zh
English Español Deutsch Français Italiano Português (Brasil) Русский 中文 日本語

Digital analytics trends: a changing industry and what's next

English

Transcript

Introduction

Tristam: Good morning, good afternoon, good evening from wherever you are watching in the world. It's another SEMrush live webinar. It's April the 30th, 2020 and I am joined by the super knowledgeable Krista. I do hope I said that correctly. Hey Krista, how you doing today? 

Krista: Hey, I'm doing really well. How are you doing?

Tristam: I'm very well and I'm super excited to hear what you've got to say. It's going to be knowledge packed I know that but a sneak peek at the presentation already. I'm Tristam, I'm co-founder of Purple Smudge and I'm going to be your host today. And Krista, if you could just give us a little introduction of who you are. I'm sure a lot of people know who you are. But for anyone who doesn't, Krista, who are you? And why are we here today? 

Krista: I have been in the digital analytics space for a very long time. I spent almost seven years at Google doing analytics and optimization as a practitioner and then I was on the Google Analytics team, helping to build Google Analytics, helping you build the new app on the web, which I'll talk about today.

As well as being the external evangelist for Google Analytics. I did a lot of teaching, speaking, training, traveling all over the world to tell everyone about Google Analytics, and how to use it. Really exciting.

Most recently at the end of last year, I actually started my own digital analytics consultancy called KS Digital, that's where I am today. Based in my home office for now, while we are all staying at home, but it's been a really exciting ride. 

I'm consulting on Google Analytics for the most part and I'm starting to work with several clients on App + Web. That's part of what I'm talking about today. I'm going to be talking about some industry trends in research I did late last year and showing how that is actually leading into some of what Google is planning for the future of Analytics. All right. Just going to go ahead and share my screen and we'll get started.

Tristam: Perfect. Fantastic.

Krista: As I mentioned, today I'm talking about analytics trends and what's next. Specifically what's next coming from Google and how that is speaking to some of the trends that we're seeing. 

The first part of this presentation is going to be about industry trends. And this is a chunk of research that I did last fall. But I want to share some of the results with you guys. Then we're going to pivot from there and look at how Google Analytics and specifically App + Web which is the new type of analytics coming from Google, is actually speaking to some of these industry trends and where it's going. 

Digital Analytics Trends: Key Research Findings

Let's dig right into these industry trends. Now, last fall, I put out a poll on Twitter, and I asked "Hey, I'm doing some research for an upcoming presentation. What is your primary form of Analytics tracking?" And I broke it up into three categories.

To me, I was looking at major vendors, are you using an Adobe or Google Analytics? Are you using more of a niche vendor, an Amplitude or Heap or Mixedpanel, something like that? Or are you doing it in-house? Are you building your own, looking at log files, using a data warehouse or a data lake? Or some combination of that? 

Overwhelmingly, as you can see the answer here was, “I'm using one of these primary platforms in Adobe or Google Analytics”. This is my Twitter, I have a lot of Google Analytics people and fans who follow me. So, it's not surprising that that would be the main thing that would come here.

What I did next was I actually put together a Google form, or a Google survey, where it was much more detailed and had several questions. And not only did I post it on Twitter and LinkedIn and Measure Slack, but I actually sent it out to leaders in other analytics tools and companies and asked them to promote it to their audiences as well. 

I wanted it to be as representative and as global as possible in terms of the answers and results that came in. So, the first thing I asked was essentially the same question: What is your primary analytics platform?

The answers I saw here were really, not just a pure or in-house solution, but often a combination. They're using data output from a major vendor, but then doing their own thing with it, which is very interesting. 

I got deeper and I said, "Well, what is that major vendor that you're using?" Now unsurprisingly, a lot of people were using the free version of Google Analytics. And then a good chunk of people were actually using GA360. After that, a number of people, Adobe Analytics, and a couple of holdouts still using IBM and Coremetrics.

I was a little bit surprised in this data about GA and GA360 were actually fairly close in the numbers. What I think was going on here was that we actually had a lot of agency people who answered the survey. If they had a client with GA360, they probably just answered with GA360s and so it was just a 0/1 answer question. 

I also asked, "What is your secondary vendor?" If you're using a secondary analytics platform, what are you using? And here the answers were a bit more varied. Still, the major vendor is Google Analytics, Adobe Analytics, but we started to see some other things come through Amplitude, Snowplow, Snowflake, Mixpanel, Hotjar, BigQuery, a lot of other smaller things as well. 

