top of page

From GA4 to BigQuery to improve product search experience

“Human37 helped us centralise our customer data, and laid the groundwork for future projects. With this first collaboration, we collected actionable and valuable customer insights.”

Amaury Gilliot, CEO Noukie’s


Over the last 25 years Noukie's has been developing a complete range of products for children aged from 0 to 8. Noukie's is an omnichannel brand that grows and progresses every day with children and their parents. Its products are designed for the harmony, comfort and well-being of our young explorers!

As a true omnichannel retailer, Noukie’s operates through three sales channels: their eCommerce website, their own brick and mortar stores (in Belgium) and their network of 1,000 resellers worldwide.


Noukie’s has always been a client-first brand. They are convinced of the importance of customer experience. In order to take its customer experience to the next level, Noukie’s took the decision to define and implement a strong customer data strategy and the underlying infrastructure to support it. The overall new data strategy needs to be designed in a way that it would address the following challenges:

  1. Noukie’s customer and product data is scattered across more than 5 systems, which are not entirely connected

  2. The customer behaviour is not tracked in a customer-centric event based approach, which makes behavioural analysis complicated

  3. Noukies is migrating some systems, including the ecommerce, and needs a system-agnostic approach on data, in order to be flexible to future technological changes

Using an iterative approach to build a sound architecture, it has been decided that the initial project to set the foundations would be a part of product analytics for which they keep facing unanswered questions: the filtering of item lists. More specifically, Noukie’s is trying answer the following questions:

  • Are my product filters being used?

  • Which filters are the most used?

  • Are filters used simultaneously?

  • Are the people who are using the filters more likely to convert?

  • Which filters are used with specific product categories


The Collection

Human37’s first step consisted in validating that the data required to answer the above listed questions was being collected. Through a series of audits and discussions, Human37 could conclude that some website behavioural data was missing. In order to complete the data collection, Human37 decided to implement a complete Google Analytics 4 (GA4) instance and have it directly connected and dumped into Google BigQuery through the native connector.

The Ingestion

Within Google BigQuery, two additional data sources got ingested and unified: their data product feed data through the Google Merchant Centre and their transactional & customer data through Noukie’s CRM.

The Unification

Once the automated ingestion data pipeline was up and running, several transformation pipelines were created in order to (1) unify users (known and unknown) identities, (2) create consistent and user-centric data models and (3) account for human errors and typos in the data. These pipelines were created with SQL models in DBT.


In order to move from cleaned, user-centric data to insights, we took the following actions:

  • Construction of the analysis - compilation of data

  • Visualisation

  • Recommendations

  • KPI baselining to automate continuous follow-up

Overall, the filter analysis brought clear understanding of which filters are used, which are business critical, useless or nice to have. All those learnings are now gonna be used to improve the new website Noukie’s is developing and will be releasing in the coming months.

In the medium term, post website release, it helps Noukies to source data to further improve the filtering unit. The data allows design, UX and CRO teams to build and validate a hypothesis through A/B testing before committing it to a permanent change to a product release.

WHAT's next?

Using the infrastructure that got created, Noukies can leverage the existing base in order to build and expand upon the following elements:

  • Ingestion - this corresponds to the ingestion of new data sources in the data warehouse Here are examples: CRM, Mailing System, Database, ERP, Customer support solution, etc.

  • Activation - this corresponds to pushing data from the data warehouse to communication destinations (Mailing system, Website, Media platforms, etc.)

  • Analysis and/or reportings - this corresponds to the analysis and/or reportings to help Noukie’s transform data into insights and recommendations, visualise it and democratise access to those insights for the entire team.

Interested in supercharging your own analytics stack? Want to drive more insights from your data? Don't hesitate to reach out to us at

Recent Posts

See All


bottom of page