Otto GmbH & Co KG
On track to becoming a platform
OTTO is one of Europe’s most successful e-commerce companies and Germany’s biggest online retailer for Furniture and Home Furnishing products. Now it is on track to becoming a platform.
OTTO generates well over 90 percent of its combined revenue online. 2.8 million article positions and 6,800 brands from the assortment areas of Fashion, Living, and Multimedia make OTTO the go-to provider for almost any wish. In the past financial year 6.6 million customers ordered online from OTTO. At peak periods OTTO.de receives up to ten new orders a second. Like virtually every other retailer with an online presence, OTTO competes against Amazon for customers and mindshare. To hold their own, they knew they needed to use and analyze data to know their customers better and to sharpen their pricing strategy for bidding on onsite ads.
Struggling with islands of data
Their existing IT architecture, however, made that difficult because it consisted of a large Teradata database and many islands of data. Some used SQL, some PL/SQL, some were Oracle-based, and some used PERL scripts. The result was that OTTO had to do a lot of data transformation to, for example, determine what kind of ad would be the most effective in enticing a customer to make a purchase.
“We decided to standardize on a single integration tool”, says Michael van Ryswyk, Product Owner BRAIN Realtime, OTTO. After analyzing their requirements and evaluating possible solutions, OTTO selected Talend for a number of reasons, including its open source nature, pricing model, and ability to support the Continuous Integration/ Continuous Delivery (CI/CD) process of software development that OTTO also wanted to implement.
“When we decided to use Talend, we also decided to build a completely new data lake on our Cloudera cluster”, says van Ryswyk. OTTO saw Talend acting as the bridge between the new data lake — which stores data from the OTTO website and from partner websites — and the company’s 250TB Exasol database. “We also knew that one of the advantages of using a solution like Talend would be simplified software maintenance. Like many other companies, we had a lot of integration scripts we could no longer maintain because the original developers were no longer here. So Talend is ideal for standardizing integration when you’re building a new data lake from scratch.”
“Also, Talend gave us the option to use our own CI/CD process, which none of the competing solutions did,” says Rouven Homann, Management Partner at cimt ag, OTTO’s IT consultant. “With CI/CD, we can continuously integrate data sources using Talend and continuously develop and deliver new, small upgrades, updates, and applications,” says van Ryswyk. “That means we can deliver requested functionality, make changes and correct issues much faster, and just overall reduce the risks associated with developing new software and putting it into production.”
Homann adds that the CI/CD strategy has also improved the quality of the software released. “We’ve established best practices for all Talend jobs,” he says. “We’ve created test environments, processes, and everything a developer needs to put new software into production much faster.”
Sharpening pricing strategy and stock
Developing software using the agile methodology and Talend in a CI/CD environment makes it possible to respond to requests from business units in hours instead of weeks. “We have 90 developers working on different teams and different projects, like developing 360-degree customer profiles, or determining which customer is connected to which mobile device and which pc or gaining visibility into what products a customer is currently searching for on our site,” explains van Ryswyk.
In addition, OTTO is using machine learning that relies on the integrated customer data and product descriptions and photos to make suggestions to customers. “For example, if we don’t carry the exact item the customer is searching for or it’s out of stock, the system can identify an item that’s close to it and present that to the customer,” says van Ryswyk.
OTTO is also analyzing referral paths and click paths from ad partner websites to see where customers are coming from so it can show more relevant ads to promote purchases. Complying with new GDPR requirements is another perfect example, says van Ryswyk. “For instance, a business unit may tell us that a customer wants his or her data to be removed from the OTTO system. In the past, that would have been a very time consuming request to fulfill and would have required searching several databases. But now, Talend can communicate quickly with the data sources and simplify the process of deleting the data from Exasol and the Cloudera data lake.”
[[ This case study was originally published at Talend.com ]]
Would you like to learn more?