While adidas group generates an incredible amount of transactional data, selling roughly 600 million pairs of shoes, units of apparel, and sporting equipment in any given year, its tightly woven association with sports goes beyond merely moving products off shelves. The company prides itself on keeping the pulse of competitive and recreational athletes around the world, perceiving how they want to look as well as what motivates them.
That isn’t by accident. As the world’s second-largest sporting goods and apparel manufacturer, with brands that include Reebok and TaylorMade, adidas group devotes significant resources into understanding the pull its products have on the general populace. Capturing data helps adidas group remain a trendsetter on the football (soccer) pitch and beyond, allowing it to produce the right products at the right time and at the right price, year after year.
By design, adidas group structured its technical landscape to maximize this analytical framework. The global organization has, until recently, had four global business warehouse systems in place, including SAP NetWeaver Business Warehouse (SAP NetWeaver BW), to support comprehensive analytics in four corresponding business channels: wholesale, retail, marketing, and planning. These four dedicated systems integrated with roughly 70 back-end databases and reported up through corporate headquarters in Herzogenaurach, Germany.
With SAP NetWeaver BW Accelerator speeding up reporting for these four channel areas, adidas group had a satisfactory structure in place. Its warehouse segment, for example, had the reporting tools it needed to effectively manage more than 44,000 global partners, touching roughly 100,000 points of sale.
But with anticipated growth on the horizon — adidas group had revenues of nearly €15 billion last year and aims to reach €17 billion by 2015 — the organization believed that simplifying its analytics platform would increase its operating margin, thus clearing the path for rapid business growth. Clearly, fewer IT operating costs would lower the total cost of ownership. More importantly, looking ahead to what was available in the IT marketplace, adidas group believed that consolidating its business warehouse and eliminating some latent data redundancies would allow advanced analytics across the enterprise, such as predictive analytics and the modeling of what-if simulations.
According to Miguel Voegele, Vice President of Group Functions for the global IT organization at adidas group, having multiple BW instances, while effective, was proving increasingly inefficient as adidas group cemented growth plans. “Replicating data up to four times created a very complex data model. It was a time-consuming process to react to new business requirements or change requests because IT was the only department aware of the data model and complexity,” Voegele says. “Simplifying complexity will ultimately lead to self-service business intelligence (BI), giving more flexibility to the business users to create reports, dashboards, and analytical tools they need without having to understand the data model in the enterprise data warehouse.”
Getting Off the Bench
As adidas group began to explore its options for architecting its new landscape, and with simplification as the overarching goal, the company quickly discarded some in-memory appliance solutions that required adding multiple components, such as new interfaces and data transfers, to its already crowded analytic environment. Instead, adidas group heavily considered the fact that, as an SAP shop, it already had existing SAP solutions and applications integrated with its SAP NetWeaver BW environment.
This was one of the reasons why adidas group ultimately decided on SAP NetWeaver BW powered by SAP HANA — but did so with one eye downfield or, to use a football analogy, it kept the attack on the dribble while calculating downfield moves. Implementing multiple SAP HANA appliances, each powering a separate BW instance, would certainly accelerate analytics, but adidas group thought it was a short-sighted option. If, later on, the company wished to have an in-memory database power its ERP system and run enterprise-wide solutions, it made sense to align its reporting environment accordingly.
“We envisioned a future of embedding reporting back into the transactional and operational processes of the corporation,” Voegele says. “From that point, based on the SAP roadmap, we decided that’s how we wanted to align our reporting, moving our entire business warehouse environment on a single SAP HANA platform.”
Toward the middle of 2012, adidas group began a six-month proof of concept using real-time data from its BW systems and, after testing assumptions, confirmed that SAP HANA could integrate with its non-SAP systems. In January 2013, the company started the move to the SAP HANA platform, and plans to go live with its wholesale business segment on SAP HANA by the end of 2013, followed by retail, marketing, and planning.
