Understand Fredhopper Suggest Search

This document describes how you can improve the business performance of your eCommerce shop by using Fredhopper Suggest Search. The following scenarios illustrate how Suggest Search converts your users faster and how Suggest Search sells more to each user.

Introduction

Fredhopper Suggest Search supports users in asking the right questions to Fredhopper based on the former behavior of other users. The module provides targeted suggestions as-you-type (Ajax). In particular Fredhopper Suggest enables:

• Marketers to push products and categories;
• Users to spell words, like wii or Gabriel García Márquez;
• Users to express what they want in the catalog language (avoiding very costly editorial changes to the product catalog).

Uplift conversion

Fredhopper Suggest Search helps users find their way through the catalog by highlighting popular shortcuts, thereby reducing the number of steps between customer intention and buy.

Refine searches

The most frequently used keywords are often broad search terms like shoes or toys. In many cases these broad searches lack relevant information even if the user is aware of them, e.g. users will know whether they are looking for women's or men's shoes whether or not they state it. Without the information, Fredhopper is forced to present women as well as men shoes including the facets to enable refining the full result further. Post-search refinements not only force the user to do one more action but also make it harder to present strong promotions on the search results page.

Fredhopper Suggest Search can present the user with common follow up actions like refinements for sub categories or brands. This means the suggest box shows likely facet values that the user may want to refine on further:

Since the space is limited in the suggest box, Fredhopper makes a smart selection of most interesting/most profitable follow up actions. Even if the user may not be interested in the offered option like microsoft, nintento or sony, it gives him a hint as to what type of refinements are possible, e.g. sega console.

Make lower level categories visible

Analyzing the user behavior in a diverse set of eCommerce shops shows that the optimal categorization for search & browse is 2-3 levels deep with a maximum of 7 sub-categories for each category. But reality often looks different. In particular, categorizations that were developed for the mail order world are often 4-7 levels deep and have up to 15 sub-categories. Additionally, it is often not obvious where a sub-category should be placed, e.g. shirts could be part of men as well as of fashion or even offers. The right category may depend on the customer's background and the shop owner's interpretation.

Fredhopper Suggest Search makes lower level categories visible by suggesting them directly. If a user looks for shirts, he sees directly that they are available in men > shirts and women > shirts & tops:

Suggesting lower level categories can be a cost effective alternative to re-organizing the categorization.

Increase basket size

Fredhopper Suggest Search guides your users by highlighting profitable shortcuts. The average eCommerce customer will see about 30 products during a visit -- compared to 300 in a high street store. Suggest Search enables you to merchandise even before a user executes a search by showing more products and guiding the user based on the first information provided.

Up-sell customers

Aside from keywords and categories, Fredhopper Suggest Search can also bring forward selected products. This is typically used to up-sell customers. When the user searches for playstation, he may for example get a bundle offer containing the console with a game:

Promote white labels

Fredhopper Suggest Search can recommend multiple products for a search term. This can be used to compare a premium label product and a cheaper but higher margin white label product next to each other, e.g. when searching for shirt:

Usage of whitelists or profanity list

Fredhopper Suggest Search provides the opportunity to include or remove/blacklist terms from the suggestions. These can be non-product related terms or "negative" search terms.