Dan Antonson: The Power of Google Tag Manager

IVAN STEGIC: Hey everyone! You’re listening to The TEN7 Podcast, where we get together every fortnight, and sometimes more often, to talk about technology, business and the humans in it. I’m your host Ivan Stegic.

My guest today is Dan Antonson, who is a marketing technology consultant for Blueprint Metrics. Dan’s been instrumental in helping me understand Google Analytics and Google Tag Manager, and today we’re going to spend some time talking about a project we’re working on together for one of our clients. In a previous episode of the podcast, we talked to Megan Glover from Solhem Companies, the environmentally friendly real estate developer in Minneapolis. You should give that one a listen when you get a chance to better understand them as a client. Today, we’re going to dig a little deeper into the technology that we’re implementing for them from a marketing perspective.

Hey Dan, welcome back to the podcast. It’s good to have you.

DAN ANTONSON: Hey, thank you. Good to be here.

IVAN: I’m looking forward to talking about Google Tag Manager.

DAN: And I am as well.

IVAN: [laughing] Okay. I want to start at the ground level. What is Google Analytics and how is it different or the same as Google Tag Manager?

DAN: That is a great question and a great place to start. So, Google Analytics is an extremely popular web analytics tool. It’s free, it’s offered by Google. And you could find Google Analytics on just about 95% of websites that are out there.

IVAN: Ooh, that’s a lot.

DAN: The market penetration for Google Analytics is incredible. Google Analytics originally came out on the market in 2006. It might have been 2005 or 2006, but it actually came out from a company called Urchin, who is one of Google’s early acquisitions was this young web analytics company. And they acquired it and transformed it into Google Analytics, and it’s kind of amazing because a lot of the same backbone, a lot of the same concepts, like the idea of a website visit and hits and events and page views and traffic sources is all kind of still there. So actually, and a fun fact for those of you who are familiar with Google Analytics, there’s this thing called a UTM, it’s what you’ll see in the website URL. So, if you click on an ad, you’ll see UTM Source, UTM Medium, UTM Campaign, and “UTM” actually stands for Urchin Tracking Module. It’s one of the remnants from Urchin. So, it’s been a long time.

IVAN: Oh.

DAN: The more you know, Ivan. The more you know. [laughing]

IVAN: Yes, the more you know. [laughing] I knew the “UA” in the analytics number in the source code, that’s Urchin Analytics. But I didn’t know about UTM.

DAN: Oh yeah, that’s right. That’s exactly right. So, it’s kind of funny to see some of the leftover remnants, especially as you get into the guts of Google Analytics. Don’t get me wrong, it’s certainly changed a lot since it’s come out, but a lot of the same kind of inherent data model, again, the idea of a visitor in a session, that is all Urchin.

IVAN: Wow. So, you were talking about analytics and you were describing the percentage of the market, and then we were going to talk about what Google Tag Manager was, and I think I interrupted you. I apologize for that.

DAN: Yes. No worries. Google Tag Manager is a much newer product than Google Analytics. Think of Google Tag Manager, obviously tag management is in the title of the product, and that’s really what it is. It’s a tag management system, and so, Google Analytics, if you were to install it, by coding it you would get a snippet or a website tag, and then you would drop it into the website, versus Google Tag Manager, it’s really a container of those tags. It’s a management system. So, think of it almost as a content management system for your marketing technology infrastructure.

Another way you can think about Google Tag Manager, think of it as more of a router. So, Google Analytics is going to be a collection system, Tag Manager actually doesn’t technically collect any data. It’s not a storage mechanism. Instead, what tag management systems do, is they route the data to the respective products. So, Tag Manager can send data to Google Analytics, you can also install advertising pixels and trackers through Google Tag Manager. But again, Google Tag Manager isn’t storing any of that information, it is just passing it to the endpoint, like a Google Analytics-type product.

Google Tag Manager is really a game changer. Back when I got started in digital analytics, 10 years ago now, I was so much more of a developer than I was a marketer. I spent so much time coding into the actual websites the buttons I wanted to track, the things that I wanted to measure, and Tag Manager has really kind of flipped that whole script on its head and really allowed marketers to focus on making use of the data instead of collecting that data, or spending all that time, effort and energy collecting that data. So, it’s really transformed the way I work. I’m just continually impressed by how far along Google Tag Manager has come. It’s really quite an impressive product. But hopefully that answers your question, Ivan.

IVAN: I think it does. So, I am now going to think about it as something that allows me to enable analytics and control the data that gets acquired on a website from the Google interface, as opposed to doing it directly on the sites and requiring some sort of developer smarts to do that.

