Fashion retailers are increasingly turning to artificial intelligence (AI) for help in the quest to give customers what they want. AI-driven retail enables brands to compete in the economy of the 21st century and meet modern customer demands by personalizing the shopping experience.
As more retail companies are taking their business from the traditional brick and mortar retail stores into e-commerce, they are able to get more insight into their customers’ preferences to serve the demand.
Talking about AI in retail brings us to one of the most beloved brands in the fashion industry – H&M.
H&M Group and their AI-driven story
H&M Group has been heavily investing in artificial intelligence to stay on top of fashion cycles, and also to support its massive growth.
From their beginning in 1947, the fashion retailer followed rapid expansion across Europe, U.S. and globally. And as early as 1998 they began to sell online.
Just in 2004, H&M held more than 1,000 stores, a number which grew exponentially to nearly 5,000 stores in 72 countries worldwide. And as of today, the retail giant employs 177,000 people and contributes to 1.6 million jobs for people employed by their suppliers.
But how does a brand manager sustain this kind of growth, reaching across different countries, markets and even continents? How do they make sure they meet the needs of such a huge base of customers with different shopping needs and expectations? And most importantly, how do they manage to run thousands of stores worldwide?
The man who has answers for all these questions is Errol Koolmeister, H&M Group’s Product Area Lead Engineer AI Foundation at H&M Group. During his talk at the Data Innovation Summit 2019, he explained that they first started adopting AI into their business in 2016. The impact of digitalization was clearly visible and H&M Group knew they had to do something in order to stay relevant.“We, as a company, and as people, started going more and more online”, describes Errol.
“We started to see that scaling a business model of physical stores became quite hard”, Errol says. The obvious solution was that they needed AI in order to sustain their growth.
The advent of AI-driven retail
Realizing they had to adjust to the new industrial shift, H&M Group asked themselves the question “Where do we want to be as a company?” That’s when the retail brand realised they needed to up their game in AI and advanced analytics. Although they excelled at some departments, such as CRM, they knew they didn’t have it at scale.
That was the moment when they started doing their first POC’s. “We tried to see if we could extract any value from them,” explains Errol. Luckily, the use cases that H&M Group picked were successful and could be put into production, making them profitable.
“It’s all about acting really rapidly,” he continues. “If we in 2016 only said let’s do a proof of concept and then stopped, no value would have been extracted”, Errol says. This is unfortunately what happens with 90% of other cases, they just do their POC and don’t continue with production.
What proved to be successful for H&M Group, was that they were focused on production already in 2016. They were determined that if they were to start with AI, they would do it big. “It’s really about starting small, thinking big, and scaling fast”, highlights Errol. According to him, in reality, there are no AI use cases, they are all business cases. H&M Group actually uses AI to amplify their business solutions.
H&M Group’s use cases have different maturity levels. But regardless of the case, they always focus on production. For this reason, they have a defined process for realizing use cases:
- Proof of Concept (POC) – which doesn’t take long to realise because, as Errol explains, they already have their data lake in place, ready to go. This allows them to spin up virtual machines and get insights in a matter of a few weeks. If the POC is successful, it’s time for the next stage.
- Pilot – H&M Group pilots the solution in a few markets to test if the theory matches reality. They put the case into production to try it out and evaluate the results. If the pilot is successful, they roll out the solution.
- Industrialization and roll-out – the last step of the process is to roll out the solution to all 72 countries, 177,000 employees and nearly 5,000 stores. Considering the number of users, the solutions have to be extremely bulletproof and reliable.
Real-life H&M Group examples of AI-driven retail
Here are some tangible examples thanks to implementing AI solutions.
Keeping popular items stocked – H&M relies on staying on top of trends in order to be successful. With the help of algorithms, they analyse store receipts and returns to evaluate purchases in each store. This way, the fashion brand knows which items to promote and stock more of in certain locations.
Predicting market demand – Fashion retailers like H&M rely on fresh products at competitive prices. Data insights help H&M to predict what the market wants so they don’t have to discount their inventory to sell it out.
Automated warehouses – Today customers expect fast, hassle-free deliveries anytime and everywhere. Therefore, H&M Group has invested in automated warehouses that will ultimately offer next-day deliveries for the majority of the European markets. The warehouses and their free shipping, exclusive for loyal customers, are driven by algorithms and data.
Personalised offline customer experience – H&M has introduced its personalised online recommendations also in their physical stores, with the help of RFID technology. Customers get in-store merchandise suggestions selected by algorithms. They can also see if an item they have seen online is available in a physical store, and scan labels to see if an item is available in another store or online.
Tailor-made clothing – Partnering with an AI-technology platform, the Swedish fashion brand has tested on-demand production, which shows great potential to react more specifically to customer’s wishes, and to align product quantity to local demand.
What’s the takeaway from H&M Group’s story?
In today’s digitalised world, it is impossible to manage a global retail brand such as H&M solely by humans. And where human capabilities end, AI begins. Taking H&M Group’s example of AI-driven retail, retailers can gain much from enhancing human intelligence by implementing AI into their business: more accurate merchandise decisions, streamlined supply chain and top-level customer experience. This is what H&M Group calls Amplified Intelligence – the collaboration between machines and humans, science and art, data and gut feeling.