Bunting offers many algorithms which can be applied to Product Recommendations, with each having certain levels of suitability for different pages of your eCommerce website.
The following list details the algorithms and the uses for them:
- Adaptive – These are the default setting for recommendations from Bunting using a machine learning algorithm to detect what the best recommendations are given in the context of the page.
- Recently Viewed by Visitors
- Most Popular – Last 24 Hours
- Most Popular – Last 7 Days
- Most Popular – Last 30 Days
- Popular Alternatives – This is most often used on Product Pages. It takes into consideration what customers view, add to cart and buy after viewing the specified product.
- Cross Selling – Commonly shown on both Cart/Checkout pages and Product pages. This takes into consideration what other customers buy from these pages.
- Repeat Custom – This will determine what customers usually buy, based on what their last purchase was. The context for this is the order complete page, this is done cumulatively.
All of these serve a solid basis for the personalization of the recommendations which are provided and allow for extensive customization.
How does the Most Popular algorithm work?
Within the recommendation engine, each product is attributed a numerical weighting, this is corresponding to an array of factors. Many factors are taken into consideration here, however, the main ones to take into consideration when selecting this algorithm is: the number of times that the product has been ordered, viewed the click-through rate and the time frames of each factor. Together all of these factors are passed into the algorithm to calculate what is essentially a rating. A simple way of defining this rating is the likelihood of the product appearing in the recommendations.