The exponential growth of data, digitization and internet connectivity is the “backbone” of the Fourth Industrial Revolution, the digital transformation. Previously I covered this concept and how it will enable innovative business models disrupt traditional ones. We are already seeing digital transformation shake up big box retailer, Wal-Mart. Amazon changed the game. Much like retail, there are new patterns and technologies to embrace for digital scale analytics. In this article, we will explore how adidas modernized analytics platforms for the digital age.

Beginning of a revolution that is fundamentally changing
the way we live, work and relate
– Prof. Schwab, World Economic Forum


Most organizations begin with digitizing the customer journey and then extend changes to operations, products and services. Savings from automation and retiring old technology paves the way for more modernization. adidas also followed this same path. Between 2010 and 2016, adidas grew their online sales from zero to over €1 billion euros. Today they enjoy over 86 million active subscriptions, more than 2, 30 million total customers and send out approximately 123 million emails per month to 150 countries.

adidas customer journey

adidas analytics technology was designed to understand what motivates consumers and drives decision-making. Graham Sharpe, Director of Reporting & Marketing Intelligence at adidas Global IT: Digital Brand Commerce, says the role of Digital IT at adidas is to accelerate building direct relationships with customers. This falls nicely within the adidas “Creating The New” IT strategy which highlights the adidas desire to become a data driven organization through “Engineering the Digital Tomorrow”.  To facilitate this drive, data scientists at adidas created a Consumer DNA (CDNA) model consisting of re-useable, pre-fabricated analytics components to create a 360 degree Consumer View.


Analytics Digital Scale cdna_numbers


Although sales is still an important KPI, adidas also measures influence as a new currency in a digital world to create brand awareness, interest and desire with consumers. CDNA data points continuously expand to include a wide variety of external sources of information. Data science algorithms called genes then iterate through billions of records every single day continuously updating and adjusting profiles based on the latest available consumer data.

Sharpe goes on to explain that three key digital transformation questions influenced technology modernization decisions to power the CDNA.

  1. What do our digital consumers expect?
    Personalization, understand me!
  2. What are the key IT success factors enabling change?
    Agility, speed, independence and data.
  3. Do traditional IT processes and technologies match to the needs of the consumers?
    Let‘s see…

Enabling automation and efficient machine learning processes for CDNA data are critical success factors. Reviewing how much time data scientists typically spent on daily tasks was not at all surprising. As much as 80% of data scientists time was consumed by organizing, cleaning and collecting data sets. Only 20% of the time was available for high value work. Thus, one of the first technology goals was to reduce that 80% of time doing low value work.


data science tasks


All analytics technologies were evaluated by adidas team to ensure they were Actionable, Performant, Flexible and Future Proof. Looking back at adidas analytics architecture and processes from 2013 when they had 20 million customers, it was neither actionable nor scalable despite the use of market leading solutions at that time.


Start Analytics Digital Scale


Fast forward to 2016 when adidas grew to 80 million customers and notice several key analytics technology additions of Exasol, Tableau, Alteryx and Pentaho. These solutions met adidas Actionable, Performant and Flexible criteria. Exasol’s low cost, high performance massively parallel processing (MPP) database included an extendable data analytics framework to power CDNA. Alteryx reduced and automated time consuming data preparation tasks. Tableau enabled the business to get fast insights without requiring IT when and where they needed it.


2016 Analytics Digital Scale


Although adidas analytics architecture performed better than it did in 2013, it would not elegantly or inexpensively scale for a world with massive volumes of data and unlimited digital consumers.

To future-proof analytics as adidas expanded to profiling over 100 million customers, several changes were made.

2017 Analytics Digital Scale

adidas decision to embrace young, smaller, innovative analytics vendors and open source technology reflects trends seen in the worldwide analytics market. Big vendors are being disrupted in the digital transformation despite claims of bigger is better when it comes to scale in the cloud. Anyone can scale in cloud. The barriers to market entry have fallen.

“Old Guard” have continued to tumble
– Doug Atkinson, Solutions Review

Looking beyond 2018, adidas shared that they expect “build versus buy” technology decisions to become increasingly more difficult due to escalating complexity in a digital world. Data security is critical and legislation is continually evolving. Analytics platform refactoring and continuous modernization seems to be never ending and the new normal.

Sharpe shared one final tip for peers in the analytics community that are just beginning the digital transformation journey. He says as you evaluate solution architecture, “rely on cost-effective analytics standards such Exasol for massively parallel relational database needs together with new capabilities to create simple, reliable, cheap (results per €), open and performant solutions”.