3 Tips for Planning Your Migration to Google Analytics 4

Planning your migration will avoid disaster.

Google Analytics 4 was released in October 2020, and the time has come to migrate away from the legacy version if you have not already. But before you start creating new tags, you will need to make a plan that considers the following:

  • What business questions do I need GA4 to answer, and do I need to migrate everything in place today?
  • How long should I expect to continue using the old version of Google Analytics?
  • How can I prepare GA4 for the privacy measures that will be rolled out by browsers in the next year or two?
  • How do I create a transition plan that allows my team sufficient time to learn the new tool?

In this collaboration with Tim Wilson (co-host of the Digital Analytics Power Hour), we will present 3 key considerations that any business should make before installing Google Analytics 4. If you are looking for a more technical guide, see my post on Translating Website Tags from Universal Analytics to Google Analytics 4.

LAST UPDATE (Dec 30, 2021)

Updated the product roadmap.


To begin, let's start by defining our scope: 

Tip #1 – Document your Business Questions

This is the core of the plan, and it's an opportunity to step back and think as broadly and creatively about the site as you can. Unfortunately, the more broadly and creatively you think, the more work you've set up for yourself in Tip #2. But, the more narrowly you think, the less business value you will ultimately be able to realize from your implementation.

There are two main aspects of developing and documenting your business requirements:

  1. Capture the key performance indicators (KPIs)
  2. Create an exhaustive list of questions that you or your stakeholders may have about these KPI's that can be answered using GA4

First: Capture the Key Performance Indicators

KPIs are the most critical conversions or behaviors that need to be captured and readily accessible—not every possible metric that could be recorded.

When creating your list of KPIs, consider each of the main internal stakeholder groups: what are the site KPIs for executives vs. digital marketers vs. the UX team vs. the ecommerce team? It is fine (and expected) for KPIs to overlap across groups.

If you are not sure if you have captured the site KPIs for a particular group, ask them. A good way to frame the discussion is by asking, "What are you responsible for or expecting the site to deliver? What metrics tell you if you have done that?"

These KPIs are metrics only (form completions, orders, revenue, conversion rate, etc.), so for each KPI identified you should think through how the metric will need to be sliced (by channel, by device type, by campaign, etc.). This process will give you the dimensions that will be used in reporting and analysis through breakdowns, segments, and filters, and many of them will be duplicated across many of the KPIs.

Second: Create an exhaustive list of questions that you or your stakeholders may have about these KPI's that can be answered using GA4

The KPI’s that you captured in the previous phase are thought starters for this exercise. The goal is to anticipate the questions that marketers will have when attempting to understand how KPI’s are performing.

This is the tedious work that many Analysts do not do, but be as specific as possible with the questions. These are use cases that will be prioritized and used to both inform and then vet the solution design. As an example:

"How are users moving through the site?"

This is a broad and unhelpful question. On the other hand:

"What pages on the site do users visit before initiating the checkout process?"

This is much more specific, and will provide the detailed information that you will need when designing the solution (see Tip #2), because this question would remind you that you need to cleanly capture the "initiation of the checkout process" in the implementation.

Prioritize Your Questions

This list of questions can and should be lengthy, and it is fine to capture fairly niche questions. The last step in documenting your business requirements is to assign a priority to each question. This way, when you start developing your solution design, you will know that high priority requirements need to be both captured and easily accessible, while low priority requirements may be omitted if including them adds an unreasonable cost to the implementation or its ongoing maintenance. Low priority requirements can also be ones that are technically met, but that will require more work/manipulation of the data on the part of the analyst or the stakeholder.

NOTE: At the end of this process you might discover that many of the items you are tracking in your Analytics tool today are not included or are low priority. This is common, and usually indicates that too much time has passed since anyone went through this exercise.

Now that you have a good idea of how your business will use Google Analytics, the next step is to document exactly how data should be captured to make this happen. This can be done with a Solution Design.

Tip #2 – Create a Solution Design with Test Instructions

A Solution Design is a document that provides explicit detail on your Analytics implementation. The contents of a Solution Design may vary slightly depending on the company and platform, but at a minimum it should contain the following 3 components for Google Analytics 4:

  1. A list of the business questions that you documented during #1 above
    -Example: “How much does it cost me in Google Ads per conversion?”
  2. A list of the variables that will be necessary to answer the business questions
    -Example: An event should fire when the user completes a conversion
  3. Instructions for testing that your data is being collected properly
    -Example: Navigate to https://example.com and submit the form to see the conversion event fire.

To get you started I’ve created a sample Solution Design template that can be copied and modified. However, I highly recommend that you hire an experienced consultant to help create this, because creating a good quality Solution Design is more of an art than a science.

Characteristics of a Good Solution Design

A good Analytics implementation is simple, flexible, and considers user privacy. 

Simplicity is important because a high volume of tags and custom coding requires maintenance, creates poor performance, and it is almost never necessary.  But, a good Analytics implementation is also flexible enough to answer a wide range of business questions.  

Many inexperienced Analysts are eager to please, and they create new custom tags to address every business question that is presented to them.  Instead, make it a goal to create an Analytics Solution Design that finds a healthy balance between flexibility and simplicity, and recognize that doing this well requires experience.

