Automated quality assurance for complex analytics environments

In complex Google Analytics environments - whether GA Standard or GA 360 and regardless of whether Google Tag Manager (GTM) or Tealium iQ (TiQ) is used - it becomes increasingly difficult to ensure the quality of the data and therefore also the quality of the decisions that are made on the basis of the reports. If you then plan to use machine learning (ML) algorithms, you need to ensure excellent data quality, otherwise your models will not be successful and any effort will be wasted.

If all events are correct - i.e. events are sent everywhere on the website where expected - you must also ensure that this remains the case afterwards. Unfortunately, a website is "alive" and with every extension or change there can be shifts or losses in tracking. Testing everything manually with every change is not possible or would involve excessive effort. Every digital analyst knows this problem only too well.

We have tackled this problem and developed several tools. We have automated quality assurance as far as possible and massively reduced the manual effort. How do you go about it?

Parameter reference: the heart of our data quality

We create a so-called parameter reference from the tracking requirements, in which all events are defined. This definition includes the mandatory, optional or prohibited attributes, their type and so on for each event.

We automatically generate the following tools and documents from this single parameter reference.

Trackbook: The guide for developers

An error-free specification for data layer developers is essential. This avoids duplication of work and writing different documents for different target groups. This technical documentation is generated automatically from the parameter reference.

The debugger for quality assurance

Once the developers have implemented the data layer in accordance with the trackbook, it must be tested. This is done using a specialised debugger that draws its validation rules directly from the parameter reference. By simply browsing a website or using an app, it is easy to check whether all events have been implemented correctly.

Documentation for the overview

The tag management system is the centrepiece of the implementation in a tracking strategy. Whether GTM or Tealium iQ - our documentation tool reads the configuration of the TMS and generates an overview of all objects in the TMS, from load rules and extensions/variables to tags, badges and audiences. This documentation allows us to check critical settings in a targeted manner, scrutinise the current status and ultimately optimise the entire configuration, which eliminates potential errors, reduces overall complexity and increases transparency.

The event inventory: Continuous testing of target vs. actual

As it is not possible to check all events or complete regression tests by hand, this tool analyses the collected data to check whether all collected events correspond to the parameter reference. This is done at regular intervals and completely automatically in order to recognise deviations promptly and thus ensure flawless data quality.

Active monitoring

Even if all events are collected absolutely correctly, this is no guarantee that other circumstances will not suddenly lead to a drop in tracking. This is where we come in with various monitoring strategies in order to automatically and actively check a large number of tracking implementations for fluctuations or anomalies.

Would you like to find out more about automated quality assurance or about zweipunkt? Visit our website and contact us without obligation. We look forward to hearing from you.