// you’re reading...

Complex Insight

Evolution of shop visualization and behavioral modeling.

The retail industry is often at the cutting edge of technology adoption – since  increases in performance due to differentiation or performance improvements have  transparent competitive advantages and near immediate measurable bottom line impact . Way back in the 1995, I worked with a research team at Salford University and a small VR start up – then called Intelligent Systems – which eventually morphed into Virtalis. One of the projects led by then founder Prof Bob Stone (now at Birmingham Uni) and Andy Connel (Technical Director at Virtalis) was the development of a virtual superstore for Sainsbury’s. The original intent of the store visualization was to test alternate layouts and store configurations. Much has happened in the intervening 15+ years in relation to store layout and contextual product visualization. Data mining of sales data provides  mountains of behavioral trend data and insight in to customer behavior. Companies such as TNS Sorensen have refined understanding shopper motivation and behavior at the point of purchase into a distinct field of research called “shopper insight.” The new tools help  marketers develop new store formats, category management plans and promotional strategies that closely align with what today’s shoppers behavior indicates what they truly want.

TNS Sorenson’s software products PathTracker and Atlas give shop designers, architects and experience designers a different tool set to create shopping experiences. PathTracker is their proprietary electronic tracking system that captures the in-store movement and behavior of shoppers. Data analysis can using PathTracker can indicate:

  • How fast shoppers walk through the store.
  • How long shoppers ponder a single purchase.
  • How fast their purchases add up.
  • The mix of categories/products in their basket.
  • The geographic path through a store that shoppers visit  (front, back, etc.)

Atlas® uses the PathTracker data from actual  stores to help  calculate the impact of moving products within a store on shoppers behaviour. This online tool enables designers to simulate an infinite variety of category configurations and immediately see the change in weekly sales and profit margins.

With these types of tools the purchasing decisions of shoppers can be optimized and innate behaviour explicitly drawn upon to maximise sales potential. The tools themselves provide a combination of behavioral insight generated through data mining of shopping behavioral data and visual simulation of new outcomes based on inferred rule sets.

Given the emergence of deep analytical tools at the cutting edge of retail- extrapolating the application of these  tools for computational insight  to other industries not only makes sense but defines a set of  emergent market opportunities. As sensor data becomes increasingly online the ability to track data (RFID, feed and other sensors) combined with historical data mining/data modeling of this data will provide trend analysis information to manufacturers, operators and many others.  This is part of the data input feed into big data analysis using Hadoop and other tools.  Its clear that  similar to  TNS Sorenson’ss PathTracker and Atlas will be needed to make sense of sensor data patterns and trends over time.

Applying a similar approach to learning means we can begin to build models of educational efficacy, patterns of interaction and learning bottlenecks.

Discussion

Comments are disallowed for this post.

Comments are closed.