This extended project timeline is due to the magnitude of consolidating its SAP NetWeaver BW and non-SAP systems into one and then transferring it onto the SAP HANA platform. As Voegele points out, at the time the project started in 2012, companies had been using SAP HANA as a database foundation for mission-critical applications, or even had been swapping out their underlying relational databases for SAP HANA to power the latest SAP NetWeaver BW release, but no high-profile companies had yet attempted a similar project. “We wrestled with the question of how to merge two BW systems on one platform,” he says. “Other companies had always had a full back-up with their old solutions in case there were issues. But no one else had done this. It was new for SAP, and it was new for us.”
Hitting the Pavement
Unlike swapping out one underlying relational database, where optimizing an SAP NetWeaver BW environment against SAP HANA makes for a relatively straightforward project, adidas group’s concern was merging its various systems with minimal disruption. The devised approach was to build the SAP HANA platform in parallel with the live SAP NetWeaver BW environment and go live with each business segment piecemeal, with wholesale first.
A central question during the proof of concept was whether a single SAP HANA appliance could accommodate the vast amount of data flowing through the existing SAP NetWeaver BW systems that, combined, allocated for 25 terabytes of data. While the promise of in-memory computing is eliminating the need for allocated disc space, adidas group was justified in feeling uneasy about turning off the old systems. “We need to be careful in our approach. After converting to the new platform, we should be able to safely switch off the old BW system,” Voegele explains. “When we move to retail, it will be coming onto the SAP HANA platform that already has the wholesale business running reports and planning.”
In educating themselves on the intricacies of the SAP HANA platform, the adidas group IT organization wanted to avoid long response times that occurred under the old system. To avoid siloed knowledge on the new platform and foster a self-service BI environment, Voegele says that the project team proactively worked with the business side to share their internal SAP knowledgebase. “Transferring this knowledge to the business would allow us to innovate and do things differently,” he says. “A significant part of our internal BI team worked with hand-picked individuals from the various businesses.”
Points on the Scoreboard
The project is expected to come to full fruition once the planning business segment of adidas group moves onto the SAP HANA platform in 2014. In creating a one-stop reporting shop, users will be able to fire ad hoc queries against the system to answer questions that a previous report may have just generated — delivering the promise of what-if simulations and predictive modeling. “By performing these predictive models, we can look forward to analyzing what’s coming next and make better-informed business decisions,” Voegele says. “The business understands the potential this brings to unlock new business models, and that’s a huge benefit.”
Anticipating that this push toward more of a self-service BI platform wouldn’t happen overnight, the adidas group’s IT organization configured basic what-if situational reporting scenarios to both increase interest and show the inherent value, making sure that end users in the major business segments were on board.
A simplified landscape attracts business-side interest and use while, at the same time, helps to dramatically lower costs, both on the technical side as well as in operations. “IT costs now are a systems running cost, keeping things flowing,” Voegele says. “Also, previously, with the complexity of the landscape and data models, it took quite a while to hire or onboard new people into the team.” Now, Voegele anticipates that reduced complexity will lessen the need and cost of external support that was sometimes required to bring new hires up to speed.
Going into Overtime
Having now laced up its SAP HANA shoes, adidas group is exploring how it can leverage the platform for other areas of its business and looking into whether other SAP applications can share the same SAP HANA platform with its now-consolidated SAP NetWeaver BW environment.
“We will plan a step-by-step, non-disruptive approach of taking every database below our SAP systems away and replacing them with SAP HANA, but we want to ensure that our SAP systems are joining a shared platform,” Voegele says. “This is important to us because of our desire to have real-time changes in customer relationship management (CRM), for example, be automatically reflected in our reporting.”
Conversely, could this real-time CRM data in the business warehouse enable adidas group to capture data with customized reports generated within the business warehouse, without having to execute the reports? “We could then join data schemas from CRM and BW to provide new insights,” Voegele says. “This is something we’ve evaluated and are considering as a next step.”