Google Analytics is dead; long live Google Analytics

— Written by Justin Beasley, with an intro by Marlin

TLDR; Google Analytics as we know it is going away in the near future. It is time to take steps today to replace and enhance your existing tracking so you can better understand your visitor and donor motivations. As we often say, you can’t optimise what you can’t measure. For those wanting the full scoop, keep reading for a detailed account of what this means and what you should do to prepare.

On 16 March, 2022, Google officially announced that they’ll retire (or “sunset”) the previous version of Google Analytics, called Universal Analytics (UA) or more recently referred to by many as “GA3.” They will stop processing new tracking data for these properties on July 1, 2023 (about 15 months away; paid GA360 customers get until October 1, 2023). That’s a pretty big announcement considering that the vast majority of websites in the world still exclusively use GA3 for tracking!

While this news has been somewhat expected since the official launch of its successor (Google Analytics 4/GA4) over two years ago, the relatively short timeline surprised many of us in the analytics world.

So what’s a nonprofit marketer or fundraiser to do?

First, don’t panic!

As it stands today, your GA3 properties will continue to collect data. You have a moment to think through this and come up with a plan. While you don’t actually have 15 months (we’ll get to this in a moment), it’s best to always take measured action with your . . . measurement.

There are a number of reasons why Google is making this decision:

• The underlying technology for GA3 is 15 years old. That’s multiple lifetimes in technology. The platform has significant technical debt for them that hampers their ability to innovate, and more specifically adjust to the changing privacy and legislative landscape that is rapidly evolving. Ranging from tracking consent to data deletion requests under legislation like GDPR, this is a very sticky area where Google is facing legislative pressure and likely helped drive this more aggressive timeline.
• In the two and a half years they’ve given folks to upgrade to GA4, very few actually have. Despite Google saying that it was ready for primetime, many of us have had notable concerns with missing features which slowed early adoptions. Even when many of these concerns were addressed, it also added to the stigma that GA4 wasn’t ready for the masses. For nonprofits specifically, there isn’t a single donation page provider that has a native GA4 integration that we are aware of.
• Those who have set up GA4 have primarily done so alongside their existing GA3 implementations. (And for the record, this was absolutely the right thing to do and is the strategy we’ve followed and helped our clients implement.) This allows you to get historical data flowing into the new platform while not breaking your day-to-day reporting, integrations, or having to re-train everyone on GA4 (which is wildly different from GA3). Google is giving these folks a nudge to get comfortable working in GA4 as their primary source of truth, and fast.

Google has given us the solution—GA4—and has done so for free. Not only is GA4 a much more modern tracking platform with many changes under the hood, it also gives you access to a ton of features that were previously only available to paid Marketing Platform (GA360) customers. These include methods of getting raw data out of the platform (BigQuery linking), unsampled data for most reporting, custom audiences (more flexible than just segments, and can be published for ad targeting/exclusion), and “Data-Driven” attribution.

There are a bunch of features that are entirely new in GA4 as well, such as 

• The new pathing and reverse-pathing reporting, which let you easily look at what people who land on certain pages or match certain profiles do next on the site (and not just pageviews—any other events you want to view too).
• Answers to questions like “for people who gave through the main donation page, what were the steps they took previously?” This is just one of many new ways to dig into your data through Explore (similar to Adobe’s Analysis Workspaces)

Even better, their pace of innovation has been much faster with new features being added regularly. Not to mention that the model is no longer anchored to rigid pageviews which increases flexibility of what is defined as a “page view.” 

It’s also expected that while new data won’t be processed after July 2023, they’ll give folks at least several months where they can view their existing GA3 reporting (as well as access it via the API).

Some of us have been here before—I still remember the pulsing “Run logs” button of Urchin (which later was bought by Google and became what we now call Google Analytics), and then GA “Classic,” Universal Analytics, and then GA App+Web (which later morphed into GA4). 

Even before those times, I remember processing logs with Webtrends and Analog. Change is the nature of technology. The fact that GA’s model lived for 15+ years is both a testament to the platform as well as an indictment of the lack of major innovation. It’ll be okay. I promise. After all, this is pretty low on the list of issues currently plaguing our world.

So there’s your silver lining. But now we need to get back to the storm clouds and how to respond to them . . .

Second, make a plan*.

For some of you, I’m sure that I lost you at the word “plan.” Others find planning fun and exciting—my wife is one of these. I land somewhere closer to it being a “necessary evil.” Regardless of your feeling toward planning in general, necessity is the word I want to stick with you here. 

