I’m yet to work on a website that doesn’t utilise Google Analytics (GA). In 16 years, I’m yet to find one.
GA is ubiquitous.
Despite the ubiquity, I don’t think our industry speaks openly enough about a new fact of life, though; Google Analytics is now an opt-in-only resource.
If a website follows cookie laws (many don’t), then GA is rendered far less useful than it used to be.
The new opt-in world of analytics
The result is that for brands who follow the guidelines to the letter, GA data is significantly less reliable.
The context is this: In (very) simple terms, data protection rules stipulate that you need express consent from your audience to use cookie-based tracking on your website (e.g. Google Analytics).
That means user tracking is opt-in, not opt-out like it used to be.
If a user opts in, you track them. If they do nothing, you don’t.
This is a massive challenge for brands and I don’t think it’s openly discussed enough online.
Perhaps it is discussed, but I’m yet to find anywhere that puts it plainly that web analytics is an opt-in game now.
The result is that for brands who follow the guidelines to the letter, GA data is significantly less reliable. Sometimes to the tune of a loss of 80% of visitor data.
It’s worth saying here, that despite the abundance of ‘cookie policy’ messages you see online, many still don’t control cookies in the way the rules require.
The challenges of working in an opt-in analytics world
Opt-in analytics mean that your GA data is unreliable, for a few reasons:
1. You are working with a smaller sample size
The smaller the sample size, the more difficult it is to identify insights.
2. The users who ‘opt-in’ may not be representative of your audience-at-large
That’s a clunky way of saying you’ll be suffering from sampling bias. Your GA account essentially becomes a focused study of a certain type of person, rather than all types of people.
3. Summary metrics (like conversion rate) become inconsistent
It’s difficult to calculate the average of something when you cannot see the total.
Ultimately, if you follow the rules and require your audience to opt in, your GA account becomes far less reliable.
In fact, if we’re serious about the accuracy of data and honest about the decisions we make using it, opt-in analytics makes GA4 useless.
What’s the solution?
Brands are going to need to accept the new reality. Step one is going to require you to implement an opt-in model if you haven’t already.
That way, you comply with GDPR.
Secondly, your marketing and leadership teams are still going to require access to reliable data on website and marcomms performance.
There are a few solutions to consider.
Consent mode (via GA4)
Google’s solution is called Consent Mode. In theory, it provides ‘modelled’ data to fill in the gaps left by the visitors who opt-out.
It sounds plausible, however, in practice, I’m personally yet to see it working.
For now, I’m putting this into the ‘unreliable solutions’ folder.
Server logs
I’ve got a sneaky suspicion that after about 20 years, server logs are going to make a bit of a comeback.
In the days before GA, server logs paired with simple interfaces to monitor them enabled website owners to track ‘hits’ and see which pages were being viewed.
Assuming they are configured to anonymise visitor information (such as IP address), server logs present a compelling solution to the current challenge.
They aren’t user-friendly though.
Privacy-compliant analytics
It’s time to start viewing GA as an opt-in-only technology, because in reality, that’s exactly what it has become
The solution we’re hitching our wagon to now focuses on privacy-compliant analytics; an analytics solution that doesn’t rely on cookies.
We’re in the process of rolling out an Indulge-hosted, EU-based solution built using Plausible, an analytics platform.
Plausible provides a lot of the insights that GA4 is used for, but with a few key privacy-related changes:
- It uses a small script to track website visits, rather than installing cookies on the browser’s machine
- It doesn’t store user IP addresses meaning individual users are not tracked over time
- It doesn’t create persistent user profiles meaning that users are treated independently on each visit; it does mean you cannot track repeat visits, but that’s a rule to protect user privacy
- It uses referrer URLs and UTM parameters to track visitor sources, rather than relying on third-party trackers
- Data is aggregated into broad statistics, not individual user statistics
We’re actively rolling out our privacy-compliant analytics solution for clients.
As the industry shifts, it’s time to start viewing GA as an opt-in-only technology, because in reality, that’s exactly what it has become.
