How good Is your web analytics implementation?

*This blog has been reviewed in September 2024

The problem with overconfidence

In practice, we often see high confidence in web analytics data, even when the implementation has serious flaws. This happens when companies assume that setting up a web analytics tool is easy. As Microsoft’s Ronny Kohavi aptly puts it:

"Getting numbers is easy, getting numbers you can trust is hard."

On the flip side, some companies lack confidence in their web analytics data, even when it’s more reliable than they think. This often stems from small discrepancies between site data and backend systems, with an unrealistic expectation that web analytics data should be 100% accurate.

Accuracy vs. precision

Web analytics data is never 100% accurate, but it’s precise enough to generate valuable insights for your business. The goal isn’t to perfectly track every interaction but to understand visitor behavior and improve your website accordingly.

Why isn’t web analytics 100% accurate? Here are two examples:

• Multiple devices: A user visiting your site from different devices (without logging in) won’t be recognised as the same person.
• Ad-blockers: Some users block web analytics tools entirely, resulting in missing data.

But is this really a problem? Not necessarily. Web analytics is about identifying trends and improving key metrics (KPIs). For that, precision—consistent data collection over time—is more important than pinpoint accuracy.

Of course, accuracy still matters. For example, if your analytics tool captures only 70% of actual transactions—or worse, 125%—your data won’t reflect reality. So, while you should aim for accuracy, don’t stress over perfection.

Reliable web data doesn’t come automatically

A web analytics implementation won’t guarantee accurate and precise data on its own. Adding basic tracking code to a site might seem simple, but creating a system that delivers useful insights is far more complex. Mistakes happen frequently. Here are some common issues:

• Standard page views: Many websites only track standard page views, which limits the value you can extract. For instance, it won’t tell you if an article was read or how many products were sold.
• Personally identifiable information (PII): Some URLs contain PII, which isn’t filtered out correctly. This is not only unethical but violates both GDPR and Google Analytics’ terms. Failing to filter out PII can result in Google deleting your account.
• Campaign tracking: Evaluating marketing campaigns requires knowing which campaigns brought users to your site. If your tracking is off, your data may be unusable.
• Single page applications (SPAs): Many sites have SPAs, which pose unique challenges for web analytics. It’s easy for things to go wrong here.
• Broken implementations: Sometimes, an initially well-implemented setup breaks over time due to changes in website code, and no one notices.
• E-commerce tracking: Web analytics tools often require very specific structures for e-commerce tracking, and errors here are common—such as duplicate transactions or missing product dimensions.

Research on the quality of Google Analytics implementations

At the Superweek Analytics Summit in 2019, Brian Clifton presented his research on the quality of Google Analytics (GA) implementations. His findings were alarming, even among major brands.

The study assessed 75 commercial websites, all of which are leaders in their fields, with many receiving hundreds of thousands or even millions of visitors monthly. Despite this, the average Data Quality Index score was just 35.7%, indicating widespread deficiencies. Even the top performers scored no higher than 73.1%, showing that even market leaders struggle with web analytics accuracy.

PII in Google Analytics

Clifton’s research also revealed that 1 in 5 companies stores personal information in Google Analytics, such as email addresses or even passwords, which violates both privacy laws and Google’s policies.

Additional findings include:

• Only 19 out of 31 e-commerce companies had proper transaction measurement in place.
• 85% of companies had flawed or missing visitor segmentation, one of the most critical aspects of web analytics.
• Campaign tracking was often mishandled, with only 18% of companies implementing it correctly.

Take a critical look at your web analytics implementation

Too many companies pay insufficient attention to the quality of their web analytics. This is surprising, given that:

• For many, the website is the primary point of contact with customers.
• Many companies rely on web analytics data to evaluate experiments and personalise the customer journey.

So, how reliable is your web analytics data? Do you know where improvements are needed?

While automated tools like Verified Data can provide a rough check, a good web analytics implementation is tailored to each site. A thorough manual audit is essential to uncover specific issues and make meaningful improvements.

If you want your web analytics implementation reviewed in detail, feel free to contact us.

Receive data insights, use cases and behind-the-scenes peeks once a month?

Sign up for our email list and stay 'up to data':