The Ultimate Guide to Google Analytics
You know analytics are important. It’s valuable to gather data about your customers’ behavior, so you can see what’s working and what isn’t. So why aren’t so many business using their analytics in the right ways?
Why aren’t people acting on their analytics? Why aren’t people even frequently checking their analytics? Why isn’t ownership of analytics specifically stated? There are probably a few reasons:
- Analytics are undervalued. Google Analytics gives you an absolute crapload of data for free. While this is great for your business’s bottom line, the free-ness of Google Analytics probably harms its reputation. We don’t often value things that are given to us for free, no matter how valuable they may be. And with something that requires a lot of work to understand, Google Analytics has a lot working against it.
- Analytics take consistent effort. You should never stop looking at analytics – and building the habit of a data-driven strategy is difficult, especially when you have other fires to put out on a daily basis.
- The product changes constantly. Should you be using Universal Analytics or Google Analytics? Is your snippet correctly configured? What’s the purpose of ecommerce tracking?
- There’s a lot of snake oil. Most advice about analytics is untrustworthy, skeezy, or (at best) out-of-date. How do you know who to trust?
So, there are a lot of issues here. The consequence is that it’s easier to not act on analytics. And the value you’re losing by acting on analytics is not directly felt. Customers aren’t clamoring for you to gather data on them. Your competitors aren’t trumpeting their awesome analytics as a clear advantage over you.
That does not, however, make analytics any less of an expensive problem – and that sucks. If you have a bad (or nonexistent!) analytics strategy, it is actively harming your business. How? Well, think of the benefits of a data-driven strategy:
What if you could really understand what your customers were doing, rather than speculate on best practices that mght work? What if your copy strategy was rooted in research and data – just like the pros do? What if you could find all sorts of places in your business’s funnel that are leaking revenue – and take clear steps to fix them?
In this guide, I’ll explain how the best conversion experts use analytics to capture all kinds of revenue, walk you through exactly how to prepare your analytics correctly, and demystify analytics as a practice once and for all.
Google Analytics tells you that setup is as easy as installing a tracking snippet. That is… not entirely correct. You need to modify the snippet twice beyond that, and there are a few settings that you need to tweak as well. First, you need to configure a few things that Google Analytics has turned off by default.
- Turn on demographic and interest reports. Go to “Admin” in the header navigation, then “Property Settings” in the middle column, then find the switch reading “Enable Demographics and Interest Reports.” Make sure this switch is turned on. This allows you to gather demographic data about those who are visiting your site.
- On this same page, turn on enhanced link attribution. This allows you to make sense of the impact of specific links on your page. For example, if you have 3 calls to action that all take you to the pricing page, you want to know which ones proportionally convert better – and which ones are not performing as well as they could. Find the switch named “Use enhanced link attribution” and make sure it’s turned on.
In order to enable these for everyone, you will need to add the following to your GA tracking snippet, on every page GA appears:
ga('require', 'linkid', 'linkid.js’);
You’ll want to insert these two lines immediately before the final line, which should be
ga('send', 'pageview’);. Google Analytics doesn’t have these turned on by default for your property, and you should turn them on no matter what kind of business you run.
If you’re running an ecommerce site, you’ll need to configure ecommerce settings for your site. Go to “Ecommerce Settings” in the same admin column as your goals & view settings, and turn the switch on for ecommerce:
Then, turn the switch on for advanced ecommerce reporting, and hit submit:
Next, you need to update your tracking snippet again. Before the last line of your tracking snippet, add the following:
You’ll then need to add the right ecommerce tracking on your thank you page, which requires hooking into Google’s ecommerce API. Here’s the reference for doing so.
Google’s product “Search Console” (née “Webmaster Tools”) is the most effective way to understand what search keywords are bringing people to your site, as well as what pages they’re visiting – and how that maps to your revenue stream. You should absolutely configure Search Console and connect it to Google Analytics.
First, go here to configure Search Console for your property. Click “Add a Property” and fill out the prompts.
