Introduction to synonyms
This guide will explain what synonyms are, but understanding how synonyms function within search is equally important. Below will explain everything you need to know about synonyms. Search configuration influences synonyms; you may need to speak to a power user of search within your team when troubleshooting synonyms.
Prerequisites
- You need to have permissions in Fredhopper's Merchandising Studio to create a synonym
- Fredhopper's Merchandising Studio is integrated successfully and search has been configured
- Basic understanding of Fredhopper Search including search index, passes, redirects and modified queries
- Access to onsite search reports and most frequent search queries within the onsite search
- A good understanding of your customer's search behaviour will help you define which synonyms you need to put in place
- Use good data regarding actual customer behaviour not assumed customer behaviour on your site
- Good understanding of your product catalogue data
Overview
Key information about this feature:
| Purpose of this feature | Provides merchandisers with the functionality to ensure search results return relevant products. |
| Menu name in Fredhopper's Merchandising Studio | Synonyms |
Synonyms are set up in the Fredhopper Merchandising Studio to help return relevant products when a shopper performs a search, without the merchandisers having to change all the product data in general. An example of this is where a search for "tv" has been executed, products containing "tv" and "television" should be returned.
Another example could be where 'women's underwear' is categorised as 'lingerie' within the product data, so when a customer searches for "women's underwear", they should be returned all lingerie products i.e. "women's underwear = lingerie".
Therefore, synonyms can be defined as different words with identical or at least related meanings.
Getting synonyms right will improve search recall and therefore lead to more purchases!
In the context of Fredhopper, the following types of operator synonyms exist:
Operator |
Synonym Syntax |
Description |
|---|---|---|
| Contains | > |
Additional words should only be linked in one direction: "Clothing > Tops, Shirts, Skirts, Coats, Jackets, Dresses" this means that if "clothing" was the search term used by a shopper, the results would include; Tops, Shirts, Skirts, Coats, Jackets and Dresses. The term on the right has half the match rate of the term on the left: Clothing (1) > Tops (0.5), Shirts (0.5), Skirts (0.5), Coats (0.5), Jackets (0.5), Dresses (0.5) |
| Equals | = |
Both terms effectively mean the same thing, e.g. tv and television; therefore the synonym created is "tv = television", and therefore have an equal match rate: tv (1) = television (1) |
| Similar | ~ |
The first term is more relevant than the second, but all relevant products should still be returned, e.g. "Jumper ~ Sweatshirt". If the shopper searched "jumper" Fredhopper would return products for jumpers and sweatshirts, in that order. If the shopper searched for sweatshirts, then the relevant sweatshirts and jumpers would be returned, in that order. The term on the right receives a significantly lower match rate score (0.01), so will feature at the bottom of the result set, unless the match rate is blended: Jumper (1) ~ Sweatshirt (0.1) |
The majority of good synonyms exist to take inconsistent product naming and referencing into account. They can help merchandisers to return both sets of products (e.g. tv and television) to customers without having to change all the product titles.
Types of synonyms
Multiple words can be connected with the same synonym operator within one synonym, e.g. tv is identical to television and tele, therefore, you could create "tv = television = tele" or "automobile, truck > car parts > wheel, seat cover". It is not possible to mix different operators within one synonym, e.g. "auto = automobile > car parts".
Synonyms of one definition can be separated by commas: pants > jeans, shorts, bermudas, 501. The words separated by a comma are not related to each other.
Each synonym stands for itself and is not transitive, e.g. if auto is similar to automatic and auto is similar to car, it does not mean that automatic is similar to car.
