Today, every strategic decision is based on data. Yet, although analytics platforms and marketing campaigns represent an investment that most business lines must account for at year-end, the quality of the figures collected is not always satisfactory.
Collecting data is not enough: it requires correct implementation based on quality standards, regular validation and robust governance. Without these conditions, errors have a knock-on effect: distorted KPIs, misinterpreted funnels, incorrect decisions and user journeys optimised on the basis of misleading insights.
In a highly competitive ecosystem, overlooking a robust tracking architecture that is scalable is out of the question. It therefore makes sense to implement it in line with industry best practices.
Why is it essential for the business?
Marketing, sales, product and customer experience: there is no area that good data cannot improve. But what are the main benefits?
Data tracking for data-driven strategies
Navigating blindfolded is never a good idea, let alone a strategy.
Data-Driven does not mean using data as retrospective confirmation, but placing it at the heart of the decision-making process. Data provides the basis for validating hypotheses, identifying new growth opportunities, anticipating trends and responding proactively to change. Data is the best compass for navigating safely.
Understanding user behaviour
Analysing navigation flows within websites or apps is a strategic starting point. Consider, for example, the internal search function: if the data shows that many searches return no results, we can address this by suggesting alternative content or improving product indexing. The result? A smoother user experience and higher conversion rates.
Measuring business objectives
Having clear KPI that are accurately tracked allows us to monitor every stage of the funnel and act on what really matters. If, for example, the objective is to increase the number of enquiries submitted through the contact form, but actual submissions have not been tracked correctly, all downstream data will be distorted.
Optimising marketing campaigns
Without reliable data, even the best media strategy is operating blindly. An accurate attribution model – one that takes actual conversions into account, rather than clicks alone – is essential for assessing the effective contribution of each channel and allocating the budget accordingly and optimally.
Segmenting audiences accurately
When we talk about a target audience, we often mean a broad, general audience, based on factors such as age, gender or geographical area. Although useful in some situations, this approach can be vague. By contrast, segments identify shared behaviours, interests or needs.
In summary, segmentation means speaking directly to those who are interested and personalising their experience, rather than spreading resources across an overly broad audience.
Identifying technical or usability issues
Imagine tracking the purchase funnel and discovering that most users leave at the checkout page. There could be a technical issue, a UX issue or an element that creates mistrust. Without tracking, all of this remains invisible.
What to avoid when tracking data
Many believe they have everything under control, only to discover – too late – that the data collected does not reflect reality. Empty reports, inconsistent figures and campaigns with no attribution. Having active analytics tools does not guarantee useful or reliable data: the quality of the implementation is what makes the difference.
We often assume that everything is working as it should, but even minor technical errors can undermine the quality of the entire tracking ecosystem.
Common (and underestimated) errors
Here are some of the most common errors we encounter when the data appears to be there, but does not really tell us anything.
- Google Tag Manager is not installed correctly. It is a key component of modern tracking, but if it is not installed correctly, the data collected may be incomplete or inconsistent.
- Incorrect DataLayer configuration. The DataLayer is the foundation for any robust tracking setup in Google Tag Manager. If it is not designed and tested correctly in line with objectives and KPIs, it can result in incomplete, inconsistent or difficult-to-use data.
- Consent Mode missing or incorrectly configured. This is not solely a privacy issue: Consent Mode is also crucial to data quality. If it is incorrectly configured, tracking will be inaccurate, directly affecting campaign attribution and the ability to derive value from data.
- Underestimated pre-go-live testing. Once the system is operational, some steps considered secondary may be underestimated, with attention focused exclusively on key areas such as high-level KPIs. At this stage, essential checks may be overlooked, including how the platform behaves across different browsers, whether events are triggered correctly and the user experience across all devices.
- No comparison with back-end data. GA4 is an essential tool, but it is not infallible. Relying unquestioningly on its figures without comparing them with actual data (CRM, e-commerce, ERP, etc.) is risky. A typical example is when GA4 reports 100 purchases, while the e-commerce system records 73. Without a comparison, the discrepancy goes unnoticed. Comparing the data helps us validate tracking quality and identify inconsistencies that could distort key KPI.
How to set up a robust tracking system
There is no doubt that digital complexity has increased, with new platforms, increasingly fragmented touchpoints and non-linear user journeys. The opportunities have also grown, provided that data tracking is robust, consistent and free from structural errors.
An incorrect value today is more than a technical error: it leads to a poor decision tomorrow. It means a misallocated investment, an underperforming campaign, a lost customer.
This is why building a reliable tracking ecosystem is an ongoing process rather than a one-off activity, requiring standardisation, monitoring, testing and continuous comparison with the actual figures. What steps should we follow?
1. KPI analysis and measurement plan
Every tracking implementation should begin with objective-led measurement, formalised in a documentation tool that defines: what do we want to measure? Which events are genuinely relevant to us? Which business objectives do they support?
2. Clear and robust consent management
A robust consent management configuration allows tracking to operate in line with company guidelines while complying with the GDPR and applicable regulations.
3. Comprehensive mapping of the user journey
Where do users come from? Where do they click? What do they do before converting?
4. Google Tag Manager as a central hub
Managing tracking through GTM gives us an overall view and enables faster monitoring and updates, while also making many activities independent of platform development.
