The Challenge

Originally launched as an online store for Apple computers, Cyberport SE is now one of the most successful German e-commerce companies and is wholly owned by Hubert Burda Media. The mail order company for computer technology and digital consumer electronics currently operates nine stores in Germany and Austria.

Cyberport SE is able to hold its own in a competitive market, which is characterized above all by high margin pressure and increasingly extensive product ranges, with targeted investments in product data management. It is essential for Cyberport to be able to offer the largest possible number of products from the various product categories in the best e-commerce quality at competitive prices. To ensure that this challenge can always be met, Cyberport decided to introduce a new, more powerful PIM system.

How Onedot Helped

As a central location for all product information, a PIM system has become indispensable in e-commerce, as all product information is collected and managed in this system and can be enriched accordingly. The organically grown and specially developed Cyberparts PIM had to be replaced. For this reason, it was decided that the new PIM system should be based on standard software. Alongside the introduction of the new PIM system, the opportunity was also taken to adapt and revise the existing product data model, mainly to meet the requirements of modern e-commerce. To ensure that the new product data model could be correctly mapped on the new PIM system, Cyberport opted for the option of running the migration continuously and iteration by iteration via the Onedot product data platform.

With the introduction of the new PIM system, strategically important adjustments were also made to the product data model, as the product data model contains all structurally relevant information such as category trees, value lists and attributes. The main task for Onedot was to transfer the product data from the Cyberparts PIM to the target data structure of the new PIM system. Unlike a difficult and error-prone two-stage IT migration project, the continuous migration does not work step-by-step, but over many small iterations.

This means that in a first step, the new product data model is mapped in the Onedot product data platform. The existing product master is also imported onto the platform. This data can then be used to carry out the first automated product data onboarding. The data from this first migration is loaded into a test system after processing by the Onedot AI, which makes the mapping suggestions from the existing product data model to the new model. There, the mapping proposals are checked and evaluated by Category Management.

At Cyberport, the results of these checks led to an improvement in product data quality. This also allowed the product data model to be refined. The advantage of continuous migration via the Onedot product data platform is that daily iterations are carried out automatically and feedback is mapped within 24 hours. Cyberport was thus able to migrate the product master of over 40,000 SKUs daily and check it for completeness and data quality. The continuous migration is completed when the data quality in the test system is sufficiently high and a final migration to the production system can be carried out.

The Result

Onedot's product data platform enables Cyberport to map the existing product data model to a new target system quickly, agilely and iteratively and to make adjustments with just a few clicks. This is primarily possible because the intuitive Onedot user interface is designed to meet the requirements of category management and other PIM stakeholders and is easy to use.

This iterative introduction of the new product data model has major advantages: Improved filters, comparison options and the more precise display of technical attributes of the products online enable a significantly improved shopping experience. It also takes the pressure off the product data team, which does not have to make adjustments and changes manually in the PIM system, but can accompany and check iterations in a targeted manner.