A deeply held value at NextAfter is you cannot optimise what you cannot measure. So if you care about growing your traffic—ultimately your ability to reach people and build relationships that further your cause—read on.

One of the first questions in planning should usually be timeline. Google gave us a date, that’s a start. But if you plan to procrastinate until June 2023, you’re in for a world of hurt. Here’s why:

1. You’ll start off with a blank slate, with no historical data to compare with
2. You won’t have time to check your implementation and correct any tracking issues, or to ensure that all important events are being tracked (or get your platform to add support or change it if they don’t support GA4)
3. You won’t know how GA3 and GA4 reporting compare (they won’t be identical; we’ll get to that later)
4. You won’t have time to learn how to use GA4 or train others in your organization how to work with it until there’s already a time crunch/crisis

So if that isn’t the timeline, what should be?

We strongly recommend trying to get as much of this in place by July 1 of this year (only a few months away) so that you have a full year of historical data before the switch. If it were me, I would urge you to get basic tracking added before the end of this month, then plan to spend time the next couple months making sure you have a tracking strategy that works, checking things like ecommerce/donation tracking, ensuring you have the goals and conversion events you want, etc.

Keep in mind that our recommendation is still to “dual tag” your site—that is, track to both GA3 and GA4 for the time being. This gives you the ability to compare the two, make sure you have historical comparisons in the meantime, as well as time to learn (and train others) to accept GA4 as your new source of web analytics truth.

*So I lied, earlier—you actually need to make two plans. A migration plan (how you’re going to get GA4 up and running) and a measurement strategy (how to make Analytics answer the right questions, and answer them correctly).

Migration Plan

Since this is the urgent topic at hand, we’re going to start here . . . but I’m not letting you off the hook on that measurement strategy (I see you looking for the back button!). Here’s a (hopefully) helpful checklist to help you prepare to get GA4 set up and tracking data and avoid any pitfalls along the way:

• Identify who is responsible for analytics implementation in your organisation. Also ensure that this person/team has access to create a new GA4 property in Analytics under the same Analytics account/organisation as the GA3 property.
• Also identify who currently uses Analytics. These are people who will be affected by the change, and may need to know it’s coming. They are also likely people who can help to ensure that their portion of the reporting continues to meet their needs (which I’ll cover more when we talk about measurement strategy)
• Work with all of the above to identify where GA4 needs to be implemented—which websites need to be tracked, and don’t forget to include any third-party tools like donation form and landing page providers, multimedia players, etc. (I’ll refer to the combination of these as “web properties” going forward)
• For any tools or platforms which don’t currently support tracking with both GA3 or GA4, open support tickets or contact your sales representative to ensure you know their timeline to implement this functionality. (Note that some platforms may only allow this through adding custom code or using Google Tag Manager, so don’t forget to look into those options as well). Do this early in the process so it doesn’t impact your timeline!
• Take the list of web properties and determine whether they should be tracked to the same GA4 property or separately. For example, let’s say you have a main website and also use a donation provider which is linked to (or embedded on) that site. Because visitors are likely to interact with both and you’ll want to know how site visitors convert to donors, these should logically be tracked into the same GA4 property, regardless of what has been done in the past.

Now let’s say you have a separate domain for an event. If there’s little to no crossover between the site audience (say, they don’t even promote/link to each other), I’d argue that it would be better to put it on its own separate GA4 property. But if the two sites, say, share a common site navigation and a large portion of traffic flow between them, it might make sense to track them to the same property.

Other examples are more clear-cut: If you have an internal intranet site, unrelated sister organisation, etc. these should almost certainly be set up as separate GA4 properties.

• Set up the GA4 property (or properties) identified above. Google has instructions here for adding a GA4 property if you’re already using GA3 (if you did it differently; don’t worry—this just saves a couple steps). This is a good time to add other users who may need access to the property going forward.
  
  If you’re currently using the newer gtag.js method for GA3, pay special attention to the “Enable data collection using your existing tags—it may be possible to let Google automatically collect the GA4 data for you from those tags. The downside here is that it will only do so for the common tracking fields (so if you’ve done any customisation for GA3 that will not carry over. More on that here). In general, I would only recommend this option if you don’t have access to get tags on your site or that process takes a long time (at which point you can use this temporarily to get started). Also beware if you have some parts of your site tagged using gtag and others using older tags—this will only work on the gtag pages.
• Determine how each web property will be tagged. If you’re currently using Google Tag Manager (GTM), we recommend adding the tags there (this is a huge advantage of using a tag manager, but we get that can be a little technical for some folks—so while we highly recommend it, it’s okay to not have the bandwidth to learn two new things at once!).
• If you’re on WordPress like many sites, you might want to consider using the official Site Kit by Google plugin to implement the tags since it’s very user-friendly and also gives you some handy stats within the WordPress interface (as well has helps you link your site with Search Console and PageSpeed Insights, other sources of useful data). It also supports loading your tag container if you’re using GTM.

