Google Analytics has been the industry default for years. But privacy-first browsers, consent requirements and a fundamental shift in how people discover websites through AI assistants are making it harder to trust GA’s reports.
Here’s a look at the common Google Analytics tracking issues, and why its alternatives are worth a serious look.
Key takeaways
- In terms of adoption, no analytics tool comes close to Google Analytics. It powers close to 44% of the entire web. This share rises to nearly 79% among websites using any kind of analytics tool.
- Gaps in analytics tools translate directly into misread trends, wasted budget and missed opportunities, particularly as AI assistants become a significant and growing traffic source.
- The most accurate picture of your website’s performance comes from tools built for how the web works today with native AI traffic tracking, reliable cross-domain handling and no reliance on statistical modelling to fill the gaps.
- Alternatives like Matomo can be a quicker, easier solution to these problems while also providing other long-term benefits.
Why GA tracking is harder than it looks
Accurate analytics isn’t just about installing a tool. It’s about understanding where measurement breaks down and choosing a platform built to minimise those breakdowns from the start.
Google Analytics works by loading a JavaScript snippet when a page loads in a browser. But for a variety of reasons, this snippet fails to get triggered or gets manipulated incorrectly.
The most visible culprit is browser-level blocking. Safari and Firefox actively limit how GA’s tracking scripts behave. Safari’s Intelligent Tracking Prevention (ITP) caps cookie lifespans, meaning returning visitors get counted as new ones after as little as seven days. Ad blockers and script blockers strip GA’s tags even before they get triggered.
Then there’s the cookie consent banner problem. GA was built for a web where tracking was invisible by default. GA uses cookies, which require consent under GDPR. This fundamental shift has made accurate tracking significantly harder for any website with a meaningful EU audience.
The web’s architecture has also shifted in ways GA wasn’t originally built for. Single-page applications don’t reload between pages, so GA’s page-load model misses navigation unless you manually instrument every route change. Additionally, cross-domain journeys break session continuity, inflating new user counts and fragmenting funnel data.
GA4 attempts to patch some of these gaps through behavioural modelling and consent mode estimates. But this has its own flaws. The decisions are made based on figures that are a mix of measurements and educated guesses.
Google Analytics tracking issues
Google Analytics is installed on hundreds of millions of websites. But its widespread adoption doesn’t guarantee the accuracy of its reports. Here are six structural problems that create a false reliability of the GA report.
1. Browser privacy settings
GA is caught in the middle of the browser and privacy war. For instance, Firefox’s Enhanced Tracking Protection (ETP) and Safari’s Intelligent Tracking Prevention (ITP) limit cross-site tracking.
ITP caps the lifespan of GA’s client-side cookies at seven days, and in some cases, as little as 24 hours. This means a visitor who comes to your site on a Monday and returns the following Tuesday is treated as a completely new user.
Privacy-first browsers like Brave block tracking scripts by default through the Shields feature. This prevents a business from tracking how many visitors are using the application.
The systematic undercounting of returning visitors, broken attribution chains and session data make the report unreliable. It can make an entire funnel analysis difficult.
2. Consent cookie banner rejection
GA4 uses cookies that, under GDPR and other global privacy regulations, require informed user consent before they can be set. A recent study shows that around 97% of data gets lost due to cookie consent behaviour.
This makes it harder to understand the behaviour of the audience with a skewed dataset. GA4’s Consent Mode attempts to partially address this by modelling the behaviour of users who decline. But modelling is not measurement, and this leads directly to the next problem.
3. Reliance on modelled data masking gaps
When GA4 can’t track a user because they declined consent or because their browser blocked the tracking script, it doesn’t leave a blank in your analytics report. Instead, it fills the gap with a statistical estimate based on the behaviour of similar users who did consent.
Though it sounds reasonable in theory, it creates a significant trust problem. The modelling itself relies on users who opted in being representative of those who didn’t. It’s a false assumption.
People who decline tracking tend to behave differently online. They use different browsers, devices, and browsing habits. The model is built on a biased sample.
4. Limited AI-driven Insights
According to a study conducted by Ahrefs, around 63% of websites today receive traffic from AI platforms like ChatGPT, Perplexity and Gemini. Although Google Analytics has recently introduced the functionality to track AI referrals and traffic, it’s quite limited.