Go to “Admin” in your header navigation on GA, and click All Products in the middle column of your Admin page:
Scroll down to “Search Console” and click “Link Search Console”. If no site has been linked from Search Console to Google Analytics, it’ll look like this, with “none” listed:
Hit that extremely discoverable and legible “Edit” link and follow the prompts.
If you have a search form on your site – which is especially valuable on big ecommerce sites, or sites with large blogs – you’ll want to configure site search on Google Analytics, to discover what people are searching for – and whether they’re getting to the content they want.
Site search is configured on your view. Go to “Admin” in the header navigation, and then in the right-hand column, find the view you want (usually “All Web Site Data”), and hit “View Settings” right beneath it.
Scroll all the way down, and turn the switch on that reads "Site search Tracking”:
Once you do, a field will appear that reads “Query parameter.” You’ll need to enter the GET variable used for search queries here. For example, if your variable is
q=search+term in the URL of the resulting search page, enter
q in the box here. You can enter as many variables as you’d like, comma-delimited, with no spaces (e.g.
Configuring goals & funnels
Next, you need to create goals. The two biggest ones are revenue and raw sales. Goals are how you measure your conversion rate when optimizing your site, and they’re absolutely critical to ensuring that you can measure a solid ROI from any optimization efforts.
Go to “Admin” in the header navigation, and then go to “Goals” in the right-hand column. If you haven’t configured any goals, you’ll see something like this:
Hit “+ NEW GOAL” there, and create a new goal for a sale.
Let’s say you’re a SaaS and you want to create a signup goal. Hit the “Create an account” radio button and then continue.
- Make the goal a destination – for example, the first page that people see when onboarding onto your product.
- Ascribe a monetary value (optional), which should be the LTV of your customer. Keep in mind that this only works for customers on a specific plan with a relatively inflexible LTV!
- Turn on the funnel you expect people to follow, and fill in the fields for each page. This is terrific for SaaS businesses.
- Finally, click “Verify This Goal” to ensure that it’s firing correctly. If it’s not, you need to bugfix GA.
Your final form should look something like this:
Next, create a goal for revenue. Go through the same process to create a goal, using the checkout thank-you page as your destination.
Configuring the right ways to read your data
Finally, you’ll need to configure the best ways to read your data. It’s not enough to poke through the fire hose of data and see what happens; instead, you’ll need to create subsets of your data to gain the best possible insights.
Google Analytics has three different ways to create subsets of data: views, filters, and segments.
- Views are the way you understand a set of data on Google Analytics; each view lets you contain a separate set of reports and filters. You should create a view for all raw data that you never modify, a test views for you to mess with in the future (like the difference between staging and production on your web server), and the view that contains your actual filters & reports. Here’s a deep dive into views that teaches you more about best practices.
- Filters are ways to qualify out traffic from the raw data set. Think hits to your application vs. your marketing site, or views from mobile vs. tablet vs. desktop.
- Segments are how you define specific behaviors for people. For example, you might want to create a segment only of people who came in through organic traffic. Or people who, once they arrived, watched your masthead video. This allows you to vet the success of a design decision by narrowing down traffic that is both wallet-out and only evaluating the thing you’re changing.
Go to “Admin” in the header navigation, then hit the pull-down underneath “View” in the right column. You create new views by hitting this option that looks like all the other options in something that’s buried three clicks away from your property’s home page:
Note that Google Analytics has a hard-limit maximum of 25 views for the lifetime of the account, so you should be careful and intentional about what views you create!
Once you create a view, navigate to it in the pull-down on the admin screen, and modify what segments, filters, and content groups that you are using for it.
Once you’ve created a view, you can add filters to that view by going to admin, navigating to your view in the pull-down at the top right of the screen, and then heading to “filters” beneath it:
You select the parameters of your filters with a series of pull-downs:
Once you have your filter configured, you should verify that the filter works with the link “Verify this filter” at the bottom of the page.
