4 Reasons Why We Developed Our Own E-Commerce Platform

Fabian Wesner
Fabian Wesner CTO Spryker
10. February 2014 in


In this first post of our new tech blog we will explain why we did not use an existing shop system to set up our portfolio companies, but instead decided to develop our own e-commerce software platform called “Yves & Zed”.

Building market-leading e-commerce companies is one of the areas of expertise provided by Project A. By “e-commerce” we mean conventional online shops such as for example our ventures Tirendo and nu3, curated shopping communities like Epicerie, marketplaces such as Saatchi Art, and subscription models like Kochzauber.

We know that there is already quite a number of well-established and sophisticated shop frameworks out there such as Magento, Hybris or OXID eSales as well as SAAS solutions like Demandware, but after looking at all of them (and working with some in previous companies), we decided that there is nothing readily available that really suits our needs. Here’s why:

Launch Quickly, Scale Big Later

In order to test the market and to have something to show to investors and partners, we typically launch a minimum viable product (MVP) within a few weeks after the start of a new company. This usually means that we put a new design on top of some lightweight frontend code and then put together a backend from our pool of standard features (for instance a CMS or components for connecting payment providers like Payone, Adyen, Stripe or Paypal). Of course we wouldn’t need our own shop system to quickly launch a MVP, but we think that it is much harder to grow companies towards more significant user numbers using a standard e-commerce solution due to scaling issues.

First, it is well known that page loading speed has a huge impact on conversion rate. With Yves & Zed it is always a high priority for us to keep response times well below 100 milliseconds, which we think is difficult using generic shop systems. For example, with Magento we often experienced that due to some architectural shortcomings it becomes unbearably slow once a certain number of products, visitors or orders is exceeded.

Second, most shop systems are not concerned with the processing of orders beyond the “easy” cases of payment and shipment. Once a company starts receiving thousands of orders per day, manual processes for cancellations, returns, and refunds become operational bottlenecks, and usually an ERP system comes into the game, which in our opinion introduces unnecessary complexity, high costs, and missing flexibility. Therefore, we built a solution for scalable and customizable business processes as one of the key components of Yves & Zed.

Third, meaningful numbers about sales, operations and marketing are a requirement for effectively scaling a company. For this, and for building data-driven features such as recommendation engines or CRM systems, consistent and integrated data about products and user behavior is needed. Nevertheless, such systems are typically developed independently from the shop system, often even by completely different parts of the company. In contrast, we see business intelligence as a natural part of IT and therefore integrated an own data warehouse and reporting solution into Yves & Zed.

Fourth, we strongly believe that a startup must own and master its technology and have full control over its source code. The initial lack of features compared to an existing generic solution is easily outweighed by the ability to quickly implement any component that the company needs (we made easy extensibility one of the major design goals for Yves & Zed). Everyone who has ever tried to customize Magento knows how difficult and unpleasant it is to fight against their rather rigid framework. Furthermore, as a company builder we also need a practical way to exchange features among our ventures to retrieve synergy effects.

That was more or less the four main reasons why we chose not to use any of the existing shop systems out there for our ventures. In our next post, we will introduce the main components and software architecture of Yves & Zed.


This post was written in collaboration with Martin Loetzsch.

Still got questions?
Ask the author for further information.