Will Zalando stay ahead of Amazon in European online fashion?
Europe’s largest online fashion retailer, Zalando, started in 2008 as a European copycat of Amazon and Zappos.com. It’s market leader position is under pressure – Can machine learning enable the company to stay ahead?
Zalando: Imitating Amazon in Europe
Europe’s largest online fashion retailer, Zalando, started in 2008 as a European copycat of Amazon and Zappos.com.  Competitors considered the European market challenging because of its large variety in languages, regulations and tastes.  This offered a white space opportunity for Zalando’s founders. Their strategy was successful: During its IPO in 2014 Zalando was valued at $6.8 billion.  The company has seen continued growth: In 2017 it was active in 17 European countries where it made €4.5 billion in Net Sales. 
Machine Learning as competitive edge in e-commerce
With over 200 million site visits per month Zalando possesses an incredible amount of data.  Additionally, they have partnered with some brands and bricks-and-mortar retailers to get access to their stock counts.  The company is relying on machine learning to convert data to insights on all fronts, ranging from forecasting demand and targeting advertisements to optimizing their logistical processes.
One example of machine learning at Zalando is the self-teaching algorithm that they presented last month. The algorithm can identify items of clothing and then put these together into outfits. As training data, more than 200,000 outfits created by stylists were used, primarily sourced from Zalando’s curated shopping service Zalon. 
Another problem that the company is trying to address with machine learning are returns: Currently half of all sales are returned by customers because of problems with fit or style. 
The company is also looking at machine learning for more radical improvements. Reiner Kraft, VP of engineering for search and personalisation, said: “We can use AI to revolutionise this [shopping] process so that years from now, you’re buying fashion in a completely different way.” 
Zalando’s approach to Machine Learning
To reap the benefits of machine learning Zalando’s management is placing a large emphasis on tech. CEO Rubin Ritter explained that Zalando is really running three companies: A fashion, logistics, and tech company.  Their dedication is demonstrated by the scientists and researchers they hire in Machine Learning & AI fields.
In 2015 Zalando opened its Fashion Insights Centre in Dublin that would create 300 jobs in three years. Other tech hubs are located in Germany and Helsinki. Most recently, the company announced that they are adding another hub in Lisbon.  The objective of the hubs is to “re-imagine fashion for the good of all and drive Zalando-scale innovation through academic research and cross-team collaboration.” 
In true software developer style, the Zalando Research department has embraced open source by posting on development platform GitHub. For example, they published a dataset of Zalando’s article images named the “Fashion MNIST”, that could be used by the GitHub community to benchmark their machine learning algorithms. 
Zalando is also collaborating with other companies to drive innovation. In 2016, they partnered with Google Zoo and digital studio Stinkdigital to create “a design engine that incorporates the style preferences of more than 600 fashion influencers.” 
Staying ahead in European online fashion
Zalando’s position is being threatened as Amazon plans to expand in fashion and as Alibaba is expanding in Europe.  Mr. Ritter argues that Zalando can co-exist with these e-commerce giants because they target a different customer segments. “Amazon is pursuing the more price-conscious shopper, whereas Zalando is after a higher-value, more brand-conscious segment.”  He believes that customers prefer a specialized play when shopping for fashion. 
However, this argument falls short. The US has many online fashion specialists such as Yoox Net-a-Porter SpA, Farfetch and Lyst, but they haven’t stopped Amazon from being market leader in the category. 
To distinguish itself Zalando will have to offer a superior customer experience. The radical improvements to the shopping experience that their research department is working on could give the company that competitive edge. To achieve this Zalando will have to rely on the Machine Learning capabilities and expertise of their employees. Having the right talent is crucial and therefore open source projects such as Fashion MNIST are important, as they help promote Zalando as a potential employer in the data scientist community.
The question remains: Will this effort be enough to stay ahead of Amazon in European online fashion?
 “Zalando: Fashion Fight.” Financial Times, May 09, 2018. http://search.proquest.com.ezp-prod1.hul.harvard.edu/docview/2052549070.
 Mac, Ryan. “Revenge of the Ripoff: How Zalando Became A $5 Billion Retailing Sensation.” Forbes (Aug 18, 2014): 1. http://search.proquest.com.ezp-prod1.hul.harvard.edu/docview/1553176155.
 Mac, Ryan. “Zalando Worth $6.8 Billion After Pricing IPO At More Than $27 A Share.” Forbes (Sep 29, 2014). https://www.forbes.com/sites/ryanmac/2014/09/29/zalando-worth-6-8-billion-after-pricing-ipo-at-more-than-27-a-share/.
 Zalando, “Company profile.” https://corporate.zalando.com/en/company/zalando-at-a-glance, accessed November 2018
 Anonymous, “Fashion Forward; Zalando.” The Economist, Sep 03, 2016, 67-59, http://search.proquest.com.ezp-prod1.hul.harvard.edu/docview/1816231426.
 “A Fashion-Savvy Algorithm,” press release October 30, 2018, on Zalando website, https://corporate.zalando.com/en/newsroom/en/stories/fashion-savvy-algorithm, accessed November 2018.
 Hounslea, Tara. “Zalando and Google Trial Style Tool.” Drapers, Sep 02, 2016, https://www.drapersonline.com/news/zalando-and-google-trial-machine-learning-enabled-fashion-design/7011718.article.
 Bremmer, David, “Zalando-ceo Rubin Ritter: ‘Nederland is de ideale proeftuin.’” Algemeen Dagblad, September 18, 2018, https://www.ad.nl/economie/zalando-ceo-rubin-ritter-nederland-is-de-ideale-proeftuin~a2be3b39/, accessed November 2018.
 “Zalando Reveals New Tech Hub Location in Lisbon,” press release June 22, 2018, on Zalando website, https://corporate.zalando.com/en/newsroom/en/press-releases/zalando-reveals-new-tech-hub-location-lisbon, accessed November 2018.
 Zalando, “Zalando Research Jobs.” https://research.zalando.com/welcome/jobs/, accessed November 2018
 Grigorev, Alexey, Rajalingappaa Shanmugamani, Alberto Boschetti, Luca Massaron, and Abhishek Thakur. TensorFlow Deep Learning Projects : 10 Real-World Projects on Computer Vision, Machine Translation, Chatbots, and Reinforcement Learning. Birmingham: Packt Publishing, 2018.
 Neumann, Jeannette, Laura Stevens. “Amazon Is Taking Over the World? Not in Europe.” The Wall Street Journal (April 30, 2018). https://www.wsj.com/articles/in-europe-amazon-com-remains-out-of-fashion-1525080604.
Student comments on Will Zalando stay ahead of Amazon in European online fashion?
I agree that Zalando can improve its capabilities from leveraging more data and machine learning. However, the fact that Yoox Net-a-Porter SpA, Farfetch and Lyst are still growing well (Yoox was growing 20% last year) showcase that the industry is shifting and this might not be a zero sum game
While I do think that Zalando offers a different value proposition to its customers compared to Amazon, other players pose more of an immediate threat, as you point out.
I think using Machine Learning to establish how to combine clothing articles into an outfit, and how to best determine fit are not only great for Zalando as they will increase cross- selling and reduce costs (returns), respectively, but are also a much needed technological advancement in online shopping. I wonder if Zalando can expand into offering accessories or other compatible products that would “go with” an outfit. I also wonder how quickly Zalando can update the outfits they generate in order to keep up with the ever-changing world of fashion