Then I asked in more of an open-ended fashion, "What are some of the reasons that this is your secondary platform?" Why isn't it your primary platform? And I got a number of different answers. 

If somebody was using Google Analytics 360 as their primary and their secondary was Adobe Analytics, they said, "Well, my day to day job is done in Google Analytics. And Adobe is our legacy tracking system." Or, "There's a steep learning curve with Adobe. And that's why I prefer GA360." On the flip side, if somebody was primarily using Adobe Analytics, and their secondary platform was GA360, they said, "Well, GA360, is just used by less analysts." Or, "Maybe we're only using the free version." Or, "We're really only focused on the marketing audiences.

When the primary was Heap and the secondary was Google Analytics, one of the things that we heard here was that “product or product analytics is more important to us than marketing”. I actually think is interesting. Because while I love Google Analytics and spent a lot of my life actually working on and building that product and promoting it, I think that they've not done the best job of actually marketing it as a platform that can do a lot of different things.

You can absolutely use Google Analytics to do product analytics. I've done it in the large enterprise organization myself. But they don't market themselves that way. It's a marketing analytics tool. Whereas, a lot of these other niche vendors have actually come in and said, "Our primary goal is to be a product analytics tool." 

Then I asked, "Are you satisfied with your current primary analytics platform?" And for the most part, most people were fairly satisfied. You can see the average here around eight or so. If people were unsatisfied, they said it was because there was slow innovation or it was outdated, it needed more flexibility.

Then I asked, "Well, how likely are you to actually change your primary analytics platform in the next 12 months?" I think this is actually very interesting, because, for the most part, people were fairly satisfied. And yet, we see while, most people are leaning towards not changing platforms, there's actually a much longer tail here who is looking towards the possibility of changing in the analytics platform in the next 12 months.

Now, I don't know if you've ever been through an analytics platform change. I've been through many, I've helped many clients through them. It is not fast or easy or painless. There's a lot that has to be done to actually change your primary analytics platform. It's embedded in all of your tools and systems, and the organization itself is educated and knows how to use something. So changing that can be very painful. And yet, people were actually open to potentially changing platforms here. 

There's two major trends that I saw. One was either, people were moving towards in-housing or mixing their data sources, so potentially using a primary as a data pipeline, and then doing something else with that data. Or, people we're looking to streamline and go back to a single platform or something out of the box. Two diverging paths there that are ... I don't know if they're at odds with each other, but they're very different paths that businesses might want to follow these days. 

Then I asked, "Do you use a separate platform for mobile app analytics?" The majority of people said, "No." They use the same platform, or they didn't have a mobile app. But some people here did say, "Yes." They are using a separate platform. So I asked, "What is that separate platform?" Now, a lot of people said, "Google Analytics for Firebase."

Then I asked, "Are you planning to combine platforms to track both web and mobile app together at some point in the future?" And here, "Yes." Either they weren't using the mobile app, or overwhelmingly they were looking to go ahead and combine these into a single platform in the future. 

Of course, my next question was, "Well, what platform are you going to look at to potentially combine these?" A lot of people were like, "Oh, I'm interested in Google Analytics App + Web. Now, App + Web might be a little bit over-indexed. Again, because this is my survey. 

A lot of the commentary that people actually put in about why they were looking at App + Web, was, "Hey, this new thing just came out from Google, we're actually really interested to see if this can meet our problems. This could be the solution. We want to see where it goes." And so people were hesitantly interested and curious if this could be the right platform for them.

The last question that I asked was, "What trends are you seeing in the industry? Where do you see things going in the next 6, 12, 24 months?" The most common theme was around GDPR, ITP CCPA, essentially, how various regulations are making it difficult to track and use data, and this move towards a cookie-less world. People are concerned about this. It is a trend that is ongoing that people need to be aware of and thinking about. 

I've heard a few people say, they've built their own platforms and are now moving back to Adobe or Google. It turns out, maintaining a roll-your-own is hard. Segment, Snowflake, and Amplitude are pushing the trend towards more access to the raw event data, rather than living in the UI and schemas of the platform. 

I've heard this over and over again, throughout the survey responses where people were very interested in having access to that raw event data, or just raw data in general and frustrated that in some of the major platforms, it was much harder to get that data out.

The growth of both big data analysis and the power of visualization tools is making it easier to combine multiple sources of data to gain a clearer picture of digital marketing and the voice of customers. That is beyond what a single web analytics tool can provide. Again here we're speaking towards that combination of multiple data sources and using tools to be able to really visualize that and bring it together.

