Case M.Video: Success with Behavioral Clustering for Built-in Appliances
M.Video is the largest Russian consumer electronics retail chain by revenue, providing a successful service to more than 379 stores, operating in 161 cities in Russia. M.Video offers a unique store concept, providing customers with shopper friendly stores, offering the latest electronics in trend at truly great value.
In 2010, as M.Video approached an estate exceeding 200 stores, the management quickly realized a key opportunity; to group stores together according to similarities in customer behavior in order to improve the management of its assortment and merchandising strategy.
This would enable M.Video to take the next step in its evolution; moving away from an approach which utilized averages across the store’s estate to eliminate significant issues. Such issues included having an incorrect assortment and inventory balance in each store, along with incorrect interpretations of a store’s customer demand – leading to the store manager driving inaccurate replenishment.
Ultimately, by grouping similar stores together, M.Video can provide a customer offer tailored to meet demand, driving an improved stock mix with optimized inventory to improve sales and profitability.
M.Video at a glance
- Stores 379
- Locations across 161 cities in Russia
- Revenue $4.3 billion
- Number of transactions 27.9 million
- Company mission “To be the best place where people and consumer electronics meet”
In 2011, whilst working with consultants at Ernst & Young, M.Video selected RELEX, an integrated retail and supply chain planning solution provider, to support store clustering in line with customer behavior and to drive insights for optimized assortment and space. With an existing project for assortment, M.Video set-out to select a partner who could support the foundation for an end-to-end solution.
“RELEX’s solution supports us to better understand our customers, meaning we can tailor our product offering to improve their in-store experience.”
– Egor Bakharev, Head of Competence Centre of Category Management, M.Video
A rapid roll-out
After a successful pilot with top-selling destination category, ‘televisions’ – in which M.Video saw significant improvements with a sales increase in excess of 11% and stock turnover improvements of 12% – the team were keen to accelerate the roll-out of RELEX’s behavioral clustering functionality.
M.Video quickly identified the best performing 20 categories – which make up over 40% of its overall category offering – that could benefit from a more tailored approach. Having the ability to conduct detailed cluster analysis in just a few days meant that M.Video could very quickly create store groupings for each category based on actual customer buying behavior. Plus, for very fast moving categories such as ‘televisions’, ‘washing machines’ and ‘cooling’, it was now possible to refresh store cluster analysis at scheduled times throughout the year, to ensure that assortment and space remained continuously in line with M.Video’s ever-evolving customer demand.
In addition to this, M.Video identified that RELEX’s behavioral clustering capability offered a key opportunity not only with the analysis of shopper behavior across the physical bricks and mortar estate, but also to drive analysis across internet sales for a category.
By understanding customer buying variation across attributes such as price and brand, M.Video quickly realized significant differences in shopper patterns between physical stores vs. internet sales. This allowed M.Video to understand similarities in shopper demand between entering physical stores vs. those purchasing products online or utilizing the click & collect service. With product delivery sourced from the most local store, to better balance assortment and inventory in line with demand.
Seeing the benefit with behavioral clustering
A top 10 seller for M.Video, ‘built-in appliances’ comprises a variety of major categories including ‘ovens’, ‘panels’, ‘extractors’, ‘dishwashers’ and ‘refrigerators’. Recognizing a key opportunity to unlock further benefits and improve customer experience, M.Video took top selling category ‘ovens and panels’, and carried out a more formalized category management review. This helped to identify store space and demand issues, with clearly defined targets:
1. Devise a standardized space allocation for category space across the store estate
2. Understand customer demand for gas vs. electric ovens and panels, and define the demand for different colors of ovens and panels
“We’ve seen significant benefits with this latest review. RELEX’s solution has allowed us to quickly move from a generic approach focused on group averages, to one which is tailored in line with our customers’ needs.”
– Egor Bakharev, Head of Competence Centre of Category Management, M.Video
Devising a more standard space allocation
Historically, assortment and shelf space allocation had been completed at a group level, however, in the absence of a formal category review process for select categories, the exact space at each store remained unknown, and after calculations were carried out, M.Video identified approximately 200 space combinations possible across the estate.