5. Server-side tracking
Today, it is no longer a “nice to have”. It is the standard, and the benefits are numerous:
- bypasses Ad Blockers and ensures more accurate data collection, particularly for ADV platforms;
- improves website performance by reducing loading times, as tags are managed by the server rather than the browser, reducing the load on the user side;
- provides greater control over several aspects, particularly the management of privacy and bot traffic, as well as the option to anonymise data and apply other customisation policies;
- reduces the risk of exposure by keeping API keys and sensitive endpoints hidden in the backend;
- allows us to integrate data from external sources such as CRM, offline sales, loyalty systems and other internal databases, providing a complete view of the user’s interaction with the brand, regardless of the channel.
6. Google Ads
Connecting Google Analytics 4 to tools within the Google ecosystem, such as Google Ads, Search Console, AdSense and others, is always best practice for gaining a broad overview of digital performance.
However, to optimise campaigns effectively, across Google Ads as well as Meta, LinkedIn and other platforms, the first step is to track conversions correctly. Only then can we accurately measure their effectiveness.
In this context, configuring Google Tag Manager (GTM) is essential for flexible, centralised event tracking (purchases, registrations, page views, etc.), while also simplifying data transmission to the various advertising platforms.
By integrating with Google Ads, GTM enables the implementation of advanced features such as dynamic remarketing, which displays personalised ads based on the products or content viewed by users, and enhanced conversions which, with the user’s prior consent, allow additional data (such as an email address or telephone number) to be collected, even when cookies are limited or unavailable.
This process is therefore useful both for ensuring tracking accuracy and for increasing the effectiveness of the bidding strategy.
7. Integration with BigQuery
BigQuery allows us to overcome the limitations of GA4, which only provides aggregated data. The tool gives us access to 100% of the raw data, eliminating approximations and enabling a detailed analysis of each individual event.
With BigQuery, we can:
- conduct advanced analyses that cannot be performed using standard reporting tools;
- easily integrate data from external sources (CRM, offline sales, internal databases) to provide an omnichannel view of the user;
- optimise queries and data structures, improving the speed and stability of reports in Looker Studio, Microsoft Power BI and Tableau;
- automate advanced dashboards and reports, with no limits on customisation.
In summary, if data is at the heart of decision-making, BigQuery is the engine that drives it.
8. Monitoring
Tracking is never a one-off exercise. Any change to the website, new feature or technical update can directly affect data collection. It is therefore essential to have a continuous system for control, monitoring and alerting.
From theory to practice: examples of data tracking ecosystems
Until a few years ago, with Universal Analytics, the data collection and analysis process was fairly linear and easy to manage: it mainly involved configuring the code correctly, defining a few basic goals and monitoring the results through a relatively simple and familiar interface.
A process we can summarise as follows:
Website → GTM client → Universal Analytics → Report
Today, with GA4, server-side integration and BigQuery, the range of options increases considerably. These tools support more complex, customised and scalable data flows, opening up new possibilities for advanced analysis and campaign optimisation. Some of these possibilities are outlined below.
Efficiency to taste
website → GTM client → GA4 → Looker Studio
This solution is suitable for small brands or companies with limited budgets. Data is tracked through client-side GTM, sent to GA4 and then displayed in Looker Studio.
There are, of course, some limitations, including:
- Ad blockers that can prevent tracking;
- limited analysis;
- reports that are often slow and inaccurate, particularly when measurements are complex.
Budget-friendly Data Analytics
website → GTM client → GTM server → GA4 → Looker Studio
A good balance between cost and performance. By adding server-side tracking (via Google Tag Manager Server), we improve reliability and accuracy. Tags are managed by the server, reducing the risk of data loss.
Although this results in more complete and secure data that is less affected by ad blockers, the solution still has some limitations. Reports in Looker Studio can still be slow and inflexible when handling a large number of events.
Data analysis at scale
website → GTM client → GTM server → BigQuery
As the volume of data to be measured grows, so does the need for more advanced tools. In this setup, although the data passes through GA4, it is automatically exported to BigQuery, where it can be manipulated, enriched and analysed flexibly.
Some of the available options include:
a. BigQuery → Databricks → Power BI
A typical example of an enterprise workflow, ideal for organisations already working with a Microsoft stack or requiring complex processing.
b. BigQuery → Ad hoc tables → Looker Studio
By creating optimised tables and custom queries in BigQuery, we can then feed Looker Studio with accurate, fast and ready-to-use data. This setup resolves performance issues and provides genuinely tailored reporting.
c. BigQuery → tailored KPIs
In many cases, BigQuery is used specifically to calculate tailored KPI aligned with business objectives (e.g. abandonment rate by device type, revenue by user cluster, etc.).
These queries can work effectively alongside Looker Studio, Power BI, Tableau or other visualisation tools, ensuring scalability, customisation and report automation, and are probably the most effective solution currently available.
A data tracking system that has not been implemented in line with best practices does not generate visible errors or trigger alerts, but it silently undermines every analysis, every dashboard and every strategic decision.
If you are not certain that your data collection system is configured according to principles consistent with quality standards, it is time to take action. There is no need to start again from scratch: we simply need to build a robust tracking infrastructure, which is monitorable and aligned with your business objectives.
Because today, more than ever, data is not simply what we measure. It is what we can turn into action.
