Client testimony

“Astral Forest successfully and efficiently accomplished the following: 

• BI environment assessment (Power BI, Azure, SQL Server, Synapse, Oracle, Databricks).
• Set of recommendations and a roadmap for the environment modernization.
• Refactoring of the BI solution (over 2 billion rows).
• Implementation of a modern data architecture based on the Medalion Architecture approach.
• Creation of a cloud-based Azure data solution.
• Construction of reports in Power BI (500 users).
• Optimization of infrastructure/license costs.

I highly recommend collaborating with Astral Forest.”

Head of Business Intelligence and Big Data

About the Client 

A French automotive parts dealer based in Paris. They offer parts for both passenger and commercial vehicles, as well as workshop area. As part of a larger capital group, they bring together companies involved in trade across Western Europe. With over 60 years in the market, they have managed to establish a network of over 200 stores. 

Background

The company’s long-standing presence in the market and maturity have allowed for the development of not only the organization but also internal competencies. As a result, the client has onboarded a dedicated person for Business Intelligence management (Head of BI), as well as a company that has prepared an Azure environment and initial reporting. 

Challenges

One of the main problems the company faced was the failure to meet specific price-related key performance indicators (KPIs). The client knew that there was a gap somewhere that was impacting profit margins. As a network of wholesale and retail stores, salespeople often manipulated margins, which had a drastic effect on sales results. Our main goal was to gather all the necessary data and identify the bottleneck. Based on the generated reports, we wanted to identify, among other things: 

• Which branches experienced the largest declines? 
• Which salespeople were responsible for low margins? 
• Which product groups had prices undercut? 
• Which manufacturers had the lowest profitability? 

Another challenge directly resulting from this problem was the quality of the available data. Inconsistencies in the reports not only hindered drawing conclusions and taking further steps but also significantly affected trust in the tool and the willingness of employees to work with it. 

Lastly, but not least, the quality of the initial reports provided by previous vendors was another significant issue. Based on those reports, our client couldn’t determine the source of the problem, which is why they reached out to us for assistance.

Solution

After coming on board and learning about all the challenges the company was facing, we decided to start by conducting an audit of the existing solutions and internal situation. Based on the audit, we identified limitations and proposed solutions for them. Initially, we focused on implementing quick wins so that we could later shift our attention to long-term solutions. 

Within a month, we managed to accomplish and implement the following: 

1. Power BI Model Optimization – as a result, we achieved better performance right from the first reports. 
2. New architecture; we switched from Azure SQL DB as a gold layer directly to Databricks – as a result, we achieved monthly savings of 3000 euros. 
3. Migration of an external web application from OVH to Azure resources – as a result, we significantly reduced the likelihood of data breaches. 

After addressing these issues, we could focus on further improving the platform and data environment. Together, we established plans that would help the company advance technically and in terms of business growth. 

From a business perspective, our main focus was on fostering a collaborative relationship with the pricing teams. We aimed to demonstrate the value of properly prepared reports and working with consistent data. By increasing their data analysis awareness, the client could trust their employees and their subsequent decisions. 

From a technical standpoint, we focused on catching up on the backlog accumulated over the years of the company’s operation. We planned the implementation of more pricing reports, which was not possible without addressing the technological debt. Additionally, we worked on fixing the data architecture. We improved the connection between Azure and On-prem infrastructure and streamlined the process of the data exchange, resulting in reducing data transmission time from 6 weeks to a single day. 

Technology

• Azure
• Databricks
• Power BI

Team and Competencies 

• Product Owner 
• Senior Power BI Developer 
• Data Engineer 
• Data Architect 

Results and Future Plans

Thanks to our collaboration and the knowledge we were able to impart, we have achieved all the objectives identified in the audit and implemented solutions that directly impact pricing results. Our key successes include: 

1. Creation of a Data Lake structure: We have established a scalable and efficient Data Lake environment that enables the collection, storage, and processing of data in a flexible manner.
2. Implementation of a medallion architecture: We have implemented a medallion architecture, which allows for the decoupling of different parts of the system and provides scalability and better control over the data.
3. Streamlining the time to implement new ideas: Through process optimization and infrastructure improvements, we have reduced the time required to implement new ideas and reports, enabling faster response to changing business needs. 

In the near future, we plan to fully migrate our client to the Azure environment and start the migration from MicroStrategy to Power BI.