facebooklinkedinrsstwitterBlogAsset 1PRDatasheetDatasheetAsset 1DownloadForumGuideLinkWebinarPRPresentationRoad MapVideofacebooklinkedinrsstwitterBlogAsset 1PRDatasheetDatasheetAsset 1DownloadForumGuideLinkWebinarPRPresentationRoad MapVideo
Customer Story: Retail

At Kiabi, data-driven insights hold the key to profitable business growth

Kiabi needed real-time insights into the performance of marketing programs, but its legacy data management platform could not deliver the insights that the fashion firm’s high-volume global retail environment required.
Kiabi turned to Actian Avalanche for its analytics solution. Avalanche unlocks performance features in existing hardware to deliver extreme processing performance with little or no database tuning.
Kiabi sales and marketing teams can make data-driven decisions based on both real-time and historical analytics. Actian Avalanche enabled Kiabi to analyze terabytes of sales data from more than 20 million customers, with query response times as much as 200x faster than on the old Oracle database.
About Kiabi
In the fashion world, the name Kiabi is synonymous with affordable ready-to-wear apparel for the entire family. The global retailer opened its first store in France more than 30 years ago. Today it operates in more than 15 countries and brings in more than 1.5€ billion in annual revenues. More than 20 million people visit the company’s 500 outlets and its online store to shop for quality clothing and accessories at sensible prices.

Retail clothiers are not driven by fashion trends. They’re driven by data trends. The more effectively an organization can access and act on current, accurate data—a particular challenge in the fast-moving world of fashion—the more successfully that organization can compete.

Nowhere in the fashion world has that challenge been met more effectively than at Kiabi, the global retailer of ready-to-wear clothing.

Data is the New Black
“At Kiabi, we succeed by offering merchandise with the greatest appeal to the greatest number of customers at the right price,” says Jean-Francois Rompais, Head of IT Architecture at Kiabi. “It’s a high-volume business, but without timely and accurate information about the impact of our marketing programs, how can we make our best business decisions? And the higher the sales volume, the bigger this challenge becomes.”

Rompais knows what he’s talking about. After many years of relying on what had originally been a state-of-the-art relational database from Oracle, Kiabi’s sales and marketing teams found that they no longer had access to the insights they needed to make the decisions that would keep Kiabi on the cutting edge of retail fashion. They needed access to both real-time and historical analytics, but the existing database lacked the structure, flexibility, or scalability required to support these needs on the massive scale at which Kiabi was operating.

“We wanted to run predictive analytics on our markdowns so we could improve our marketing campaigns proactively,” says Rompais. “But there simply wasn’t enough information in our data warehouse to know which marketing offers were driving the best results behind our markdowns. For example, if we sold a 10€ item for 5€, what was the source of the markdown? It might have included 2€ for a customer loyalty program, 2€ for a special promotion, and 1€ for a voucher. Or, it might have resulted from an entirely different mix of marketing promotions. We simply didn’t know.”

Yesterday’s Fashion Reconsidered
One of Kiabi’s technology partners invited Rompais to attend a presentation showcasing Actian technology, though, and what he saw convinced him to take a much closer look at both Actian and the assumptions that Kiabi had held for so many years. Actian Avalanche, powered by Actian Vector analytic database, appeared able to achieve extraordinary levels of processing performance—even using off-the-shelf commodity hardware. It also appeared capable of operating with huge datasets—whether static or streaming—while returning query results within seconds rather than hours.

Perhaps the Oracle database that Kiabi had relied on for so many years was like a suit cut for a younger Kiabi; perhaps it simply no longer fit the company that Kiabi had become.

Rompais engaged Actian to develop a proof of concept that would explore the potential of Actian Avalanche at Kiabi. “During our proof of concept, we evaluated database performance using a variety of use cases and database queries. We installed Actian Avalanche on inexpensive, generic hardware—without any manual database tuning—to see what its on-chip caching and smart compression could do. Actian Avalanche performed 2x to 200x faster than Oracle. It performed so well we that did not have to look at competing solutions.

“Actian Avalanche’s columnar parallel processing architecture would enable us to scale to terabyte datasets—both cost-effectively and without sacrificing performance—which is what we needed an analytic database solution to do. To attempt this with Oracle would have been too costly and labor-intensive.”

Figure 1: Actian beats Oracle in side-by-side performance tests

Empowering Business Users
The proof of concept was proof enough, and Kiabi decided to deploy Actian Avalanche in its production environment.

“Phase one of our project focused on pulling data from our outlets for markdown analysis,” says Rompais. “We completed that phase with very positive results and we’ve been able to move into phase two sooner than we had anticipated. That phase involves integrating our web infrastructure and pulling data from our online stores.” Kiabi also plans to extend Actian Avalanche to new use cases, such as measuring sales performance against commission, and measuring customer response across multi-channel marketing programs.

The Actian solution makes it easy for Kiabi to use of its preferred business intelligence (BI) tools, including Business Objects and Tableau Software. Integration between Actian Avalanche and BI tools is transparent to the business users, enabling them to focus on analyzing sales and marketing programs to gain deeper insights and improve decision-making.

“Today, we can handle 800 million records and beyond,” says Rompais. “Our business users are able to analyze all of our markdowns, and with a much finer granularity. They can drill down and gain insights into sales and marketing performance that weren’t available before we adopted Actian Avalanche. Queries execute very quickly, and we’ve reduced data volumes by an order of magnitude.”

Empowering IT As Well
Performance is only part of the benefits Kiabi has realized in deploying the Actian Avalanche Platform. Kiabi and its partners have deep roots in SQL culture, so having full ANSI SQL support has been important. “Over time, we’ve adopted some new Java-based frameworks and tools,” says Rompais, “some of which hide the underlying SQL. But for our developers and IT staff, having knowledge of the database and its structure is important. We appreciate Actian Avalanche because it offers transparency and gives us access. We’re very efficient because we are compliant.”

Rompais notes that Kiabi had been using Actian Avalanche in production for a full 18 months before its IT team needed any support. From day one, the products had been quick and easy to deploy and had “worked like a dream,” which had not always been his experience in prior deployments.

“Our relationship with Actian is more like a partnership. We recently brought in the Actian professional services team to perform a Technical Audit. Our IT staff found them more than willing to explain database particulars and provide useful advice. They’re coming back to train our staff on best practices as we upgrade to the latest version of Actian Avalanche.”

"We installed Actian Avalanche on inexpensive generic hardware—without any manual database tuning—to see what its on-chip caching and smart compression could do. Actian Avalanche performed 2x to 200x faster than Oracle. It performed so well that we did not have to look at competing solutions."
Jean-Francois Rompais, Head of IT Architecture, Kiabi
Ready to Talk to an Expert?