Some of these platforms strip referrer headers entirely or use non-standard signatures that make GA4 difficult to track AI chatbot referrals. This makes traffic figures unreliable as AI referrals grow. For instance, a spike in direct visitors might mean your brand awareness is increasing, or it might mean Claude started recommending your brand more often. You have no way to tell.
Besides, automated traffic from AI agents such as bots and crawlers sent by AI systems to gather information goes completely unreported.
Matomo now tracks both AI assistant referrals and AI agent traffic separately, giving you a clear view of where human visitors are actually coming from and filtering out automated activity that would otherwise distort your data.
5. Cross-domain checkout journeys
In practice, a user journey is typically shared across multiple domains. For instance, a visitor might browse a product on one domain, be redirected to a third-party payment platform like Stripe or PayPal, and land on a confirmation page on another subdomain or an entirely separate domain.
GA4 may record this interaction as two separate users. By default, browser cookies cannot be shared across different domains. To overcome this, GA4 uses the _gl parameter, which appends a user identifier to the URL when a visitor moves between domains. But in case the link parameter _gl is not configured properly, it can inflate user count and distort conversion rates.
Cross-domain tracking in GA4 can be configured to work more reliably, but it requires careful implementation, ongoing maintenance and a clear understanding of every domain. This creates a significant technical overhead for most teams.
6. Behavioural analytics gap
GA4 tells you a visitor arrived, clicked and left. It has no native way to show you what that experience actually looked like to the user. It has no heatmap features to show up where attention is clustered, no session recordings to show where users hesitated or gave up.
This leads teams to diagnose conversion problems and UX failures entirely from aggregate numbers such as bounce rates, exit rates and engagement time without ever showing behavioural reality behind them.
A page with a high exit percentage could have a broken button, a confusing layout or content that simply doesn’t match visitor intent. This is a critical gap for SaaS products and subscription businesses where retention matters as much as acquisition.
Can you fix these tracking issues and still use GA4?
Yes! You can. But the fixes for GA4’s core tracking issues need significant ongoing investment. Server-side tracking, proper cross-domain linker configuration, custom AI referral filters, and consent mode tuning each require technical knowledge, developer time and active maintenance.
Though it’s possible to get GA close to reporting accurate data, there are architectural issues like no native way to separate AI assistant referrals from direct traffic or to identify and isolate automated AI agent visits. Even a well-configured GA4 implementation still relies on modelled data to fill consent gaps.
Product teams should question whether the tool they’ve inherited is still the right foundation for it. The effective strategy is to track analytics with a privacy-first analytics platform like Matomo, one built for how the web works in 2026. Most teams that do it don’t go back. Some teams often run Matomo alongside Google Analytics for their product insights.
Matomo not only solves common tracking pitfalls with GA but also frees developers’ time to ship awesome features to their applications instead of fighting with and configuring analytics.
Why marketing teams are adopting GA alternatives
GA’s tracking issues aren’t new. Stricter privacy regulations, consent economics and an AI-driven traffic landscape are pushing conscious marketers and product teams to look for alternatives.
The compliance problem pushed the game
GDPR and equivalent data privacy laws have fundamentally changed what GA can measure. When around 97% of data is vanished, the data you see on analytics reports is a prediction. Teams making channel attribution decisions, planning campaign budgets or diagnosing conversion drop-offs are doing so on a partial picture, often without realising how partial it actually is.
Several European data protection authorities have already ruled against GA use, with Austria, France and Italy among those to take action.
AI traffic made the attribution problem impossible to ignore
GA was built around the decade when search dominated the web. The core tracking principle was that when someone clicks a link, a session should start and a source should be recorded. It worked well enough when traffic sources were predictable.
AI assistants broke that model. Though Google Analytics recently introduced a dedicated AI Assistant channel, understanding the full impact of AI on customer journeys remains challenging. Not every AI source is recognised automatically and many user journeys begin long before a visitor clicks through a website.
The organisations best positioned for an AI-driven future will not simply have access to more data. They need to have access to trustworthy data they own, understand and can confidently use to support decision-making.
