The SWE algorithm is based on a feature-based clustering mechanism. These features include:
Social network parameters that describe a person;
Store parameters that describe the product;
Offer parameters that describe the discount applied
Since the objective is to optimize a discount, maintaining a balance between best sale margin (minimum discount) and biggest sale volume, a brief analysis is done in order to define the system's performance when facing a random strategy and an optimal one (best margin sale).
The SWE mobile app was developed with the physical store panorama in mind. It took advantage of QR codes and the customer's social network data in order to provide a personalized discount in store. All the customer had to do was to download the app, install, and follow the steps. It was developed in HTML5 and CSS3, in order to overcome possible cross-platform issues that could appear in iOS, Windows or Android platforms. During development final stages, the app was subjected to a simple usability test, in which both users registered no usability errors.
The SWE web app allows the store manager to check his store's recent activity, viewing who bought what, always knowing the discount applied in that case. He can also check particular stores or products performances and compare them, based on their conversion rates. Adding to this feature, there is an Analytics page, which enables cross-store and cross-product comparisons, powerful information to attain in order to optimize product prices, which can also be controlled in the web application.
In order to ease product's implementation on the market, guaranteeing wider availability for digital stores' managers, two plugins were developed. Two major E-commerce web platforms were chosen, both Magento and Woo-Commerce. These integrations simplified and helped the widget installation for most store managers. Most importantly, both plugins rely on the same (Open) API, which proves that it can be used to expand the application to other platforms.