Luxury fashion house gets deeper insights into product line sales performance

A global luxury fashion house needed to replace a 100-hour per month manual process of creating sales reports from multiple internal and external sources, so they turned to Pythian to integrate, transform and unify multisource sales performance data.


Critical Issues

The company deals with billions of dollars in sales from a range of channels: their own bricks and mortar retail stores, their online stores and third-party retailers. They needed to cut through this complexity to get a clearer picture of omnichannel sales performance across widespread regions to make more informed decisions about what to sell, where to sell and to whom, and how to promote their products. The company’s marketing department was experiencing the paradox of having vast volumes of data at their fingertips but getting few reliable insights from it. They had all the data they needed to see the complete picture of sales performance across regions and sales channels, but lacked the ability bring it all together to get the timely insights they needed to improve on all aspects of their marketing efforts - from developing products to meeting market demands, to promoting their products to reach the right buyers at the right time. The problem was that their data was in multiple systems and databases, and didn’t follow a common format or reside on the same platform. Further, their existing manual extraction process and Excel-based analysis created two problems: first, it consumed an enormous amount of time and effort to extract insights, and second these insights were unreliable as they were subject to human error. Often they were full of anomalies, incomplete records and duplication. By cleaning, deduplicating and integrating their data in an efficient way, they would gain the insight they needed for a better understanding of which product lines were selling well, where they were being sold, and through which channels. This would help them focus their marketing efforts so they could reach the right customers with the right products at the right time. To accomplish this, the client knew they needed to find a better way to integrate their complex data sets than to continue with their time consuming manual process of creating reports in Excel. They knew that Google Cloud Platform (GCP) could provide a cost-effective means of integrating data efficiently. They also knew that GCP would allow them to scale on demand, collaborate more effectively around data and focus on getting insights rather than managing infrastructure. And they understood the power of Google BigQuery to provide the insights and integration with Tableau, a tool they had recently decided to adopt for better visualizations. The client understood that they couldn’t do this alone. They didn’t have in-house expertise in big data integration or in GCP. They needed to partner with a company with deep GCP expertise and experience integrating complex, multisource data from a range of on-premises databases and external systems. Pythian was the logical choice.

We Provided

The client chose Pythian and an enterprise data and analytics platform built on GCP that incorporates both services and automation to streamline processes and workflows. The Pythian team created a set of pipelines to effectively integrate data on a single platform while streamlining once time-consuming manual processes like cleaning, deduplication and data format unification. Pythian was able to help the client integrate a range of data sources, including sales transaction data residing in an Oracle database, an inventory management system and Google Analytics. Pipelines were created to export data from each source system, transform it, transfer it to Google Cloud Storage, then load into Google BigQuery. An automated mechanism was also developed to do a daily poll from Oracle for sales transaction updates. Within BigQuery Pythian created use case-specific data marts. Pythian’s analytics team created a set of more than 50 standard reports in Tableau that would automatically draw from the data marts in BigQuery any time new data was loaded into the pipelines.


Today the client has the integrated data they need to enable their users to innovate, define and quickly test new business hypotheses. Each time new data becomes available in the source systems the client can initiate a process whereby the pipelines detect an update. This triggers the ingestion of data into the enterprise data platform. Data is then automatically transformed in the pipelines and the data marts are updated. A simple refresh by the client in Tableau updates the data within the automated reports. The company’s marketing team can explore the reports to discover how product lines are performing across their various retail properties, and across their partner channels. A hundred-hour-per-month manual process of creating sales report spreadsheets now takes just minutes with GCP and Tableau.


  • Google Cloud Platform
  • Google BigQuery
  • Google AppEngine, Google Cloud Storage, Google Datastore, Google Compute Engine, Virtual Private Cloud, Cloud IAM, Google Stackdriver
  • Apache NiFi
  • Tableau