Furthermore, the assortment was analyzed based on group averages, and as such, remained untailored to customers who shopped within each store. This led to inconsistency in the assortment offer and a lack of understanding of customer preferences at head office and therefore store level. Although store replenishment was possible, this lack of clarity meant that store managers would often misinterpret the needs of their store e.g. replenishing the store based on an assumed value led shopper, when in reality, the customer requirement was more premium driven. In addition, this led to issues with chosen assortment and replenishment options e.g. selling ‘retro’ style oven and panel products in a single store but not of the same color, which impacted the shopper experience as they could not find the two products they wanted to purchase in the same color.
As a first step to category optimization, M.Video decided to standardize the space allocation for this category across stores by using the following approach:
- Understanding the current space allocated across the store estate
- Analyzing the unique sales for the category; looking at the number of unique SKUs sold during the month on average across both ‘ovens’ and ‘panels’
- Analyzing the efficiency of space bands by looking at the percentage of unique sales against the current space allocation
With effective and standardized space bands identified, M.Video was quickly able to define the optimal space allocation for each store and was able to apply store level efficiency calculations.
This initial space operation enabled M.Video to get a fast balance between operational and sales efficiency within allocated space bands. This also minimized the variation of space across the store estate to make it possible to generate a more targeted assortment mix per space band.
Understanding customer demand
With space identified, a key next step in the category review process was to define the level of customer demand across different product types; ‘gas’ vs. ‘electric’ and different product colors for ‘ovens’ and ‘panels’.
Utilizing RELEX’s behavioral clustering functionality to analyze customer purchasing patterns across attributes of type and color by sales revenue, M.Video were very quickly able to identify clear variation in customer demand across the store estate for ovens and panels.
Five clear store groupings were quickly generated, in the table below, these have been analyzed against product type:
Cluster analysis identified a distinct group of 47 stores that showed an average mix between gas vs. electric ovens and panels, as seen in column ‘Cluster 50/50’. Other clusters highlighted trends such as ‘Cluster Gas’ with 68 stores with an indicative trend toward gas and ‘Cluster More Gas’, where customers showed a much stronger preference for gas ovens and panels.
Working together with category managers, the M.Video team were able to utilize these key insights to drive the best space break allocation for these categories per store in line with sub-category product types e.g. an even split for gas vs. electric in ‘Cluster 50-50’ vs. a higher proportion of space for electric products across the 80 stores in ‘Cluster more electronic’.
With sub-category space allocated in line with the trends identified for product type, M.Video were able to further optimize assortment with cluster analysis by product color; identifying the percentage share of revenue for each product color across each of the 5 clusters. The resultant analysis identified key customer preferences which M.Video was able to utilize to drive more targeted assortments, with an optimized product mix to meet customer demand in each store grouping and improving key identified issues at store, such as ‘ovens’ and panels replenished with incorrect color which is now matched to customer preference.
improvement in inventory turns
total saving across the store estate for the category
The allocation of new space bands along with the use of behavioral clustering to support the category management review for ‘ovens and panels’ meant that M.Video were able to accurately allocate stores to the right cluster in line with demand, unlocking a number of significant benefits.
- Improved understanding of customer behavior across ‘ovens and panels’
- Improved inventory turns of up to 12% in some stores, giving a total saving of over $260,000 across the store estate for the category
- Improved assortment mix in line with identified customer demand
- Improved logistics efficiency with more accurate replenishment
- Improved customer experience at store
Related case studies
- Case Kolonial.no: Winning at Fresh Food Etail
- Case Musti Group: Thriving in a Complex, Multinational, Supply Chain Environment
- Case Akademibokhandeln: Centralizing Processes for Better Christmas Management
- Case Minimani: Hypermarket Chain Cuts Food Waste Through Supply Chain Optimization
- Case Derome: Ensuring a Successful Implementation and Change Management During Rapid Growth