The alternatives have matured
Years ago, switching from GA meant accepting significant trade-offs in features, integrations or ease of use. That’s no longer true. Matomo now offers full-featured analytics from session recording, heatmap to funnels and multi-touch attribution, all without data sampling, without reliance on consent-based modelling or modelling dependencies that question GA’s reliability.
Critically, Matomo tracks AI assistant and AI agent traffic natively, with no complex configuration needed. You can see exactly how much of your traffic is coming from ChatGPT, Perplexity or Gemini, and separately identify automated AI agent activity that would otherwise pollute your behavioural data.
The teams moving fastest aren’t waiting for GA4 to solve these problems. They’re building their reporting on foundations they can actually trust.
From tackling issues to better decisions
The six Google Analytics tracking issues covered in this article contribute to a more serious problem: the numbers in your analytics reports aren’t accurate. Incomplete data isn’t just an analytics problem. It’s a budget problem, an attribution problem and a strategy problem.
Every decision that flows from a report built on partial observation leads to an unmatched outcome. The practical step is to measure the gap before deciding whether it matters for your business.
Want accurate and unsampled analytics that give you the full picture of user behaviour, including heatmaps, session recordings and A/B testing? Start your 21-day free trial of Matomo Cloud today. No credit card required.
FAQs
How do I check if Google Analytics event tracking is working?
The quickest way is to use GA4’s built-in DebugView. It shows a real-time stream of events firing from the browser, so you can confirm whether a specific interaction, such as a button click, form submission or scroll depth trigger, is being recorded as expected.
You can enable debug mode by installing the Google Analytics Debugger Chrome extension or by setting debug_mode to true in the GA4 configuration tag.
Another way is to use the real-time reporting in Google Analytics. It shows events as they fire.
A few things to check if events aren’t appearing are:
- Confirm the tag is published in GTM and not just saved
- Verify the trigger conditions match the exact interaction you’re testing
- Check that your browser isn’t blocking the GA script via an ad blocker or privacy extension
What are the disadvantages of Google Analytics?
The most significant disadvantages of Google Analytics are:
- Data sampling: Once your data exceeds certain thresholds, particularly in custom reports and explorations, GA4 stops analysing and works from a representative sample instead.
- Consent requirements: Consent requirements under GDPR and similar regulations mean that anywhere between 40-60% of visitors who decline your cookie banner become invisible to GA entirely.
- Cross-domain tracking: Google Analytics cross-domain tracking is fragile by design. The _gl parameter that stitches sessions together across domains breaks silently when redirects or JavaScript navigation strips it from the URL, leading to inflated new user counts and broken funnel analysis.
- AI traffic: AI traffic is often miscategorised as direct traffic because most AI assistants don’t pass standard referrer headers.
- Google’s data ownership: Your analytics data lives on Google’s infrastructure, is subject to Google’s terms and contributes to Google’s broader data ecosystem, which is an issue for organisations with strict data governance requirements or operating in heavily regulated industries.
What data is Google Analytics unable to track?
Google Analytics usually measures the event that happens in a browser after a page loads. However, GA4 also does offer server-side tracking to handle backend transactions, offline conversions and API-level events. Though it still requires significant developer resources to implement it correctly.
But Google Analytics doesn’t track how people actually experience your website, where frustration builds, where they hesitate over something confusing, or repeatedly try something that wasn’t working. That level of visual detail requires heatmaps and session recording, neither of which GA offers, but Matomo does.
In GA’s report, a visitor who clicked a broken button four times before leaving and a visitor who found exactly what they needed look identical.
Why does GA4 feel slower than the old version?
It’s because GA4 really is slower than Universal Analytics by design. GA4 processes data through a more complex pipeline. Event-based data collection and machine learning processing for modelled metrics are the major factors contributing to reporting latency.
Standard reports typically reflect data with a 24-48 hour delay, and exploration reports can take longer. If real-time data matters to you, GA4’s Realtime report remains fast, but it’s limited in what it can show.
Can I use Google Analytics for free?
Yes, the standard version of GA is free. However, its enterprise tier, Google Analytics 360 (GA360), is a paid product aimed at large organisations with advanced needs, but the vast majority of websites run on the free version.
It’s worth knowing that a privacy-friendly alternative to GA, Matomo On-Premise, is also completely free. You host it yourself, own your data outright and get core analytics functionality, including features like event tracking, goal tracking and audience segmentation.
