Wednesday, August 21, 2019

Machine Learning and Image Recognition in Retail

image recognition in retail merchandising
References to machine learning seem to be everywhere these days. From new university-level courses to magazine articles and social media posts, daily we encounter new information on artificial intelligence and machine learning. But how can those new technologies be applied to improve retail operations?

One of the main ways that machine learning can help retail is when it comes to quantifying data that previously came as individual pieces of information while surveying stores, such as images.
Image recognition in retail is one of the best examples of the use of machine learning in merchandising applications. 

Machine learning allows images to be fed by programmers into a mobile merchandising system in order to train a neural network to identify product faces on outlet shelves.

Then, field staff can simply take pictures of the store shelves on their mobile merchandising apps and the system will be able to identify - with accuracy that reaches 95% - the number of faces of the target product, as well as compare it to competitor brands and automatically calculate share of shelf.

Merchandising managers and other stakeholders can, in real-time, access the results of store audits, filtering the quantified data and generating reports and visualizations according to their needs on KPIs such as out-of-stocks, share of shelf, product availability, competitor availability, and promotions, among others, and make faster mission-critical decisions.

BrandML is the image recognition technology developed by VisitBasis, creator of the namesake mobile merchandising software chosen by many market leaders in the CPG and product distribution industries.

Want to know more about machine learning and image recognition in retail? Contact www.visitbasis.com and ask about the new BrandML technology.


Wednesday, August 14, 2019

Share of Shelf with BrandML


image recognition for retail merchandising
Calculating share of shelf for retail goods used to be a tedious and error-prone process: It involved merchandisers counting the number of target product faces, then the number of competitor product faces, for each and every store where the product was being sold.

Machine learning has recently introduced image recognition technology in retail and is now transforming how merchandising audits are performed.

BrandML is the image recognition technology developed by VisitBasis, creator of the mobile merchandising software chosen by many market leaders in the CPG and product distribution industries.

BrandML allows merchandisers to calculate product share of shelf simply by taking a picture of the category shelving at each store.

The neural network that supports BrandML is first "trained" with pictures supplied by the client. This allows the system, with around 95% accuracy, to recognize the target products in the pictures taken by merchandisers and automatically generate share of shelf and other KPI reports and visualizations, including out-of-stocks, product availability, competitor availability, promotions, among others.

VisitBasis combined with BrandML becomes an extremely powerful tool for merchandising, sales, and channel management, as VisitBasis already provides GPS-stamped visit scheduling and planning and also allows building custom forms for in-store activities.

Interested in using image recognition to automatically calculate share of shelf for your product lines in the different outlets? Contact www.visitbasis.com and ask about our new BrandML technology.