IRIS Home // Case Study // Zeeman
Zeeman: Moving from intuition to data-driven expansion
The challenge
For years, data played a minor role in determining new store locations at the textile chain Zeeman. “It was done the Vicky the Viking way, mainly based on gut feeling. Open a little shop and move on to the next,” says Micha Candel, Manager Real Estate at Zeeman. Since 2019, those days are over: Zeeman now uses data and smart algorithms to determine the best locations.
Zeeman’s expansion
Since its founding in 1967, Zeeman has evolved into more than just a textile chain. Its range includes baby and children’s clothing, household textiles, underwear, and non-textile products such as groceries and cleaning items. These products are offered in physical stores and the webshop.
The strategic focus on quality basics and textiles has been fruitful; Zeeman welcomes 80 million customers annually across 8 countries, generating a turnover of 931 million euros.
Zeeman also takes its social responsibility seriously, being careful with people, the environment, and society. This is reflected in the use of sustainable materials (>75%), selling second-hand clothing, and a top-10 ranking in the Fashion Transparency Index. In 2022, Zeeman won the SBI ‘Initiative of the Year’ award for numerous sustainable initiatives (e.g., the ‘Frugal campaign’).
For this leading player in the European textile market, the era of intuitively opening stores is over. There really isn’t an alternative, says Candel. “We have 1,300 branches in eight countries, most of them outside the Netherlands, with only a headquarters in the Netherlands. You can no longer rely on your gut feeling; you have to rely on data.”
The solution: data-driven location strategy
In 2019, Zeeman started their collaboration with Kyden (formerly known as Kirkman Company) and The Big Data Company to apply smart predictive algorithms. These algorithms are ‘trained’ based on large amounts of data about potential customers in the catchment areas, the characteristics of the store surroundings, and information about the turnovers of current stores.
“We look at factors such as the purchasing power index and the composition of households in the store’s catchment area,” says Sander Kloppenburg, Data Experience Leader at Kyden. “But also other relevant matters, such as the presence of bus stops, a school, or another ‘point of interest’. Morevover, we use data that shows how busy a certain place is at a specific point in time. We combine the data and use it to train the algorithm.”
These algorithms are then made available on the Interactive Retail Intelligence Scout (IRIS) platform. Users can click on a spot on the map to virtually open a store, and within minutes, the estimated turnover pops up, plus an estimate of the impact on existing Zeeman branches in the vicinity. Supporting visualizations also clarify how the revenue prediction is constructed. Daan Kolkman, Partner at The Big Data Company, explains:
“We aim for an accuracy of 90% for all our clients. Together with the Property & Expansion Managers of Zeeman, we discussed which factors they find important when identifying a new location. We compare their input to the feature importances identified by machine learning models to create a model that is accurate, widely supported, and transparent.”
The main sales driver for most retailers is the number of people in the catchment area, but this often accounts for only 40% of the accuracy. The remaining 50% is determined by the characteristics of the population within the catchment area, properties of the store, and the store’s environment.
The result: better decisions
The system provides “a critical mirror for the gut feeling.” Preventing an early and costly closure of a new branch is crucial, says Candel: “We are now a phase further. We have seen what works and what doesn’t. By doing, we learn.” According to Candel, the chain wants to continue expanding: “We want to maintain and, where possible, accelerate the growth we have experienced.”
In recent years, Zeeman has calculated hundreds of scenarios with the IRIS platform, including openings, closures, changes, and relocations. In some countries, the textile chain is expanding rapidly, while in others, it aims to optimize the existing store portfolio.
Not only the algorithm needs to be trained, but Zeeman’s team itself is also developing. The interplay between predictions from the IRIS platform, internal budgets, and gut feeling leads to better decisions. Candel:
“The team from Kyden and The Big Data Company brings data applications to the ‘normal’ world with refreshing expertise in the field of transformation.”
Daan Kolkman
The Big Data Company
Inspired by Zeeman’s success with data-driven decision-making? Discover how your company can also grow with the help of data and smart algorithms.
_
