As a leading quality fashion discounter, Takko has ambitious growth targets. In addition to its existing 2,000 stores across Europe, a further 300 locations are set to open over the next three years, primarily in the core markets of Germany, Austria and the Netherlands. At the same time, the company needs to continuously increase sales and profits in a competitive industry environment. Against this backdrop, the efficient handling of the extensive company data available is essential.

The Takko branch in Tönisvorst. Over the next three years, 300 new locations are set to be added to the existing 2,000. (Takko Fashion)

“Our challenge is to generate as much knowledge as possible about our customers, locations and products from our data, in order to achieve our business goals,” says Dr Dirk Mannweiler, Takko’s Manager of Data Science. However, the previous data landscape severely limited this endeavour. ‘Until recently, our data analysis was based on a classic Microsoft DWH in our own data centre. We built our reports with Reporting Services and performed analyses using Excel. Machine learning automation and the use of artificial intelligence were out of the question.’

For these reasons, those responsible decided to make a fresh start. A central data and AI platform based on cloud technologies will form the foundation. Further steps will see the connection of a modern CRM system for customer analysis with Salesforce. The leading industry tool, RELEX, is planned for supply chain management.

Simplifying data handling

Measured by company size, Takko has a relatively small BI team. Therefore, ease of use was a top priority when selecting the platform technology. “We need a solution that is easy for us to use and administer,” explains Dirk Mannweiler. With this in mind, the Takko team evaluated various providers. It quickly became clear that the solution would be a Databricks Lakehouse, in which all company data would be centralised.

“Whether it’s reading and processing data, developing AI models, automating processes, or connecting new systems, Databricks is perfect for our project to harmonise Takko’s data and data science worlds as effectively as possible, thanks to its intuitive functionality.”

The new Data & AI platform is designed to enable employees at all levels of the company to use data more effectively to achieve business objectives. (Takko Fashion)

The next step was to find the right partner to structure and implement existing ideas and plans. This is where the data and AI experts at ORAYLIS came in. “We talked to various service providers. The ORAYLIS consultants understood us right from the first meeting. We also got on very well with them right away on a personal level.”

Data strategy assessment as a project entry point

The project began with a data strategy assessment. The joint workshops helped the Takko team to structure their requirements and goals.

“In principle, we knew where we wanted to go,” explains Dirk Mannweiler. ‘But there were still a lot of unanswered questions, such as: What can we realistically achieve? Have we prioritised our issues correctly? What technologies do we need exactly? What does a concrete architecture look like? What are the next steps? All of these questions were answered to our complete satisfaction during the workshops. Afterwards, we had a clear idea of how we could realise our Databricks vision with the Microsoft Azure Cloud services. At the same time, it became apparent that the planned technology stack would optimise many of our use cases, such as our planned Salesforce CRM.

At the end of the assessment, those responsible received a decision document. This clearly summarised the results and outlined the next steps. It also confirmed that ORAYLIS would be responsible for the upcoming implementation of the solution.

Enable reports on your own initiative

Another important finding from the workshop series was that the new platform would enable Takko’s specialist departments to work with data independently, thereby increasing their efficiency. To this end, ORAYLIS introduced a managed self-service approach to the project. Sophia Pommerin, Senior Project Manager at ORAYLIS, explains the details: “Employees can create reports and applications independently within defined parameters. In addition, the specialist teams are responsible for maintaining and providing specific data pools. Meanwhile, infrastructure, strategy and governance issues remain the responsibility of central BI.’

Sophia Pommerin, Senior Project Manager at ORAYLIS. (ORAYLIS).

The Databricks Lakehouse is supplemented by Microsoft Fabric to provide employees with the necessary tools. This innovative software includes all the components required for building modern data and AI platforms. Takko uses Power BI exclusively as its company-wide analysis tool. “This approach saves Takko a lot in licence costs,” says Sophia Pommerin. “The project is based on Fabric F64 capacity. This means that any number of employees can use Power BI to analyse data for their work.’

Integration of Salesforce and RELEX

In parallel with platform development, the project team began migrating the old CRM environment to the Salesforce system. Databricks proved its worth here as an automated data hub that consolidates CRM data from various sources centrally, processes it, and feeds it directly into Salesforce. In return, the CRM tool provides live daily data on customers and sales. Combined with master data and geodata, this enables more comprehensive reporting insights.

The marketing and sales departments in particular benefit from this. “We are now working with AI-supported segments and clusters, which give us a much better understanding of the customer journey,” says Dirk Mannweiler. “As a result, we can address our customers in a more targeted and personalised manner with bespoke offers. This enables us to provide them with a shopping experience that is tailored to their needs.” The Data & AI platform also provides valuable support in connecting to RELEX. Existing ERP data is linked locally in Databricks with product master data, which is then made immediately available to the new planning system.

AI-supported software solutions optimise both customer engagement and the product range in stores. (Takko Fashion)

The Planning & Allocation department can use RELEX’s detailed AI forecasts and recommendations for optimal merchandise allocation. These insights are then fed back into the platform and made available as corporate knowledge. This contributes to expansion goals by enabling the company to advance its own geo-analyses quickly and effectively.

Data and AI in harmony

Application scenarios demonstrate how data and artificial intelligence are becoming increasingly inseparable at Takko, forming an integral part of its business activities. Dirk Mannweiler is delighted with the sustained acceleration of the innovation process: ‘Our AI projects can now be implemented much faster and with higher quality. We can process and analyse huge amounts of data in real time, and our solution scales automatically. This means that questions from our specialist department, such as those relating to out-of-stock scenarios, can be answered using AI in the shortest possible time.

Accordingly, the agenda for the near future is already packed. The team wants to expand the CRM area and further refine customer segmentation. Customer communication is to be gradually automated. At the same time, product topics need to be explored in greater depth, and expansion plans require increasingly detailed information. Last but not least, the team has set itself the goal of tapping into the potential of generative AI through the use of Databricks Genie, with the aim of supporting work processes increasingly with AI assistants and agents.

The decision to continue the partnership with ORAYLIS has already been made: “From data strategy assessment to platform development, everything worked really well and made perfect sense. We have realised that we need this expert support in many areas and look forward to continuing our collaboration.”

Turn your data into value.