If you’re using something else, check their documentation or contact Support to figure out how to get tracking added (and ensure you can load both GA3 and GA4 tags so you have that continuity until the sunset date). 

Again, start this process sooner than later in case there are any surprises.

• Get the tags installed (or the tracking IDs added if your platform handles the tagging themselves) and verify that they’re working—and that GA3 tracking didn’t break. There are a number of ways to verify this, but I’ll give you some favorites:

1. DebugView in GA4: This is a fantastic new feature, as it allows proper debugging right in the GA interface of your browser. Julius Fedorovicius of Analytics Mania has an excellent detailed breakdown of using the feature here. This is the preferred way to check your GA4 implementation.
2. Google Tag Assistant (GA3): Before DebugView, this great Chrome extension from Google was the only way to see a lot of that data. It also shows what Google tags are loaded on a page, as well as pageviews and events being fired—and how GA3 hits are received and processed by Google. You can also record multi-page sessions to look for tracking issues. Here’s a link to the extension as some more advanced use-cases with the recordings feature.
3. Real-Time Reports (GA3/GA4): Another long standing way of answering the age-old question of “is this thing on?” These reports show data coming into your GA property in (near) real-time. Here are the notes on how to use those for GA4 as well as GA3. (I’ll note that these reports are increasingly unreliable in GA3, so I wouldn’t recommend actually reporting from them until you’re using GA4.)
4. Omnibug: This Chrome extension works similarly to Tag Assistant, but also shows other popular marketing pixels such as Facebook, Google Ads, Adobe Analytics, etc. 

Be sure to check specific features by making test interactions. For example, if you were previously using ecommerce tracking to track donation conversions in GA3, make sure that’s all still working correctly (and if you aren’t, you should be!). We’ll get to the GA4 side of that in a moment.

You’ll find that in different setups and scenarios, one of these may work better than another (or sometimes one won’t work at all). So hopefully this gives you a toolbox for making sure things are working as expected.

• Build your measurement plan (described below)
• You’re done—pat yourself on the back and keep being awesome for your nonprofit!

See? That wasn’t so bad, and maybe you learned a couple things along the way.

Measurement Strategy

This might not be the sexiest part of data analysis, the old quote from Zig Ziglar rings true: “If you aim at nothing, you will hit it every time.” That’s equally true in web analytics, where there are no second chances to collect data correctly when it doesn’t happen the first time. That’s the purpose of a measurement strategy.

I get it—I know you don’t need one more thing right now. You’re still reeling about the loss of your beloved Universal Analytics. But I promise, you’ll thank me one day for this one some day.

At a high level, a measurement strategy defines:

• What’s important to your organisation
• How you measure/report that
• And what data you need to do that

Simple, right? Well it would be if we had a single goal with a single measure coming from a single source of truth. But in reality, there are many. Econsultancy put it this way: “A measurement strategy in itself is a definition of how data will be used to understand how [organisational] goals are being met. [Organisational] goals and the wider context of how they drive the [organisational] need to be collated, appraised and prioritised.”

So first off, this isn’t a one-time exercise—it’s a living, breathing process that needs to evolve with your organisation. But documenting it is often incredibly enlightening, and helps to build consensus around what’s important. As an analyst, there’s nothing worse than presenting a bunch of data and the stakeholder saying “so what?”

Data is a critical tool to drive action toward what’s important in your organisation. But if you wait to track donation transactions in GA until someone asks for it, it’s too late. Similarly, if a volunteer signing up is essential to your cause, or pledge signatures are important to initiative X, you must first start tracking those things. Because the inevitable follow-up questions come: How did that donation ask on Facebook perform? What’s my best source of volunteer signups? Is Y a better return on investment than Z?

Data collection is by far the most important part of web analytics. The quality of collection directly affects the quality of the analysis, and also the questions it can answer. But it’s the part we spend the least amount of time thinking about and maintaining.