DAN: That’s exactly right.

IVAN: I love that you called it a content management system almost. That makes a great deal of sense. So we’re doing, as I mentioned in the intro, this project with Solhem Companies. At a very basic level what are we trying to solve for Solhem?

DAN: Solhem’s an interesting client just because they have a lot of different website properties. So, if you’re familiar with them, they essentially have a website for every single property. So, it’ll be juliampls.com and then there’ll be tulampls.com, and in that environment, analytics and tracking and really marketing technology become a little bit challenging to scale. You have to go into each individual site and add a snippet for Google Analytics, you have to set up unique accounts for each one of these properties.

And the specific problem that we’re solving with Solhem is, we are setting up a single Google Tag Manager container, or a single bootstrap if you will, that will literally run on every single site, but also enable the right domain to be routed to the right place. So that way, instead of going into each site one-off, we have one central content management system for marketing technology that we can interact and track all of these sites without having to reengineer for every single site. It’s a really unique situation.

Most of the projects that I work on, it’s a single website, or there might be a subdomain or two, but it’s one experience. And instead with Solhem, we’re literally tracking 12 different sites, and we’re going to be tracking 12 different sites with one container, and that makes it really unique.

IVAN: Just so that I can turn it around and understand it  from a business perspective, Solhem has buildings that all are created and developed consistent with their own values as a company. And so what they see from a tenant perspective is that once a tenant leases in one of their buildings, they actually sometimes find tenants that will re-lease in a different building, because it’s more convenient in a different part of town, but they will specifically seek out the other building because they’ve had such a good experience with the building that they first chose.

But from a technology perspective, we actually don’t have any way to know that that’s happened on the website. And further to that, even with tenants that haven’t even leased with Solhem yet, we have no idea if someone is visiting one of their properties in one of their buildings and is then also comparing them to other buildings of their own, that they still own. So, we don’t have that cross-domain information. I think what you said was the client will now have that, or potentially has the ability to then see which people visit more than just one building and more than just one website.

DAN: Yeah, that’s exactly right. And to get into the specifics of why and how, that’s challenging. So, Google Analytics is a first-party cookie tracking system. What that means is, when Google Analytics initializes on juliampls.com, it stores the session data or the client ID—essentially you’re kind of not thinking of it as an anonymous visitor ID, it stores that at the first-party level. And so the site that creates the cookie is the site that the cookie is stored on, and that’s what’s known as a first-party cookie.

Third-party cookies are the opposite, where a website will create a tracker on a website that it doesn’t own, and third-party cookies are starting to go away. More and more browsers, for instance, with Safari and Firefox, they’ve essentially said, “We are not allowing third-party cookies anymore.” So, what makes Solhem so challenging is that every site that has Google Analytics running on it, juliampls.com is creating first-party cookies for itself, tulampls.com is creating cookies for itself. What we’re doing with Google Tag Manager is, we’re setting up Tag Manager to create links between those sites. So, when somebody for instance clicks from juliampls.com to the Solhem property site, back to tulampls.com, that information gets passed in query parameters, so we’re decorating those links between the sites.

And then on the back end within Google Analytics, we’re stitching all of that information together, so we get one consistent view of that potential applicant across all of those different sites. That’s really exciting, obviously, because the benefit is that we can start to see, what are the properties that are being compared? What features do they have in common? How do we create a better product or a better experience for the potential applicant? And that’s just visibility that Solhem has never had before.

IVAN: Yeah, I’m looking forward to seeing the outcome of all the work that you’re doing right now at some point in the next few months when we actually start collecting data and when we can start turning the data into knowledge.

DAN: Absolutely.

IVAN: So, where did you start? We have a number of sites, as you’ve mentioned, and my understanding is that we had different Google Analytics accounts, one on each of the properties, because that just made sense to do at the time. Now you’re trying to weave all of these accounts together, put all this information together. How do you start even thinking about how to solve this?

DAN: [laughing]. That’s a great question. Like any normal developer I started by hitting my head on the desk a few times, you know, What did I do? Why did I sign up for this? [laughing] And then I found myself on Stack Overflow and realized that nobody had tried this before and then I hit my head even more. No, I’m kidding. So, for me, like I said earlier, Tag Manager’s a newer product, so unfortunately—there’s resources online, you can find people who’ve pieced things together. But for me, what I’ve always found helpful is when in doubt, just try something.

