Google Analytics

Module 3: Next-level site and app analytics
In this week’s module, we’ll revisit the concepts of conversions and goals, and introduce a few new ones—segments and filters.
Conversions, goals & beyond
In your first course, The Digital Audience, you used the terms goals and conversion quite often.
As you learned, there are business goals/objectives that are often a bit vague, and then there are SMART goals—clearly defined goals that are specific, measurable, achievable, realistic and time-bound.
And then there’s conversion, which is an actions users take that moves them from potential customer to actual customer. Conversion is also the last step of another concept you’ve previously learned about — the funnel:
At the top of the funnel is awareness: When users have actually been made aware of your organization or brand. There will be many, many people who are aware of your existence but who never take your relationship further.
Next, there’s interest/consideration: These are users who are interested enough in your organization or brand to interact with it, or have thought about your company or your product, or how it might be of value to them.
If you successfully lure users past the interest/consideration phase, they’ll end up in the evaluation/intent stage, where users are actively considering making a transaction with you. For example, they’re browsing your T-shirts (maybe even putting one in their cart), previewing your music, checking out your upcoming events, and reading up on how subscribing to your e-newsletter would benefit them.
Conversion is when a user takes an action that moves them from potential customer to actual customer. The key word here is “action.” The action that defines a conversion can be anything — it all depends on what your product is. If your organization sells cookies, the conversion point could be a completed online sales transaction (they filled in their credit card information and clicked “submit”). If your organization is a nonprofit, a conversion could be a donation, or an email signup. If your organization produces news stories, a conversion could be as commerce-oriented (e.g., became a digital subscriber) or content-oriented (e.g, viewed two stories and clicked the ‘share’ button once). Think about the physical action users can take that represents a commitment to your company, and that’s your conversion point.
At the bottom of the funnel, there are the audiences who have not only reached conversion, they have liked what they bought/read/heard/joined so much that they become loyal. At the loyalty stage, audiences not only continue to engage digitally and with your brand, but they also may become repeat customers.
A few of these loyal folks will make it down to the narrowest part of the funnel: advocacy, where audiences are so loyal, they’ll talk about your organization/brand in their digital communities. When they talk about their experiences with your brand on social media, review sites, blogs or at the dinner table, they are now doing the hard work of getting new users into your funnel. Think about people who use Twitter to share how much they loved a movie. They not only paid to see it themselves, they are now actively encouraging others to do so as well.

Google Analytics allows audience data analysts to record specific conversions—and when and where they take place. These data allow analysts to understand how it is that their company would like to see its performance evaluated (maybe it’s a purchase, maybe it’s a newsletter subscription, or maybe it’s both), and then gives them the ability to actually gauge its performance.
With that said, there are two main types of conversion:
Purchase conversion
Goal conversion.
Sometimes the point of conversion is a completed sales transaction—that’s a sales conversion or a purchase conversion, and that’s the easiest conversion to wrap your head around. But other conversions include an app download, a form submission, a series of steps taken—say, viewing four articles on a news site. Those are all valid conversion points for sites/apps that don’t involve ecommerce. Though these sound trickier when it comes to measurement, there is actually an entire section in Google Analytics dedicated to this sort of conversion tracking. Once you set up conversion tracking in Google Analytics (by simply identifying what the conversion action is in the account’s settings), you will get lots of rich information in the Conversion reports section (right below “Behavior”).
This sort of conversion tracking is incredibly important to media stakeholders, as they often want to track and measure specific actions or series of actions, even though they’re not the organization’s purchase conversion point. In other words, media companies often want to understand how their audiences are using their offerings in ways that are outside of the point of purchase.
Subaru may want to see, for example, how many of the visitors to its site have download a guide for one of their models. An online clothing store may want to keep track of the number of times its users take advantage of its live-chat functionality. And the developers of an online video game may want to know how many players have reached the highest level. These observations can help media organizations improve their products, by, say, making the live-chat function more visible if it’s clear that audiences aren’t aware of its availability. Alternatively, they can help the organizations make more money. If hundreds of thousands of people have shown that they are dedicated enough to a video game that they are sticking around past all the most challenging levels, then its publisher might consider charging a premium for users who get past its midpoint.
