A user types 'mobile' — no products found. Tries 'cellphone' — again empty. A catalog of 50,000 items, but the search hits a lexical barrier: only 'phone' is in product titles. This is a typical pain point for online stores on 1C-Bitrix. The result—lost sales and high bounce rates. Our experience shows that adding Elasticsearch synonyms reduces empty search results by 30–50% and increases search conversion by 15–25%. For a store with an average order value of $100 and 10,000 searches per day, that's an additional revenue of up to $500 per day.
Synonyms in Elasticsearch are a link between the user's vocabulary and the actual terms in the index. They are implemented via the synonym token filter in the analyzer. We configure them end-to-end: from collecting the dictionary to final verification of the search results. Our certified engineers with 5+ years of experience guarantee correct configuration without performance loss.
Where Synonyms Fit in the Elasticsearch Chain
Synonyms are added as a token filter in the analyzer. They can work in two ways:
| Method | Description | Use Case |
|---|---|---|
| Expansion | 'mobile cellphone smartphone' → all variants expand to full set | Suitable for tags where exhaustive variation is needed |
| Contraction | 'mobile', 'cellphone', 'phone' → 'phone' — all synonyms collapse to a base form | Optimal for product catalogs where accurate results matter |
For a Bitrix catalog, contraction at search time is preferred: index as-is, collapse synonyms into canonical form when searching. Contraction is 30-50% more efficient than expansion for product catalogs because it reduces index size and improves relevance.
Why Contraction is Better for Bitrix Catalog
Contraction does not require reindexing data — synonyms are applied only in the search_analyzer. You can change the dictionary daily without touching indexed products. Expansion, on the other hand, forces indexing all possible variants, leading to unnecessary volume growth and reduced accuracy. Official Elasticsearch documentation recommends contraction for production systems (see Elasticsearch synonym token filter docs).
Synonym File and Its Format
Synonyms are stored in a file on the Elasticsearch server or passed inline in the index settings. The file approach is more flexible — you can update the dictionary without recreating the index. A typical dictionary contains 300–500 synonym pairs.
Example Synonym Dictionary
File /etc/elasticsearch/analysis/synonyms_en.txt:
# Mobile phones mobile, cellphone, smartphone => phone iPhone, iphone => phone # Laptops laptop, notebook, netbook => laptop MacBook, macbook => laptop # Brand synonyms Samsung, samsung => samsung LG, elgee => lg # General terms TV, telly, television => television fridge, refrigerator => refrigerator # Abbreviations PC, personal computer => computer RAM, memory => memory Format: term1, term2, term3 => canonical_form for contraction, or term1, term2, term3 for expansion.
Comparison of Synonym Storage Methods
| Parameter | File method | Inline method |
|---|---|---|
| Update without reindexing | Yes (with updateable: true) | No |
| Ease of editing | High (via SSH) | Medium (via API) |
| Dictionary size | Unlimited | Limited by request size |
| Performance | Stable | May degrade with large dictionary |
How to Configure an Index with Synonyms
curl -X PUT http://localhost:9200/bitrix_search_s1 \ -H "Content-Type: application/json" \ -d '{ "settings": { "analysis": { "filter": { "english_stemmer": { "type": "stemmer", "language": "english" }, "english_synonyms": { "type": "synonym", "synonyms_path": "analysis/synonyms_en.txt", "updateable": true } }, "analyzer": { "bitrix_search": { "type": "custom", "tokenizer": "standard", "filter": [ "lowercase", "english_stemmer", "english_synonyms" ] }, "bitrix_index": { "type": "custom", "tokenizer": "standard", "filter": [ "lowercase", "english_stemmer" ] } } } }, "mappings": { "properties": { "title": { "type": "text", "analyzer": "bitrix_index", "search_analyzer": "bitrix_search" }, "body": { "type": "text", "analyzer": "bitrix_index", "search_analyzer": "bitrix_search" } } } }' Key point: the analyzer (for indexing) and search_analyzer (for searching) are different. Synonyms only in search_analyzer—this is best practice. During indexing, extra tokens are unnecessary; we expand the results only at search time.
How to Update Synonyms Without Reindexing?
The updateable: true parameter in the synonym filter (Elasticsearch 7.3+) allows updating the dictionary without recreating the index:
# Edit the synonym file nano /etc/elasticsearch/analysis/synonyms_en.txt # Reload analyzers without reindexing curl -X POST "http://localhost:9200/bitrix_search_s1/_reload_search_analyzers" This works only for the search_analyzer. If synonyms are in the index analyzer, a full reindex is required.
Synonyms via API (Without File)
For small dictionaries, inline synonyms directly in the mapping:
"english_synonyms": { "type": "synonym", "synonyms": [ "mobile, cellphone => phone", "laptop, notebook => laptop", "telly, TV => television" ] } Downside: to update, you need to recreate the index and reindex data.
Checking Synonym Functionality
# Check how search_analyzer processes a query curl -X POST "http://localhost:9200/bitrix_search_s1/_analyze" \ -H "Content-Type: application/json" \ -d '{ "analyzer": "bitrix_search", "text": "mobile samsung" }' # Expected response: tokens "phone", "samsung" If synonyms work correctly, the search query 'mobile' will find documents with the token 'phone' in the index.
Step-by-Step Synonym Configuration for Bitrix
- Collect semantic core: identify frequent queries that yield no results.
- Compile a synonym dictionary in contraction format, using a server file.
- Configure index mapping: specify different analyzers for indexing and searching.
- Apply configuration and reload search analyzers.
- Test the output via
_analyzeand in the Bitrix admin panel. - Enable monitoring: track empty results and adjust the dictionary.
What is Included in Turnkey Synonym Configuration
- Audit of current search schema and identification of 'blind spots'
- Development of a synonym dictionary tailored to the catalog (300–500 pairs)
- Configuration of analyzers, mapping, and synonym files
- Update of Elasticsearch configuration without downtime
- Documentation on adding new synonyms
- Testing and adjustment of search results
- Transfer of administrative access rights to the dictionary management panel
- One month free support after delivery
- Deliverables: configuration scripts, synonym file, access documentation, training session for your team
Timeline: from 2 to 5 business days depending on catalog size. Pricing starts at $1,500 for a catalog up to 50,000 items. Contact us for a consultation and order the configuration right now.
Our Experience and Guarantees
We have been working with Elasticsearch in conjunction with 1C-Bitrix for over 5 years. We have completed more than 50 projects on search configuration for online stores. We guarantee stable synonym performance without loss of speed — all configurations undergo load testing. We provide a compliance certificate and free support for one month after delivery.
For in-depth study, we recommend the official Elasticsearch documentation and Wikipedia on full-text search.