Step 1 - Specify your Suggest requirements

This page enables you to provide Fredhopper with all required information to create a tailor-made Suggest index for you.

Step 1.1: Define types of suggestions

Fredhopper Suggest Search can provide suggestions for:

• Search terms, e.g. suggest the most relevant search terms that start with son;
• Categories, e.g. suggest the most relevant categories that start with wom;
• Products, e.g. suggest products that start with ad;
• Other types of content, e.g. suggestions for videos, blogs, countries, themes or styles.

For concrete examples see the page Understand Fredhopper Suggest Search.

Please inform your account manager about your requested indexes based on your Suggest designs.

Step 1.2: Define data sources

Fredhopper Suggest Search can use both product and behavioral data. For each type of suggestions defined in Step 1.1 above, please specify the data source.

To suggest products/categories, please provide the standard Fredhopper-XML, which contains product/categorization data.

To suggest search terms please provide behavioral reports containing the following fields:

• Search term, e.g. shoes;
• Frequency, e.g. 523;
• (optional) Number of search results, e.g. 382;
• (optional) A Fredhopper location defining the scope to which the search should apply, e.g. //catalog01/en_GB.

  If you want to use a search term in multiple scopes you must specify it for every such scope .

If you wish to suggest other types of content, please discuss this with Fredhopper directly and ensure you can deliver the requisite data in CSV format to the Fredhopper Managed Service. The report should contain at least the top 10,000 search terms, but preferably at least 100,000. Every Web Analytics system, like Google Analytics or Coremetrics, can provide this report.

Step 1.3: Define business rules

Fredhopper Suggest Search can use various business rules for sorting and filtering the suggested information. For every section defined in Step 1.1, please deliver all relevant business rules. The tables below list some example business rules:

• Business Rules for keyword suggestions
  At most 10 keyword suggestions should be shown.
  Keywords are sorted first by frequency, then lexicographically.
  Keywords can be blacklisted.
  Keywords can be whitelisted.
  Synonyms should be taken into consideration, keywords with highest frequency of a synonym always win.
  Non-catalogue redirect keywords should always have 1 as the number of results.
  Catalogue redirect keywords should always list the number of results as if these keywords are not redirects.
  Singular and plural should never both be suggested, the version with the highest search frequency should always be used.

• Business Rules for product suggestions
  The suggest should show max 3 product suggestions.
  Products that are suggested are never more than seven days old.
  Products are sorted by Ranking Cocktail popularity.
  Product suggestions should be shown based on the title of the product.

Step 1.4: Define required response information

For each of the type of suggestion defined in Step 1.1, please specify the data you require back from the Fredhopper Suggest service, so that you are able to properly display the information. See the tables below for an example specification:

• Fields required for keyword suggestions:

  Field	Comments
  searchterm	The keyword entered by the user
  nrResults	The number of results returned if the user were to search for that product

• Fields required for product suggestions:

  Field	Comments
  Name	The product name
  Brand	The product's brand
  Secondid	The product's ID
  Price	The product's price
  Image	A small image (thumbnail) of the product