Using Google Analytics to Track Goal Conversions
Because goal conversions vary depending on the organization, they need to be established by the person analyzing the data. Goal conversions can be tracked in Google Analytics by setting up Goals. If you look at the Cronkite News Analytics (go to CONVERSIONS > GOALS), you will see that we have implemented two goals:
GOAL 1 is called “3 pages”—users achieve (or “trigger” or “meet”) this goal when their session includes three or more pages. So, let’s say a user clicks on a Facebook post and get to a story on our site, then they click on the homepage, then they visit another story. That’s three pages in one session, and the user has met the goal. But if another user comes to the homepage, clicks into a story, then exits from that story, that’s only two pages, and that user did not meet our goal.
GOAL 2 is called “Stayed 5 minutes”—users achieve that goal when their session exceeds 5 minutes.
Sometimes, users complete both goals in a single session. In these instances, that’s recorded as one user achieving two conversions.
You can set up goals in any of these categories:
DESTINATION: When a specific URL loads. For example, you want to track how many times the “Submission complete!” page loads.
DURATION: Sessions that last a specific amount of time or longer. For example, you may want to track how many times users spend more than 20 minutes reading what should be easy instructions.
PAGES/SCREENS PER SESSION: A user views a specific number of pages or screens. For example, you may want to know when visitors have looked at 10 or more stories on your tech blog.
EVENT: A social share or video play. For example, you may want to know how many times users click the “Tweet this!” button.
One big advantage of Google Analytics for those interested in understanding audience behavior is that it provides the opportunity to analyze conversion paths (Links to an external site.) and goal funnels (Links to an external site.) so that it’s easier to make educated guesses surrounding the conversion—and where along the way people abandon the path.
Of course, as we’ve discussed previously and will continue to touch on, these data provide some insight, but are not totally comprehensive. If you see people spend a certain amount of time on a page before deciding to subscribe to your newsletter, you might assume that page’s content is what persuaded them to join. However, it could just be that they’d already decided to join, and just happened to be on that page when they remembered to actually fill out the form.
That’s all to say that when it comes to audience analytics, you can see what people did, and you can use it to guide your assumptions about why they did it, but at the end of the day you’re dealing with guesswork as much as you are actual data.
To learn more about conversions start here (Links to an external site.).
Segments
Segments are subsets of your audience data. Applying a segment to your Google Analytics reports let you analyze an isolated group of sessions or users. You can use a segment to look at only users from Chicago who used a desktop, or only sessions acquired from search engines, or only sessions from users aged 18 to 24 that started on your homepage, moved to your sign up page and ended on your email submission page.
It’s important to note that segments are quite different from filters and secondary dimensions.
SEGMENTS ARE NOT FILTERS: You can create filters in your Google Analytics account settings. They are similar to segments in that they allow you see a subset of data, but quite different in that filters are “destructive”—they alter the underlying data. Once you use a filter to narrow down an audience, you can’t get the “filtered out” data back! Segments are “non-destructive,” in that they only isolate a certain set of data until you remove the segment. Think of segments like a pair of glasses you can take on and off, whereas a filter is more like LASIK.
SEGMENTS ARE NOT DIMENSIONS: You already know how to apply secondary dimensions to your analytics reports. Secondary dimensions give you more detail; they break down your data further. Segments, though, change the initial data set. You can look at data through a segment, then apply a secondary dimension to get more detail on that segment.
Audience data analysts can use segments in lots of really, really helpful ways. They can be used to isolate sessions from mobile users acquired by Facebook to analyze how only readers who come from that channel behave with our news content. Or they can be used to isolate only customers from Florida to see how they behave on a “FloriBama Shore” blog, compared to the average user. They can also be used to identify what media device audiences are using. For example, segments can be used to isolate only mobile users who ever reached a company’s Submission page.
I encourage you to go into either Analytics account and toy around with the segment builder to see what the possibilities are. Go to, say, your Audience Overview; at the top, click the big “+Add Segment” button, just next to “All Users.” You can click through some of the pre-loaded segments under the “VIEW SEGMENTS” list, or click the big, red “+New Segment” button and make your own. Click the blue “Apply” button, and now you see comparisons between All Users and only the users in the segment you chose. It’s an awesome way to drill down audience behavior by more specific groups.

Module 4: Finishing up Google Analytics
This week’s lecture is designed to give you the time and resources you need need to move through your final GA training and take the exam. What follows will focus the following: Conversion and Campaign Tracking, reporting, and alternative analytics programs.
