How Lampenwelt used Mara to build a Data Warehouse

How Mara, Project A’s open-source data warehouse framework, helps our portfolio companies optimize marketing planning

By Alexandra Deichsel

Having developed from a small eBay seller to a leading eCommerce platform for lamps and lighting, our portfolio company Lampenwelt is a great success story. Based in Schlitz, a small picturesque town near Fulda, Lampenwelt has grown immensely in the last few years. Founded in 2004 their yearly revenue increased to over 90 Million EUR in 2018. Lampenwelt currently offers a broad selection of approximately 50 000 products from more than 300 manufacturers, serving more than 1.5 million customers located in 15 different countries across Europe.

Project A’s Data Team has been supporting Lampenwelt to scale their Business Intelligence (BI) Infrastructure to get better transparency on their online marketing efficiency.

When we started our collaboration, Lampenwelt had a good and automated solution to report sales and finance data. What they wanted to build, was a single source of truth and by automatically retrieving data from different sources to generate insights on the contribution of different marketing channels to acquire customers and conversions.

Together with their BI Engineering team, we unified Lampenwelt’s marketing campaign structure and successfully applied Project A’s recently open sourced Data Warehouse (DWH) framework “Mara” to integrate data on marketing costs from different marketing channels, web analytics, and transactional data to consolidate and aggregate it in a marketing touchpoint cube.

Mara, Project A’s open sourced Data Warehouse (DWH) framework

Knowing the consecutive chain of marketing touchpoints for each customer enabled us to apply a position based marketing attribution model. Thus, according to their position within the customer journey each marketing channel gets its weighted share on metrics like acquired customers, purchases or revenue, a meaningful improvement compared to commonly used last click attribution models, in which the last channel gets all the credit.

“Project A’ s ample experience with other leading European digital start-ups and “grown-ups” helped us to design and implement our marketing automation and enhance marketing capabilities in a short time frame.”

Janina Herber, Head of BI & Analytics at Lampenwelt

Today, Lampenwelt uses Mara not only for integrating marketing and customer data, which was the initial scope of our project, but also for all other reporting purposes. Mara’s data pipeline infrastructure combines a postgres database with Python scripts to govern all Extract Transform Load (ETL) processes and has built-in documentation features that are very valuable to create transparency.

Using a DWH as a central repository of different data sources with standardized definitions of measures and dimensions, and data cubes that are optimized for querying data is essential for a company to facilitate analysis and generating insights based on data.

In the meantime, Lampenwelt has successfully implemented data cubes for other business questions for example on product performance and logistics.

If you’re interested in working for Project A or for one of our Ventures, make sure to check out our Jobs page!