And so for me, what was really nice about Solhem was that there was no Google Tag Manager infrastructure. So, I didn’t have to necessarily design with something else in mind, I could really start with a blank slate. And since Google Tag Manager, like Google Analytics, is a free product, I was able to spin up a single sample container that I could just build for myself. What makes Google Tag Manager really slick—for those who are listening who haven’t checked out the tool yet—is, like a content management system, you create a tag or a series of tags, you create the rules in which you want the tags to run, you create custom JavaScript to apply linkers. But you can do that all without actually pushing to a production environment, per se. You can run it in what they refer to as just a “preview” mode where just in my own local browser, I can hit preview, and it’ll put a little window in the bottom of the actual website itself, and I can go through and test and tinker and make sure things work.

Because you have the ability to do that, that rapid prototyping, that’s really where I started was just, Hey, here’s a single container, I’m going to try to run this across two sites, and let’s see if I can make a rule that works across anything. And just within a weekend, I had this makeshift prototype ready to go. And then from there, I cloned things and brought in stuff that I knew would work, and from that shell is really where the foundation came from. But what made again, Solhem really advantageous or just a lot easier is you don’t have this technical debt that you have to account for, I was, again, really able to start with a blank slate, which are my favorite types of projects, so it’s been a lot of fun.

IVAN: [laughing] I’m glad to hear it. So, now we know where you started, now we know what we’re trying to do on these many different property sites. And add to the mix now the tool, the software that Solhem uses to actually manage their business and to manage their tenants and their leases. A tool that is made by a company called Entrata, and they effectively use it as a CRM. And now they have Google Analytics and Tag Manager capability as well, so it seemed a natural fit for us to track and to look to see what happens when visitors are outside of the Solhem property and on, say, one of the property portals that Entrata makes. So, let’s talk about how you’re managing to accomplish that transfer, and what the benefit of doing that is.

DAN: Sure. So, we’ll use the juliampls.com as an example. That website is a Drupal site, and when a prospective applicant wants to go one step further and actually apply for a lease or get more information, there’s an “apply now” button. And when you click on that it brings you away from the main site, from juliampls.com into what they refer to as a “prospect portal,” so it’ll be juliampls.prospectportal.com. What makes that again challenging, back to this whole first-party, third-party cookie thing, is that prospectportal.com can’t read the cookies that are on juliampls.com, so there’s no way for me to stitch those things together. Unless you use what I referred to earlier as a “link decorator,” where we’re actually using a query parameter to pass information.

So, what we’re doing is, with that same Google Tag Manager container, we’re dynamically appending those query parameters, so when a user clicks on “prospect portal” it strings that information, it transforms the URL that they’re actually going to, and brings them to the prospect portal page.

And now, Entrata, we are able to work with them, they were able to let us install Google Tag Manager on the prospect portal. And what that allows us to do again is to stitch the user behavior, or that journey, from one site to the other. So, now for the first time, Solhem actually has the ability to see, ok, of the people who look at the floor plans, what percentage of them even make it to the apply page? And when they get to the apply page, how many of those visitors are actually taking action? Are there parts of that experience that people are getting stuck on? Or, how can we make that experience better? So, that’s visibility that we didn’t have before, and that’s been a big part of this project, again, is just stitching all of these different sites together, but again designing so it’ll work across all of Solhem’s sites and digital experiences.

IVAN: So, that sounds like one of the use cases, actually getting marketing data and analytics information from their property websites and into their CRM. What are the other use cases?

DAN: Sure. So, the first use case of course is just tying experience together, but now that we’re able to pass information from juliampls.com into the prospect portal, what we have the ability to do now is pass information directly into the CRM itself. So, of course, Google Analytics is going to measure people moving between the sites, but that information’s just within Google Analytics.

So, one of the first things that we’re working on is how do we pass source information, keyword data, campaign information, landing page information. How do we make that jump from Google Analytics into the CRM? That’s been really—just in my own experience within digital analytics and marketing—don’t get me wrong, I certainly value the information that comes out of Google Analytics, but the reality is that Solhem is running their business out of Entrata. They’re not, nor would I argue that it should be, Google Analytics, so a big part of this project has also been integration. How do we pass that information back and forth, because we want to help Solhem align their marketing data to their actual outcome?

So, did this person actually apply? Where do actual leases come from? Where do the applicants who aren’t the right fit, how do we optimize, how do we invest less on the marketing front end to enable better outcomes downstream? So, there’s two pieces to the CRM. It’s how do we make Google Analytics work better, but also, how do we bring that information to the CRM? That’s the number one use case we’re thinking about, but there’s certainly other opportunities as well.

