Filtering should be able to be done at any stage of the search as customers use the categories. When they do this, they’re giving input in two different ways. This type of search isn’t only helping them, but also giving the machine learning the feedback loop it needs to help subsequent customers find what they’re looking for. Common filters include cost, sizes, features, and manufacturers, but the list can be configured any way the business needs it. For example, a memory card might be filtered by speed, power consumption, cost, and connection pins. Clothing can be size, color, or style.
Filters can also be used to match specific parts to a particular product. Vehicles are a good example. Customers can start off with the filters option and check the make and model of a car, and then tell the filter what type of part they’re looking for. This last filter ensures that not every available part for the vehicle will show up in the results if the customer was looking for tires specifically.
Filters are an excellent example of why product descriptions are such an important part of every item an eCommerce business sells. If the product descriptions and the filters are aligned, then many products will be lost when the filter has no way to find a product.
Here’s another aspect of filtering that’s often overlooked: popularity. There is a definite hierarchy for the categories that helps customers find what they’re looking for. If machine learning finds that most customers for a clothing site want to see brands before colors, brands should definitely come first. After all, if the filter number of buttons is up first, customers may just throw up their hands and leave the site.