The following table shows how to translate common business cases into synonym syntax:
Business case |
Synonym Syntax |
Search results |
|---|---|---|
| auto is identical to automobile; both are super concepts of car parts |
auto = automobile ------------- also create: auto, automobile > car parts |
Products with auto and automobile would return the same results if either auto or automobile were the search query
If auto were to be the search query then, products for auto, automobile and car parts would be returned, in that order. If automobile were to be the search query then products with automobile and car parts would be returned in that order. If car parts were to be the search query, then products with car parts in the data would be returned.
|
| city bike is similar to a mountain bike | city bike ~ mountain bike | Products with city bike and mountain bike in the data |
| every 501 and 511 is a jeans is a trousers (taxonomy) | trousers > jeans > 501, 511 | If trousers were to be the search query, then all products with trousers and jeans and 501 and 511 would be returned in that order. If jeans were to be the search query, then all products with jeans and 501 and 511 would be returned in that order. If 501 were to be the search query, only 501 will be returned, and if 511 were to be the search query, only 511 will be returned. |
| every jeans is a trousers (but not all trousers are jeans) | trousers > jeans | Products with trousers and jeans in the data. If trousers were to be the search query, then all products with trousers and jeans would be returned in that order. If jeans were to be the search query, then all products with jeans and trousers would be returned in that order. |
| tv is identical to television | tv = television | Products with tv and television in the data |
| tv is identical to television and tele | tv = television = tele | Products with tv and television and tele in the data |
How search works, the basics
Synonyms work within the search functionality. You do not need to set up synonyms for every conceivable search query, as Fredhopper will follow the process below when a search is executed. Setting up synonyms will help guide shoppers to the most useful products when required. To understand synonyms, it is best to explain where they fit into the search process within Fredhopper.
Let's say a search query for a leading brand of "Red shoes by Chloé" is executed on a site. Fredhopper carries out the following:
- Linguistic processing includes stemming & tokenizing
- Stemming ensures that search has the purest form of the word or search phrase to match against
- Forces lowercase = "red shoes by chloé"
- Removes or transforms special characters = "red shoes by chloe"
- Identifies words= "red","shoes","by","chloe"
- Removed long words = "red","shoes","by","chloe"
- Removes stop words (e.g. by, the, and, that, are, this, etc..) = "red","shoes",".","chloe"
- Identifies root words = "red","shoe",".","chloe"
- The number of words that result from stemming have been 'tokenized'
- Stemming ensures that search has the purest form of the word or search phrase to match against
- Next Fredhopper looks to see if any synonyms are in place for this search query
- Is there a redirect in place for this search query? If yes, then the shopper will be taken to the redirect. If no, the process continues:
- The search query then goes through a pass and search indexing process
- Search index is a body of structured data that search refers to when looking for results that are relevant to a specific search query. Indexes are a critical piece of search since they must be tailored to the specific information retrieval method of the search's algorithm.
- Finally, the match rate, ranking rules, result modifications and facets are applied. Click for more information ranking rules and ranking cocktails.
The match rate will depend on the search pass. The match rate is calculated by adding the match rate of each of the terms in the query.
These checks result in the product lister page being displayed with the relevant products.
Things to remember
- Having two rules; "purple = violet" and, "violet = lavender" does not imply purple and lavender are the same
- Synonyms can combine multiple words
- "purple, violet > lavender" reads purple and violet are generalisations of lavender, however, purple and violet are not the same
- Synonyms are not necessary for:
- Singular/plural i.e. "dress = dresses"
- Misspellings i.e "dresses = dreses"
- Refinements i.e "shirts = dress shirts"
- Special characters i.e. "big & tall = big tall"
- Different word order i.e. "pants linen = linen pants"
- Stop words i.e. "shirts for men = shirts men"
- As these are dealt with via the linguistic analysers
Best practice
- Check your site or Preview if there are any products missing in a search result before setting up any synonyms
- Have good knowledge of your product data, which makes it easier to predict how a search will influence search results
- Have good knowledge of your shoppers search queries, what shoppers expect to find, and the products that are actually returned when a search is executed
- Don't create too many synonyms
- A good synonym can help bring the most relevant products to the top of a page, while an unnecessary synonym can give undeserved emphasis on less relevant results
- Check synonyms in Preview before publishing
Are you ready to create a synonym? Click here to open a step by step guide.
Your feedback is important to us. Please take a moment to let us know if you found this article helpful by clicking "Yes" or "No" in the "Was this article helpful?" section below or by adding a comment.
Comments
0 comments
Please sign in to leave a comment.