IVAN: We had talked about look-alike audiences as well. Can you describe what a look-alike audience is and how that might be useful?

DAN: Absolutely. So, Solhem already knows who their customers are, what their leases look like. They already have information about them, they know who they are, and one of the things that we’re also looking to enable for them is to allow them to take their first-party data, the names, emails, the information they already have and push them back into the platforms. So, we’re saying “Hey Facebook, here’s what our leases look like. Here are the people who are already renting from us. Help us find more people that look like this.” And that’s what a look-alike audience is.

We can do that from Solhem’s own data, but we can also do that from the behavioral data on the website as well. So, for instance, if we’re able to feed an audience into the Google platform where we say, “Hey, here are all of the people who are active within our Entrata application. These are the people that are thinking about leasing, are leasing or have already leased.” And we can take that information and actually feed that back into Google, and we can say, “Hey, again, here are the folks that are doing the things we want. Help us find more of these.” Then the Google machine learning AI kicks in, and their platform helps us find more of those people. So, as part of this build, we’re looking to really enable that help Solhem find more audiences that kind of fit their criteria to again, help grow their organization.

IVAN: Sounds like a big black box when you say, “Hey Google, help us find more audiences like this one.” Do we know anything about how it works or is it actually a black box like I mentioned it to be?

DAN: [laughing] Yeah, there absolutely is a black box. It’s right in the lobby of Google when you walk in. No, I’m kidding. [laughing]. So, Google and Facebook—it’s a really interesting time for privacy now. Certainly I, as a practitioner, and somebody who is also on the web, I certainly want to enable and to be a good internet citizen, and privacy is really important. Now, with Google and Facebook, who have both, especially over the last few years, really come under fire for some of their practices, but the stance that, at least in the current state is that Google and Facebook are saying, “We will not give you the actual information about the people using their platforms.” So, there is no way that I can go to Facebook and say, “Hey, here are my customers, please give me a list of names, phone numbers, emails and addresses of the people that you think are look-alike to my audience.” Facebook doesn’t do that, and neither will Google.

However, when we say look-alike and feeding it back to Google, what we’re saying is, “Hey, please take our data, join it with the data you already have and help us connect with those people through the advertising platforms.” But at no point do we ever get to see who those individuals are. Those individuals still, of course, can one: opt out of advertising completely; or two: not take action. No advertisement is necessarily going to force you to sign a lease or buy a product. But what we are saying is, “Hey Google. Hey Facebook. Here are the signals that we have. Please help us find them.” But knowing exactly how they’re doing that, like what is Google targeting specifically, that’s the part that certainly does become a little black box. We have no idea how they’re doing that or exactly what is in that audience.

IVAN: It feels like everyone’s doing this. Is that an accurate statement? Or is this special somehow?

DAN: It’s a good question. It’s hard to say. Put it this way: like I said earlier, 95% of websites have Google Analytics applied, and if the website owners who are managing Google Analytics wanted to create these types of audiences to remarket to create look-alikes from, you certainly could do that. But are all of those websites doing that? The answer is obviously no, right? You have to go through the time to set those up and to be smart about it, and then at the end of the day, those website owners, they have to pay for it.

So, is it everywhere? Well, just considering where Google Analytics is and every time you see a “like” button or the Facebook “share” button, that’s hooked into the Facebook platform, so, yes, it certainly is everywhere. But you also have a lot of control as a consumer in terms of what you’re going to opt in or opt out of.

By using Facebook, you’re obviously opting into sharing your information with Facebook and trusting that they’re going to be responsible for that, which is certainly up for debate. But you can certainly opt out of those advertisements in the Facebook settings. You can certainly do that within the Google settings. Where things get a little bit more fuzzy in terms of what you can and can’t control is when you start looking at the Experians of the world. You know, Experian sells household information to advertisers both digitally and on TV. If anybody knows of a way to fully opt out of that let me know. But it’s everywhere. It’s tied to credit. It’s tied to these databases. It’s all there, but in terms of how it’s being operationalized, that’s really up to the owners of the products or the people spending money on the platforms.

IVAN: I think the thing that we’re trying to do for Solhem is to allow Solhem to learn more about its own consumers and its own demographic, so that it can become a better company, so that it can offer better services and build buildings that are even better than the ones they have right now. And, also so that they can grow. I think this gives them an advantage. I’m always trying to figure out where that very fine line is of doing something that is privacy-enhancing and not privacy-busting, and doing something that is beneficial to the consumer, but also to the clients that we serve.