 I think this survey was very interesting. I enjoyed doing it, I enjoyed reading and analyzing the responses. I think it gives us some insight as to where the industry is potentially going, and especially this diverging path that we've seen in terms of either going towards a single platform and wanting it to bring it back towards a major vendor for ease, or wanting access to that raw data so that you can really manipulate it yourself and do something more in-house. 

Google App + Web Overview

Now I want to turn and look at what's next. And specifically, I'm going to focus on Google Analytics and Google Analytics new App + Web. It looks similar to what you've seen in Google Analytics all of these years. But it also looks different. 

If you're familiar with Google Analytics for Firebase, or previously, Firebase Analytics, it looks similar to that as well. I really want to look at, how does App + Web stack up to some of these industry trends that we were seeing in this data?

First, it combines multiple sources of data. You can have your app data and your web data and multiple different streams of data, as Google is calling it, coming into the same property in Google Analytics. It's called data streams. You can see my setup, you can add streams as you want. Next, it ties identity across platforms. So, not only can you do this by a device, but you can also do it by user ID. Right. It's really looking at how we can get that more holistic view of the user across devices as users move from desktop to tablet to phone. 

Next, and this one I think is really important. App + Web is fully event-based. Everything in Google Analytics App + Web is an event. Even a page view is considered an event. I have in parentheses here, like many of the product analytics platforms. Essentially many of these product analytics platforms have been for some time fully event-based. And Google is starting to get in on this trend.

That's where Firebase Analytics was in the beginning, and how it's evolved into App + Web today. And you do have access to that raw event data, via a free export with BigQuery. And this one is huge, right? People have been saying in this data and outside of this data for quite some time, they want access to this raw event data. They want to be able to manipulate it and analyze it the way they want to. When you're doing that in BigQuery, you can start to build your own models and get really, really sophisticated with this.

One of the comments we saw earlier about Snowplow, in particular, was that people were using it for data science and data mining. And you can do exactly that with this export into BigQuery. This I think is a huge game-changer for the way people are looking at Google Analytics data. And you can also go ahead and explore and visualize that raw event data, via Data Studio. So even if you don't have a strong SQL skillset to manipulate that data in BigQuery, you can go ahead and do that in Data Studio today. 

How App + Web Differs from Google Analytics

Now I'm going to talk about App + Web specifically, and what makes it different. What sets it apart from the Google Analytics that we all know and love today. 

First and foremost, App + Web is a new data model. As we've said, it is fully event-based. And it has three types of events that I'm going to go through. The first is automatically collected events. There's a bunch on the screen: first open or first visit, screen view or page view, ad click, user engagement. 

These are things that as soon as you either implement the SDK on app or put that code on your page with web, these are going to be collected out of the box. 

Next, we have recommended events or suggested events. Now, for these, there's a list in Google's help documentation in the developer docs of an extensive look at different verticals and what you might want to consider. The reason this is so important is because if Google can anticipate or know the types of events that you're going to send it, it can actually provide you a lot more reporting out of the box. 

You can still send your own events, that's not an issue. But again, think carefully about what might fall into recommended versus a custom event. I would suggest optimizing towards having more of that available in the reporting UI with recommended events if you can. 

There are some limits to keep in mind. Specifically, you get 500 unique events per property in App + Web. And event parameters, you get 25 per event. Now, this event parameter thing is new with App + Web, and I think it is very interesting.

You can essentially send 25 additional unique pieces of information with every event to Google. And I'm going to give you some examples of that in a second. But, 500 sounds like not a lot if you're used to using a lot of events in Google Analytics. But it is very different in how you're going to think about this. 

But I just want to lay the groundwork of these limitations so you keep that in mind as we go through. Then on the configuration side, you still have things like audiences. User Properties is a new feature in App + Web. These are similar to what you might have used as a user scoped custom dimension in Google Analytics, and a few other things on here. 

App + Web Implementation Best Practices

I'm going to talk about some implementation best practices. And this is where I'm really going to start to explain some of those limitations and how to actually think about this differently. So we have a couple of different examples here that I want to run you through. So the first one is, think about a property booking app. So something like hotels.com or booking.com or something like that. 

Maybe it's an app and you're looking on your phone or a website; doesn't matter. App + Web collects both. And let's say you do a search for property, like I'm looking to go to Amsterdam. I type that in the search box and I get a whole page of listings of different properties in Amsterdam that we might want to book. 

An event here that we would actually want to consider is, when somebody actually views one of those properties that have come up, and I'm using a recommended event here. There's a recommended event called view item. If I'm wanting to customize this, I might want to call it something like view property, but since I can use a recommended event, I'm going to go ahead and do that for the reasons I mentioned earlier in terms of making sure I have more available and reporting. 