Conversion & Campaign Tracking
Although you’ll learn more about these concepts in your GA videos, I wanted to introduce you all to an example of what campaign tracking actually looks like.
This example comes from Cronkite Professor Jessica Pucci, partially because she’s better at explaining how these things work, and partially because the ads Facebook directs at me are significantly more boring (the first one I saw when I scrolled through my newsfeed was for a home alarm system).
As Prof. Pucci explains:
I just logged into Facebook and screen-grabbed the first ad I saw. It’s an ad for Class Pass (what are you trying to tell me, Facebook?!). Look at the ad, and notice that it contains an image and a post containing a link:

If I click the link in the ad, I go here… notice the long URL.

Here’s the URL itself: https://classpass.com/try/2week5newyear1?utm_source=facebook&utm_medium=paidsocialrtg&utm_campaign=acq&utm_content=12680&utm_term=rtg (Links to an external site.)
It contains the URL, plus four UTM tags:
URL: https://classpass.com/try/2week5newyear1
?utm_source=facebook : This tells GA the traffic came from Facebook. Class Pass needs to use a different URL in its Instagram ads with the source as “instagram,” to tell GA that Instagram traffic came from Instagram!
?utm_medium=paidsocial : This tells GA that the traffic not only came from social media, but it was paid. If we just saw a bunch of sessions derived from Facebook, we wouldn’t know how many of those sessions were paid or organic, and we wouldn’t know whether the budget we’re spending in our paid social efforts is actually converting to site sessions!
?utm_campaign=acq : Class Pass can name its campaigns whatever it wants… in this case they’re just calling it “acq,” probably for “customer acquisition campaign.” You might see campaigns called “Winter2018” or “Sweatshirts”… whatever the advertiser wants!
?utm_content=12680 : Class Pass makes lots of different social graphics (just like you probably did in a previous MCO course!), and they assign each a number. Doing so lets Class Pass see which content actually drives sales/conversions.
?utm_term=rtg : This UTM parameter is for paid keywords only (for search ads), so it’s essentially void here.
Making these URLs looks complicated, but it’s actually really easy: you can use Google’s URL builder (Links to an external site.) (or any of the other zillion URL builders) make them… you just plug in your link and your terms and it makes it for you! Here’s more information  (Links to an external site.)on tagging URLs for campaigns, if you’re interested.
Using tagged URLs means better analytics, and less traffic in the “direct” bucket… since, if you don’t tag your URLs, most of your paid traffic (from social and ad campaigns) will end up there, and then we’re blind to it! We can then create segments and dimensions for particular campaigns and campaign content. Further, we can use these to drill deeper down into our Conversion analyses. Sure, we can already know what traffic source people used to find our products/content and follow their steps through the conversion funnel, but this lets us know exactly what social/email content or campaign got them there, and (of course) how users who saw different campaign content behaved on our site. For example, as a Google Merchandise Store marketer, I could create two different videos for our Spring YouTube Mug campaign, and ultimately see which video ended up driving more revenue. That helps me create better videos and social campaigns in the future!
And yes… if you clicked on that link, you have officially screwed up Class Pass’s analytics! Because even though you are here in Blackboard, you have clicked on a link tagged to tell GA that the traffic came from Facebook. If you simply click on https://classpass.com/try/2week5newyear1 (Links to an external site.) (the URL without the UTM tags), GA will record your visit to the site as referral traffic from Blackboard, as it should!
Tagged URLs and other elements you’ll explore in your final GA training are also what helps organizations do what’s called remarketing… or that creepy-ish, hyper-targeted content delivery. When you visit, say, a site that sells baby blankets but don’t take a conversion action during your session, GA takes note of that. Or, if you visit a news site and read a story about crime but don’t subscribe for further content, GA remembers that, too. Marketers, then, use Google Analytics to strategically remarket the site and its content (or its baby blankets) to users who haven’t converted yet (and sometimes, those who HAVE, but could be converted again!). You’ll learn lots more about remarketing in your final GA videos.
Analytics Reporting
We’re now going to move on from analytics collection to analytics interpretation. An important and challenging part of the work that audience analysts do is communicating their findings to others in a way that is both persuasive and understandable. In order for the information gleaned from these data to be useful to organizations, they need to first make sense to the people in charge.