You can add segments to a view by going to the “reporting” tab. Segments appear at the top of this pane; “All Users” should be selected by default. Hit “+ Add Segment” and either create a predetermined segment by Google, or create one of your own:
How & when to check in on your data
Once you have everything configured, you should check in on your data at least weekly. You have two goals during your check-ins:
Ensure everything is working
All sorts of stuff can go wrong in just a week. For example, someone can move or rework your snippet on the site during an otherwise-innocent code deploy, mess with the URLs that are tracked in common goals or funnels, or deploy pages that simply don’t contain analytics tracking at all. Additionally, other team members (bless their hearts!) can always muck with your account such that views, funnels, segments, and other key data-gathering sources are now broken – or reporting inaccurately or unexpectedly.
Every few months, I go through the Google Analytics health check that ConversionXL put together – which provides a solid deep dive into your account, ensuring everything is working correctly. On a weekly basis, I:
- Check all goals & funnels to ensure nothing broke in the past week.
- Check all new pages – including products, new features, landing pages, etc that were deployed on the repository in the past week – against analytics, to ensure that hits to them are being reported correctly.
- Check all segments, views, goals, and funnels to see if anyone added or modified anything.
In my experience, I’ve usually found one thing worth correcting every week. Sites change, and analytics always needs to keep up.
Find revenue-generating design insights
Bugfixing, of course, is the boring-but-necessary part of the job. Far more important – and the reason you’re here – is to pull revenue-generating insights out of the data you’re gathering. You can do this two different ways: by figuring out where you’re leaking revenue, and by figuring out opportunities to improve conversion rates.
Fixing revenue leakers
The former is easier, at least at first. First, we’re going to go through and see what browsers and platforms are leaking revenue. This is precisely what it sounds like: the boring, journeyman work of browser testing and bugfixing.
You might be displeased at the prospect of more thankless, crummy work of browser testing. But browser testing is absolutely essential before you start any A/B tests. Why? A truly shocking number of revenue-killing bugs surface on comically popular browsers. Leaving them unchanged is no excuse. You must audit your revenue-generating funnel at least once a quarter in order to kill browser bugs before they become a problem for your business.
Here’s the most painless way to do that.
On your Google Analytics dashboard, go to Audience -> Technology -> Browser & OS. Hit the bar graph that’s second from right (the one with zig-zag values), which shows changes relative to the site average:
Then, take the pull-down at right and change it to “Bounce Rate”, “Revenue”, or “Conversions”. You should end up with something like this:
In this case, most people are coming in through Chrome and not bouncing (good), but most other browsers are underperforming (bad) – and they comprise over half the traffic in aggregate (worse).
Mobile, Tablet, Desktop
Go to Audience -> Mobile -> Overview, and switch your view to show relative performance again. This view shows broad-stroke performance of mobile devices and tablets:
The top row is for desktop, 2nd is tablets, 3rd is smartphones. Here, I’m using bounce rate – but you can always switch that pull-down to reflect the actual conversion rate that you configured in your goals during the previous lesson.
Note that the bounce rate is much higher for mobile browsers, especially in-app browsers. There could be an opportunity to test a callout that captures the customer’s email in these circumstances, so you can begin a customer relationship more easily.
For specific devices, go to “Devices” below “Overview” in the left-hand navigation. Discard “(not set)”, because there’s no way for us to know what that is. Combine specific platforms if and only if there’s very little difference in variation – specific devices may end up leaking revenue, and they could even operate at volume!
Specific screen resolutions could also be a conversion killer, especially on Android. Explore “Screen Resolution” under “Secondary Dimension” in order to assess this:
Note: even after looking at this report, you should absolutely get a device out and test your site natively, in order to understand what’s possibly causing the elevated bounce rate. Trust me, you can find one. And yes, it’s worth the effort.
Go to Behavior -> Site Speed -> Overview. Then, scroll to the bottom of the page, click “Browser” at left, and then hit “view full report” in the bottom right corner:
You should end up with something like this:
Why is Android’s browser underperforming here? And why is Mobile Safari struggling? It’s probably because the page load times are longer on cellular networks. Do what you can to improve the speed in these browsers: by using CDNs, more mobile-friendly layouts, and shorter page load times.