If you've used event tracking in the current version of Google Analytics and the Universal Analytics, you'll know that it's a hierarchy and there's three pieces of that event to track. There's a category and action and a label. But each combination of category, action and label, is unique. So you use it once, and you don't use it again, because you want to be able to track that unique instance. That's completely different in App + Web.

Here, I'm telling you this view item event, this is one event out of your 500. But you're going to use it 20, 50, 100 times. You're going to use it every time somebody clicks on a property on that page to actually view. What's different about it, what distinguishes which property somebody is viewing or items somebody is viewing is going to be the parameters. 

In this case, as I mentioned earlier, you get up to 25 parameters per event. I can send something like the item name or the hotel name, the property name, the location ID of that property that I'm looking at, the number of nights that somebody wants to stay. I can send a whole lot of additional information with that one view item event. 

This is so much more than it would have gotten from the original combination of category action label within events in Universal Analytics. 

Let me go through one more example for you. Let's say you're publisher, maybe the New York Times. As a publisher, as the New York Times, you have a lot of content, a lot of articles on your website. And when somebody views an article, we want to send an event. We want to let ourselves know that somebody has actually viewed that article. 

We're going to send an event called view content. And the parameters with that event, in this case, will be something like the content title or the content author name, content author ID, the content location. A bunch of different things.

What's really important here is that as a publisher, you might have the same article showing up on your homepage, and the technology page, and the local Bay Area page for me. It could show up in a lot of different places. Before, you would have tracked that with just a page view that somebody saw this article but, it would have been much harder to actually combine all that data to really get an understanding of how many people saw the article since it would have happened on three potentially different page views. 

Here, we have the same view content event for all of it, we can easily then see how many people saw this article, which for a publisher's probably the more important thing to know. But you can still break it down by the page path where somebody was when they saw this, to understand the location that the article might have showed up in. 

I forgot to mention, but there are a couple of automatically collected parameters with every event, and that page location or screen name will be one of those. You'll always know where somebody was when they triggered an event. 

Okay, I know that was a lot, definitely throw in questions to the chat if you have them, and I can cover that in the Q&A on how this works a little bit more.

App + Web Unique Features

Now that I've walked you through how App + Web stacks up to industry trends and how to think about this and how it's different, I want to actually show you some of the cool new features that you're going to get with App + Web.

The first (and this might be my favorite, although I feel like I have a few favorites in App + Web) is something called enhanced measurement. This is really freaking cool. Because what it does essentially is if you click that little settings icon in the bottom right of that box, it opens up this panel. And it's all of these different things that you can toggle on or toggle off, to tell App + Web to just go ahead and track out of the box for you. 

Pageviews is what it's always going to track but, scroll tracking, for example. Scroll tracking used to be very difficult to set up. Then Google Tag Manager came up with a trigger that made that much easier, but you still had to set it up. Here, all you have to do is literally come in, toggle that box over to on, and App + Web is going to track scrolls for you. Same with Outbound links, site search, video engagement and file downloads. 

I feel like this is really democratizing the data more throughout an organization. Because previously, these all would have been things that required additional implementation or the use of a developer or somebody that really knew how to set this up, either hard coding the code on your site, or with Google Tag Manager. And now all you have to do is come in and toggle this little box. 

Speaking of Google Tag Manager, there are a couple of new tag types in Google Tag Manager to help you launch App + Web. The first one is the configuration tag. This is essentially where you would put your measurement ID, which today would be like your universal analytics property ID. In App + Web, it's called your measurement ID. It's just your little codes and a bit, your unique identifier that tells Google to send the data to your property. 

There's also a new metric called engaged sessions per user. This is pretty simple. It's just a calculation per user of engaged sessions that we just looked at. Then probably my favorite new metric is engagement time. I don't think time on page is an accurate metric and I don't look at it for my analysis. I encourage my clients not to either.

Engagement time first was a concept in Google Analytics for Firebase. And the way that it worked is essentially how much time in the foreground is somebody spending in your app. So it was much more accurate because generally, when you're on an app, you only have one app in the foreground. So we now have this new metric called Engagement time, which I think is much more accurate than time on page. 

Analysis has been around for some time in Google Analytics 360, but it wasn't available to the free version. However, it is now available to everyone for free in App + Web. And there's a lot of different techniques that you can use in analysis, exploration, funnel. One of my favorites because I spent a year of my life building it is pathing.