For example: If a managing editor of a newspaper looks at all the visitation data for their site throughout the week, but that person doesn’t know how to read the numbers, they won’t know how to act on them. He or she may come up with faulty logic for why traffic spikes at some times and not others. But someone who knows how to read audience data could look over those same numbers and say, “Hey, it looks like traffic peaks from 7am to 9am, then from noon to 1pm, likely because people are either having breakfast or lunch then, so let’s make sure we’re pushing our best stories out on social media during those two windows.”
That’s just a long way of saying, it’s not enough for the metrics and the insights you derive from them to be stuck in your heads. You need to also be able to share that knowledge with the stakeholders in your organizations. This is why you’re tasked with “working” with stakeholders in your Worksheets.
Often, though, you’re not sharing your insights in an email to one or two people, but in a report routinely distributed to a specific department or the entire staff of an office. For instance, a journalist might receive a morning email that lists all the top performing stories of the day before. This is called analytics reporting. Analytics reporting essentially creates a routine out of analysis. Like any routine, reporting has its benefits, but also its challenges.
Benefits
Analytics reporting introduces consistency and uniformity to understandings of audience behavior. By sharing the same metrics once a week/month/quarter, audience data analysts are able to show a client/organization what metrics are getting tracked regularly, and whether they are or are not improving.
Reporting also gets people who aren’t analysts used to making data-driven decisions, and understanding how their roles, questions and work/output connect to larger organizational goals.
Challenges
When people get used to seeing only the metrics shared in regular reports, they begin to privilege those measures above all others.
When people get reports on a routine basis, it can just become another email they ignore. Many people in the media world don’t think about their audiences, and for those people, audience data reports don’t automatically register as relevant to their daily practices.
So how can audience analytics reporting overcome these potential challenges?
I’ve said it before, and I’ll say it again. “Good” analytics reporting is
Understandable, and
Useful.
If you create an an audience data report that non-analysts have trouble reading or comprehending, then you’re wasting your time. It’s an even bigger waste of time to create reports that aren’t ultimately useful or “actionable” (meaning they can be used to come up with some sort of plan, like improving social media outreach to grow traffic or changing the placement of the comments section to improve engagement). That’s why every good report includes actionable insights. These are syntheses of the data that recommend a particular action.
Sometimes the actionable insights are part of the report itself (as will be the case in this unit’s Worksheet). Sometimes the insights are given in addendum to the report, via an email or presentation. Regardless of where they appear, these recommendations should always be informed by the data, and with strategic goals in mind. There is no point sharing data or actionable insights that do not ultimately connect with SMART goals or organizational objectives.
For example, if, while working as a journalist, I received a note from our audience analytics firm letting me know that responding to more readers on Twitter would likely lead to more time spent on Twitter by my readers, I would be very confused. Why? Because how much time my readers spent on Twitter had nothing to do with me or my site’s success. If, on the other hand, the report encouraged me to add more photos to my stories because doing so would encourage people to share the stories more, then I would have thought, “More readers? Great, I’ll do it.”
In short, choosing which metrics to share in a report is pretty easy if you set SMART goals to begin with: You’ve already done the work of identifying your KPIs—how you’ll measure whether your goal is successful—so reporting becomes about tracking how close you are to reaching/meeting/exceeding those goals.
Other Analytics Programs
By now, you’re familiar with Google Analytics, and how we can use it to track our digital audiences, content and businesses. This means you already have an advantage when it comes to understanding how audiences behave in today’s media world. However, there are other useful analytics products out there. Some are comparable to Google Analytics in their scope, while others are smaller and more specialized. Others have totally different goals and methods for collecting their data. All of these other offerings can be used by audience analysts to piece together the story of content performance.
Historical analysis
Google Analytics is what we call a historical analytics tool: We can use it to explore audience, content and ecommerce data across chunks of time large and small. And, these tools connect business data (like revenue) to digital audience data. We can manipulate the data to show us metrics and insights regarding activity that has already happened—a day ago, a week ago, a year ago—and that data helps us describe site/app/business performance over time. Google Analytics is by far the most popular and widely used analytics tool of its kind, but there are others (e.g, Adobe Analytics).
Real-time analysis
You may have noticed that Google Analytics has a “Real Time” reports button above the Audience, Acquisition, Behavior, Conversion report navigation. Try clicking it to see real-time analytics in action.This shows what audiences are doing on your site right at that moment.