How detailed this needs to be depends on the size and complexity of your organisation, but start with this:

1. Ask who your audience is. I’m not talking about personas, just about thinking through some of the types of people who are interacting with you online. Discuss this broadly, and also talk about why they’re coming to your site and what success looks like for them. This helps to anchor your plan in empathy for the people you’re actually wanting to help (or who are actually wanting to help you!)
2. Collaborate with key stakeholders in your organisation on what important actions people might perform on the website (include items your website lacks if leadership has already identified it as important). Usually you’ll end up with 5–10 items (some freebies for nonprofits: “Donate” and “Sign up for email” are must-haves).
3. Rank these items in order of importance or preference. Put a star next to any that your organisation couldn’t function without (for an advocacy organisation, this might include pledge signatures. Or for another, child sponsorship—the point is to understand essential vs. valuable, and also prioritise your efforts).
4. Discuss what you need to know about these items. What questions would your future selves want to ask? If you looked at this data in a year, what would you need to know if it was a success or failure? What would you want to drill into to see what led to that trend?
5. Decide what metrics (pieces of data) you would need to answer those questions as well as what segments/dimensions you would need to drill down the way you would want to. This also helps to teach stakeholders what data is available, and how to ask good questions that relate to actual outcomes (and not questions about bounce rate, which is a very, very poor metric—but I digress)
6. Determine which of these metrics are already available and which aren’t. For each, ask yourself if measuring it actually helps you achieve your mission and not just have data for data’s sake.
7. Plan what needs to be added, tracked, or researched. Set a timeline, assignment, and action items for each.
8. Iterate as you make changes, get new questions, and learn from experiences make sure to maintain and update a “living document” with this information. When questions come up, point stakeholders and leaders to the priorities they’ve agreed on, as well as what it will take to measure them.

There’s a ton of information out there on specific templates and methodologies for measurement strategy, and this barely scratches the surface—but I hope that this gets you thinking about the concept in a way that should be accessible to nonprofits of all sizes.

The main point is to not just Track Things™ but instead think about the why of tracking and to put some intentionality behind actually achieving those organization goals (and knowing when you have).

Third, take some time to learn GA4.

The changes under the hood in GA4 means that the reporting interface will feel very unfamiliar at first. Trust me, it’s not just you—everyone I know has struggled with getting reoriented to the new interface, myself included.

While Google has made several changes to bring back reports that were originally absent, things like the navigation and how you drill into reports is unlikely to change. On paper, the new model is arguably better—but it may not feel like that at first when you struggle to answer basic questions.

Like any big change, I’d recommend breaking down your efforts and focusing on a couple key areas first:

1. First, get used to how to navigate the reports and where to look for things that are no longer in the main navigation. Krista Seiden (a former Googler who worked on what is now GA4) has a simple write-up that covers some of these basics visually to help you translate between the two paradigms
2. Learn Explore reports, which let you build your own free-form analyses or choose from several useful starting templates that can then be customised. In general, you’ll find yourself using Free-form analysis for a lot of things you would have used segments for in the past (but with far less likelihood of sampling!). Optimise Smart has a good step-by-step visual guide on building your first Free-form analysis

There are also differences in the underlying data. For example, expect to see a small drop in the number of sessions reported. This sounds like not what you want, but is actually a good thing—people clicking away from and then back to your site, or actions that override campaign parameters no longer create duplicate sessions in GA4 . . . making your data more accurate.

We’re hoping to find several additional ways to help make this transition easier for nonprofit organisations like yours (such as a webinar and training course for NextAfter Institute members). Until then, I’ve included additional links to resources on other sites below this post. Many of these folks are fantastic to follow on Twitter if you want to know more about GA, analytics in general, or what’s happening in the space.

Final thoughts . . .

I know this has been a lot. And was likely unexpected. But you’ve got this. And you can take the opportunity to set the future versions of you and your organisation up for success—and that’s a worthy cause unto itself.

Best of luck, and happy analysing!

Additional Google Analytics Resources:

• Google’s official announcement
• #Measure Slack (Slack community for marketers, and a great place to ask questions and help each other—warning it takes a while to get approved and added)
• Digital Debrief/Krista Seiden (former Googler who was on the product team for what is now GA4)
• Simo Ahava’s Blog (a fountain of wisdom on GA, and THE expert on Google Tag Manager)
• Charles Farina on Twitter (GA specialist at AdSwerve, great source of GA-related information)
• Analytics Mania/Julius Fedorovicius (lots on more advanced GA topics, including working with data in BigQuery)