Finding opportunities to improve conversion rates
Next, there are a few common places where you can look to improve conversion rates. First off, let’s look at your funnels.
Assuming your funnel is working properly, you should take a look at any significant drop-off points. People frequently drop off when you send them a checkout form, and there’s not a whole lot you can do about that – other than making the checkout form itself as brainlessly easy to use as humanly possible.
Beyond that, I always look for points in the funnel where more people are dropping out than you’d expect – not only from page to page, but also from form field to form field. You can use Hotjar to examine form field drop-off rates, and per-page drop-off rates can be analyzed in your Google Analytics funnels.
This is great for isolating bugs, paring back extraneous form fields, and identifying the weakest points in your funnel.
I love analyzing traffic sources on Google Analytics in order to understand how to custom-tailor an experience to new arrivals. The “Acquisition” tab in Google Analytics helps you understand where people are coming from before they come to your site – which provides valuable context for how to craft your pitch once they arrive. For example:
- Referrals can be accompanied with a welcome message, saying that they came from a given blog or podcast.
- Social media views can have custom-tailored content based on the post they visited from, or a search for a specific hashtag that might be of interest.
- Email clicks can automatically populate discount codes for one-off sales.
Take a look at your key traffic sources, figure out their corresponding conversion rates, and see if there’s any way you can serve them custom-tailored experiences.
Speaking of custom-tailored experiences, what about people who come back to your site? What if they already purchased? What if they’re existing customers? What if they threw your product in the cart, but didn’t act?
Each one of these should have a separate game plan. Talking about what to do in each of these situations is outside the scope of this guide, but you can easily figure out the share of past customers based on how many completed a goal or funnel before coming back – which helps you prioritize what strategies to pursue first.
Finding Correlations with Conversion
What’s the main behavior your customers take before they convert? Slack quickly realized that sending 2,000 messages on a team correlated strongly to lifetime retention. In the words of their founder Stewart Butterfield:
Based on experience of which companies stuck with us and which didn't, we decided that any team that has exchanged 2,000 messages in its history has tried Slack — really tried it. For a team around 50 people that means about 10 hours’ worth of messages. For a typical team of 10 people, that’s maybe a week’s worth of messages. But it hit us that, regardless of any other factor, after 2,000 messages, 93% of those customers are still using Slack today.
You probably don’t run a workplace chat app, but the spirit remains: you probably have a behavior that, when completed, maps strongly to conversion. If you’re running a SaaS, that may involve completed onboarding or data migration. If you run an ecommerce business, that may involve getting to a specific point in the cart, or successfully customizing a product.
Analytics help you understand what correlates best with conversion, in order for you to take action.
In any organization, someone needs to have clear responsibility for analytics. Ideally, it’ll be baked into their job title; but at the very least, it should be made clear that this is a high-priority responsibility, performing it overrules other fires in their job at the time, and they should be expected to share analytics-driven design insights with the rest of the team as frequently as necessary.
Translating analytics into design insights
I use Trello to keep track of all research I’m conducting for a client. Here’s the exact template I use on Trello with all of my clients. Design decisions come in two flavors: they should either take the form of A/B tests, or they should be rolled out to everyone between tests.
One-off changes involve no debate. They include usability improvements, bug fixes, and page weight optimizations. In my own practice, everything else is tested – especially changes to copy, pitch, demographic targeting, and so on.
This guide focuses on Google Analytics, but there are many ways to gather valuable data you need, including Mixpanel, Segment, Kissmetrics, and even A/B testing frameworks like VWO & Google Optimize. As your business grows, it’s important to find the right combination of data-gathering tools that works for you.
Why? Because a continuous focus on analytics generally provides a positive ROI for any business. Following these steps, you’ll get started on the right foot with Google Analytics – which is essential for any business, and free to boot.
Most stores get CRO wrong.
They plan CRO activities at the wrong time, or they don’t know how to research new tests. Our free CRO roadmap draws on our years of experience to make CRO easy for you to implement for your store.
Sign up below to get your Ecommerce CRO Roadmap today:
Draft’s Letters is a port of actual value in an online storm of listicle posts.