There are a slew of other tools out there that allow you to see site performance live and in-action. These tools show you what’s happening now: Where’s traffic coming from right at this moment? What are people reading this instant? What products are people accessing in the last 10 minutes, and what sites or apps are referring the traffic? The value of this information is obvious for anyone working in, say, breaking news. If an editor has a news story that’s getting a lot of clicks on Facebook but none on Twitter, they can take immediate steps to balance that out.

The two major players in this field are Chartbeat and Parse.ly. The screenshot above shows the general gist of real-time analysis: Here, Parse.ly displays what’s happening right now on the Cronkite News site, what users are spending time on (within the last 10 minutes, and today overall), and how our performance at this moment compares to the site’s average day. In the Cronkite News newsroom, a version of this screen is on a huge monitor that everyone can see, so that everyone understands what’s happening… and so that everyone can jump in and share content if, say, performance dips below average.
Competitive and cross-platform analysis
One thing Google Analytics cannot do is show how an organization’s online audience behavior compares with that of its competition. Competitive analysis tools provided by firms like Nielsen, comScore (Links to an external site.), and SimilarWeb let you do precisely that. Here’s a screenshot of a SimilarWeb comparison of Cronkite News and two competitors, AZCentral.com and ABC15.com.

The benefit of these sorts of data is that it allows organizations to understand where they stand relative to their competition. If I am the editor of a film review website, I might be curious to know how many people have visited my site relative to Vulture or The A.V. Club so that I know how popular my site is compared to how wide an audience I could potentially be reaching. These audience data providers also sometimes offer organizations measures of cross-visitation, so I can actually see how much overlap there is between The A.V. Club’s audience and my own. This can be useful for determining where to place digital ads. If I know a pool of people who would like my site are visiting other film sites, I might want to place more ads on those sites to increase brand awareness. These sorts of data are also essential for advertisers, who use them to determine where their ads will find the largest audiences.
Companies like Nielsen and comScore collect data using panels, meaning they gather a large group of people (for comScore it’s a million in the U.S.) and track all of their online behavior. Then they use statistics to make projections — they use math to make best estimates of how their sample is representative of the national population. The data collection method means the data isn’t as precise as the data we get from Google Analytics, which is a definite disadvantage. However, it’s the only way to compare site data across many sites, aside from asking each individual site to give you Google Analytics access… which most organizations would never do.
Behavioral analysis
All of the tools we’ve discussed so far help us understand audience behavior to some degree. But there are other tools that let us learn more about specific audience behavior that isn’t measurable with traditional analytics tools. You may have heard of eye-tracking  (Links to an external site.)research (Links to an external site.)—that’s (labor-intensive, expensive) research that measures slight movements of human eyeballs while users look at content. We can get close to that with digital heatmaps, which monitor users scrolls, mouse movements and clicks. It’s certainly not as good as eye movements, but the trade-off is that we can collect data far more quickly from far more users. Digital heatmaps reveal behavior on that one page by device—desktop, tablet or mobile phone—and behavior.
Summary
As these different audience data providers reveal, the field of audience measurement is one that continues to grow more expansive and sophisticated by day. Now, media companies can track not just how many clicks their content gets, but who clicked on it and how much of it they actually consumed. As Nir Grinberg, a research fellow at the Harvard Institute for Quantitative Social Science, noted in his analysis of how people actually read online news: (Links to an external site.)
“Instead of just how far down the page a person got, I’m looking at what percentage of the article they actually covered,” he said. “How far did they go down the page, relative to the length of the article? If someone spent a lot of time on an article and the article is short, that’s a good signal. If they spent the same amount of time on a long article, that’s less good.”
This knowledge doesn’t always mean more clarity, however; in fact, many believe that the increase on audience metrics leads instead to an increase in confusion surrounding audience understandings. As Derek Thompson writes in  (Links to an external site.)The Atlantic (Links to an external site.), “No matter what metric we settle on, there will be reasons to doubt and editors to manipulate it.”
With that all said, it’s worth remember what we discussed at the beginning of this class, which is that even in a media landscape with countless, granular audience metrics, the measure that continues to matter more than any other is the size of the audience. “Where are most people going, and how can we direct them to us?” That’s the question that faces nearly all media providers in today’s digital world. And it’s one you’ll explore in more depth in this week’s Worksheet and in the weeks ahead.