I was given the task to rethink how pathing works in Google Analytics. And we came up with a way to be able to really expand and look at that user journey. 

So you can look at the event name or the page name or screen name. And you can start to expand these different nodes. Each node is an interaction on the site. And you can see a lot of these things, you can say that the session started and then somebody went to my homepage, and then they continued to go through multiple pages, and you can look at this up to 10 steps of actions that people take. I think that this is really, really cool, to be able to really understand how people are moving through your site in more of a clickstream type way.

I just want to say, thank you guys for having me. This has been really fun and I would love to take your questions now. 

How Should a One-Person Marketing Department Prioritize Tasks?

Tristam: Fantastic. If you could just stop sharing your screen and just stop my mind from being blown from the information I've just taken in, that was fantastic. We've got quite a few questions and we'll just jump straight in. We've got Intrinsic ID asking, what would you recommend for how a one-person marketing department should prioritize analytics, reporting, and dashboard efforts? What would you do first and what tools would you start with et cetera?

Krista: I think as a one-person marketing department, you're probably in a smaller company, you have to wear a lot of hats. I would spend time and effort to try to make sure that your analytics implementation is solid and up to par and you might want to look at bringing in some outside help to be able to do that, if you can afford it, or if you have that skill set, great. 

But, I would focus on spending the majority of your time building out scalable dashboards, understanding what your business KPIs are and being able to track towards those and when you have spare cycles, digging in for more analysis.

Tristam: We've got so many questions coming in. The next one from ... Sorry, Monica, if I pronounced your name incorrectly, Monika Mesnage. “Is there a way to control to show scroll tracking on some pages only? Is there any documentation on how scrolling will be defined?”

Krista: Totally. Yeah, I definitely get this. I think this is maybe one of the drawbacks with enhanced measurement right now is because it is just a toggle. It's a one-setting-for-all-types scenario. My suggestion for you specifically would be rather than using scroll tracking out of enhanced measurement, keep that off, and actually just implement scroll tracking as a custom events in App + Web. It is going to use up one of your 500 events. I've got some feedback into the Google team that they should consider making enhanced measurement a little bit more flexible there so that we can avoid that scenario.

But in your case, if you just create that Custom Event, you can trigger it to fire on whichever pages or screens that you need it to fire on, and you can have more control over the granularity of the percentages and intervals of its firing app.

Do You Need an App to Use App + Web?

Tristam: Then a question from Connor Cluster. Is App + Web something that someone without an app should use, and is it free like the standard version of GA?

Krista: Great question Connor. Absolutely.  I had a mobile app in my example of data streams. I don't actually have a mobile app, I set that up as a dummy stream and going in there. I use App + Web just for my website. 

The name is a little bit misleading in that way, just because it makes you think maybe I have to have both. You don't, you can absolutely use it just for app or just for web. And there's a lot of really great new things as I showed you that you can get from doing that. And yes, it is free, just like the standard version of GA.

And the goal is that it'll be free and unlimited forever. That's one of the benefits of having web is that because of that different data model, they can scale it in a much more scalable way. 

What Can Go Wrong with Digital Analytics

Tristam: Cool. And I've just got a couple of personal questions that I'm curious just to see from your experience, what are the most common things that you see going wrong? Are people not taking advantage of the platforms they have? I still see large businesses that don't have their GA set up as we spoke earlier. 

Krista: I think one of the biggest things that I often see when I come into clients, specifically on the Google Analytics side, and often when I'm coming into a client who already has Google Analytics 360, is that they're just not taking advantage of the platform that they have. And they're very interested in all of the shiny new things, but they really haven't gotten their foundation in place. My feedback is often you got to walk before you can run. Let's learn how to use segmentation properly.

Let's make sure that we have enough custom dimensions set up that we can really do some very interesting analysis here. Let's make sure we're collecting all the data we need to collect, and monitoring the things that we need to monitor. I think some of the biggest drawbacks that I see are just people not fully utilizing the tools that they have, and potentially getting distracted by shiny new things before they're ready. 

Tristam: We got two minutes to the hour. As usual, I don't know where the whole hour goes when we do these webinars. Thank you, Krista, for such an awesome webinar and presentation today. I know it's been enlightening for myself and certainly has in the comments. Krista, where can people find you online if they want to learn more if they want to get in contact?

Krista: Absolutely. So I'm on Twitter @kristaseiden. I blog about analytics at kristaseiden.com. And if you're interested in working with me, you can do so with KS Digital, my digital analytics consulting company and you can find us there at ksdigital.co.

All levels