Evolving retail analytics
- Description analytics
- Diagnostic analytics
- Predictive analytics
- Prescriptive analytics
- Autonomous analytics
Description analytics
descriptive statistics of how much each item in the assortment sold in the past period
Diagnostic analytics
a model that relates sales to marketing-mix activity (price, promo, shelf space) to explain past sales patterns
Predictive analytics
an attribute-based version of the previous model that allows one to predict the sales for new line extensions based on the calibrated model
Prescriptive analytics
the previous model + some optimization routine that searches for the best product line extension to add to the assortment
Autonomous analytics
automated algorithm used in online setting that continuously updates itself and presents each visitor with an optimized assortment
What is the effect of customer assistance on sales?
Diagnostic analytics
How can we best predict product returns?
Predictive analytics
How can we optimize an experiential store flyer?
Prescriptive analytics
Decision support system components
- Category sales model
- Very large neighborhood search heuristic
Category sales model
- Estimating the category sales response function
- Quantifying the effect of the amount of space in a store flyer on its own and cross-category space
Very Large Neighborhood Search Heuristic
- Optimizing the composition of a flyer
- Simultaneously deciding on category selection and space allocation
- Heuristic that finds a neighbor of a flyer and when an improvement moves there -> in parallel on multiple computing cores from different starting solutions. Find the optimal flyer out of 450 Mio possible ones in less than 15 secondes
New retail characteristics
- Digital supply chains
- Digital payments
- Platforms that collect lots of data
- In-store apps
- Super apps
- …
Priorities of retailers: seven themes
- Advancing analytics to build more resilient supply chains
- More localizing of decisions at the store level
- Improving measurement of in-store customer behavior via new technologies
- Reinforcing omnichannel orientation through enhanced inventory-sharing and more efficient omnichannel distribution and delivery
- Extending data access within the company across the value chain
- Using more external and contextual data
- Connecting different decision areas such as inventory optimization for product promotions, or demand forecasting with anticipatory returns
New technology focus:
- Get better at tracking assets such as inventory and shopping carts
- Gain a better understanding of what happens in-store in terms of traffic and customer behavior
- Cloud-based storage and computing to provide (near) real-time recommendations for mobile users