DAN: Sure. There certainly is a fine line. I feel, at least in my own experience with analytics, most of the organizations that I’ve worked with, exactly like you said, they want to grow, they want to do it responsibly. And I think at the end of the day the way I look at it is, Solhem wants to invest in what works and what doesn’t, and when a company is spending money on advertising and marketing and it’s not working, those costs of doing that has to go somewhere, and that usually gets passed to the consumer. So, if we can make Solhem smarter and make their investments work better for them, so they get a better return, the vision would be that they turn that into making their products, their experiences, their buildings, better in alignment to their mission.

Now, certainly on the flipside, I certainly understand the argument that while it’s just driving profits, and they could certainly put that right back in their pockets, and the consumer gets nothing. But again, if it wasn’t working, the alternative would be it comes out of the consumer's pocket. The other fine line is just what’s creepy online? You always hear that, like, the creepy advertisements that follow you around. And I certainly recognize that, but I think again, the goal for Solhem is to find their audience and just to be efficient about it, and it only feels creepy when it’s not the right audience. So, if we can make that more targeted and speak to the right person—like, Solhem doesn’t want to put an advertisement in front of you that isn’t going to resonate with you, so anything we can do to help them with that, I’m all for.

IVAN: You said that Google Analytics had 95% of the market, so, what’s the other 5%? Is there a competitor? Is there something that competes with Analytics and with Tag Manager? Or is Tag Manager basically the thing?

DAN: That’s a good question. So, the other 5% is obviously a mixed bag, and I would say the most common scenario we see is just nothing. There is no Google Analytics. There is no Google Tag Manager. If I had to make a bet on what the 5% is, that’s what it is. But that being said, the competitors out in the digital analytics and tag management space, there’s absolutely competitors. And in the analytics category, the products that you see are going to be things like Adobe Analytics, Mixpanel, you might be able to put a product called Segment IO in that category which helps you aggregate. They all have their own spin on it.

But what’s really ironic about those competitors is that, even when I work with a client who is using Adobe, there’s a good probability that they’re using Google Analytics as well. It’s free, why wouldn’t you use it? And on top of that, Google Analytics allows for a lot more Google-style integrations than you get with Adobe Analytics, and those products certainly come with pretty steep costs. Adobe Analytics, you’re not going to get into that product with less than $20,000, $30,000 a year type of investment. So, the tools are not cheap. Mixpanel is much cheaper but it’s a very different style of digital analytics. So, I would say, Google Analytics is certainly in a category by itself. Again, even if you’re using one product, you’re probably using that anyway.

IVAN: Yeah, that’s interesting. That’s something that I’ve seen as well with some of our clients who use Siteimprove. Siteimprove has their own analytics product, and clients that use Siteimprove also put Google Analytics on their sites as well, in addition to Siteimprove Analytics which, just because it’s free, I often wonder what the parity is like, and if they get the same or similar data between the two of them.

DAN: I’m not super familiar with Siteimprove’s analytics specifically, but just knowing what I’ve seen with other product comparisons, what’s really important to remember about Google Analytics is that it is a very marketing-centric product. It connects to Google Ads, and it is very much about measuring source, like where do people come from and tying that to a conversion. So, the whole data model behind Google Analytics is it serves marketing. Now you could use it for tracking other things, but it’s designed for marketing.

Now, products like a Siteimprove more than likely, they’re going to focus on things that are more focused on usability: error tracking, user satisfaction and user experience. More of a journey-mapping kind of experience, with very specific use cases. There’s analytics products that are designed 100% around tracking errors that happen, like server errors, JavaScript errors, any error you can think of. That is their sole purpose. Now, Google Analytics can do some of those things, but not to the same extent. A lot of the products that are out there, the Mixpanels, the Adobe Analytics, they all have their varying flavors, but it’s kind of more about the use case than it is, do they overlap or not? That’s how I’ve always thought about it.

IVAN: Well, it’s definitely been fascinating talking to you about Google Tag Manager and what we’re trying to achieve and how we’re achieving it with Solhem. So, I’m looking forward to seeing some of the reports and learning something new about the sites come the beginning of 2020.

DAN: Perfect. Thanks Ivan, this was super fun. Thanks for having me.

IVAN: Thank you. Thank you so much for spending your time with me. It’s been really a pleasure as always. Dan Antonson is a marketing technology consultant for Blueprint Metrics, and you can find him online @danantonson, and that’s on Twitter. You’ve been listening to The TEN7 Podcast. Find us online at ten7.com/podcast. And if you have a second, do send us a message. We love hearing from you. Our email address is [email protected]. Until next time, this is Ivan Stegic. Thank you for listening.