Data-Driven Marketing: Our partnership with Brico

Recently, we helped Brico embrace a data-driven culture. Brico needed better insights into its customer base composition and evolution, while becoming more data-driven to help its customers with personalised marketing to propose the products that the customer needs at that specific moment.

This effort covered several solutions, spanning the components of intelligent data-driven organisations (data foundations, information management, business intelligence, and data science).

Our data strategists, who master all components of data-driven organisations, have guided Brico in:

Analytical translation, to make sure that all business needs and priorities were covered in the analytical solutions (analytical base table, models, and dashboards)

Data project management over different data domains.

Preparing a high-level overview of potential analytical solutions on Brico’s path to fully personalised marketing.

Eventually, we integrated several data solutions into one environment:

• Analytical base table describing customers in hundreds of parameters to describe purchase & engagement history.
• Predictive models (artificial Intelligence) to target customers in need of specific products with personalised marketing.
• Deployment & scheduling to integrate all of these solutions and deliver them at the frequency that Brico needs them to realise additional value.

Below, we shortly present you the different data solutions individually.

Analytical base table

Python Predictions supported Brico in creating an analytical base table (sometimes called feature store) to unlock hundreds of features that help Brico with better a understanding of their customers and that data scientists can use for artificial intelligence.

We brought several sources together into one table to describe customers by creating all the necessary data pipelines, with transparent and well-documented definitions of all parameters. This table is refreshed daily to allow marketers to immediately use insights to tailor campaigns to customers and it is stored historically for data scientists to use Artificial Intelligence. This frees up to 80% of the data scientist’s time to focus on what matters, preparing algorithms to personalise all communication with customers!

Predicting customer interest

Python Predictions has helped Brico in using artificial intelligence to personalize marketing for its customers.

We have created two explainable analytical models to better understand the retailer’s customers. These analytical models are tailored to (1) be understandable and fully transparent and (2) include parameters that capture the human expertise that has been built over the years. These analytical models are scored automatically at the frequency that is required for marketing campaigns.

The targeted customers were so triggered by the communication that was sent to them that the open rate of the e-mails increased to a whopping 75%.

Model deployment & scheduling

Brico wanted to have structural insights into its data and customer base and wanted to leverage analytical models for marketing campaigns.

Python Predictions developed a solution to structurally:

• Ingest, process, and serve data for their analytical base table, used for reporting & analytics.
• Score analytical models to target the right customers.
• Prepare required views for PowerBI dashboards.
• Automate all data pipelines.

While making sure that the solution was cost-effective and tailored to the needs of Brico’s marketing and CRM teams. The solution was created to allow rapid redeployment when data sources were changed or when the analytical models were updated with new data.