleted. I can’t delete them either. Maybe someone will turn them back on.

— We might actually need them.

— Exactly. I’ll ask someone. But there’s no one to ask. I can’t even see who set it up. It shows an email address. But I don’t know what the folders are for. We ask the brand. Sometimes they don’t know either. We say let’s delete it. We delete it. Then we might need it again. That’s why using GTM properly is important. As we wrap up, did we miss anything else?

— Not really. We covered how to use the tools. We talked about what’s right and wrong. We discussed if we should use GA4 or not. We covered GTM as well. We covered how to use both tools correctly. This session focused on the dos and don'ts. We tackled the most common questions. It feels like we've covered everything.

— Exactly. This session felt like looking through a brand's lens. We discussed Consent Mode and how Google Analytics 4 behaves. I hope you found it useful. Check out our measurement content on ZeoBlog. I highly recommend taking a look. Thanks for joining us today. See you at the next ZeoTalks. Take care.

— See you later. Take care.

This transcript was generated using AI speech recognition and verified against subtitle style rules.

Yiğit Konur

Content author

Yiğit Konur

both

Contact us

How we handle your data

Open as a full page