Finally, a good Solution Design considers user privacy.  Your plan should capture only the data that you need, and it should also anticipate more restrictions on browser storage and IDFA’s to be imposed in the future.  You can do this by remembering the three “nots”:

NOT NECESSARY: Data that is not necessary to the business should not be tracked. Being “data driven” does not mean that you “track everything”.

NOT ALLOWED: Make sure that you properly disable tracking when requested by the user, and add tests to validate that this is working as expected.

NOT FOR PERSONALIZATION: Some of the data you capture in Google Analytics 4 might not be appropriate for use in personalized advertising. Make sure to mark these events as “NPA” from the “All Events” report.

When your Solution Design is complete, the actual implementation is as easy as building a table from Ikea.

Ikea paradox

The most challenging part of migrating to Google Analytics 4 begins after you've deployed the tags, so this is where Tip #3 comes in:

Tip #3 – Create a Vision and Set Expectations

It is difficult and time consuming to familiarize yourself with any Analytics tool, so the most effective leaders create a vision that gets team members motivated to put in the effort, and set expectations with stakeholders about how long this should take.

To make this more complicated, as I write this in November of 2020 the team at Google is working on a very large number of features that have not yet been released to Google Analytics 4. As a result, your migration plan might be delayed or prioritized to allow the product team time to roll out product updates.

Here are three important factors to consider when creating your vision:

  1. What are your key priorities for the next 6 to 12 months?
  2. When will it be feasible for the business to fully adopt Google Analytics 4 without relying on the legacy implementation?
  3. How much time will it take my team to be sufficiently proficient in Google Analytics 4?

Let’s discuss these one at a time.

What are your key priorities for the next 6 to 12 months?

This question is more broad than Google Analytics, because it’s about determining where the GA4 migration is prioritized within your overall plan for the year.

If you have a good handle on this already then you are free to skip to the next section. Otherwise, imagine what a highly functioning Analytics team would look like in your organization across the following categories:

Data Integration Do data silos currently limit our ability to answer the business questions documented in #1 above?
Standard Reporting & Dashboards Do those stakeholders who rely on our team to measure progress toward KPI’s have easy access to the data they need?
Experimentation & Analysis Is our team helping stakeholders ask good questions, define testable hypotheses, and find answers that drive the business forward?
Governance Do we have clear control over who can access data?
Culture and Adoption Do the decision makers in our organization depend on the data we provide?

You may add other categories to this list, but the idea is to identify weaknesses and focus your efforts where they will have the largest impact. Migrating to Google Analytics 4 will very likely help improve data integration and analysis, but it will not solve the problems that you may have with culture or governance.

When will it be feasible for the business to fully adopt Google Analytics 4 without relying on the legacy implementation?

Several key features were not included with the launch of GA4, which has led many Analysts to conclude that the product was taken out of beta too early. I disagree with this assessment, but I am also advising my clients not to plan to completely remove the legacy version of Google Analytics for several months after GA4 is installed.

Google’s official product roadmap is confidential, but to assist with planning I'm including my expected timelines for when major features will be fully available (I will keep this updated as things change):

H1 2022 >= H2 2022
SA360 and DV360 Integrations X
Management API X
Attribution Modeling X
Salesforce Integration X

As you can see above, organizations that rely heavily on remarketing audiences will likely be the last to fully migrate away from Universal Analytics. However, these organizations should not wait to install Google Analytics 4, because the tool will need time to build audience pools anyway.

The recommendation that I provide my clients is to start small. Install the basics today and plan to build out your event tracking slowly over the course of 2021 while you continue to report from a legacy Analytics property.

How much time will it take my team to be sufficiently proficient in Google Analytics 4?

As I’ve stated many times on this blog: Google Analytics 4 is very different from the legacy versions of Google Analytics, and it will take a combination of formal training and hands-on experience for the typical Analyst to become proficient in the tool.

To get started, I recommend that organizations hire a company like Search Discovery to conduct GA4 training for your team. And of course, I’ve created https://ken-williams.com/ to be a resource for Analysts ramping up on GA4. For those who are just starting out I recommend exploring the 4 topics below (and signing up for my newsletter):

1) The Event Driven Data Model – Google Analytics 4 uses Firebase Analytics on the backend, which means that data is formatted in a new structure that I have described in detail HERE.

2) Analysis – The Analysis Hub in Google Analytics 4 contains very powerful tools for exploring your data, and any Analyst who relies on GA4 will need to invest the time it takes to familiarize herself with these.

3) Campaign Reporting – The dimensions (ex. “source / medium”) and metrics (ex. “bounce rate”) you use to evaluate campaign performance have changed with GA4. I have detailed the changes to Bounce Rate HERE, and be sure to sign up for my newsletter to read an upcoming article that explains how to use the new dimensions.

4) BigQuery (optional) – The BigQuery schema is very different for GA4, so if you are currently using this feature you will need to understand these changes. Fortunately I’ve created some helpful content to get you started HERE.

So, back to your question: “how long will it take?”. This obviously depends on skill level, experience, and the time that the Analyst invests. Here is the advice I have for managers:

  • Get each member of the team excited about the place where they fit into your vision
  • Set high expectations, and provide the support needed to meet them
  • Expect your team to fall back on the old version of Google Analytics for at least 6 months