# Similar Matching – Definition, FAQs & How Dedupely Helps > Similar matching explained: how to find duplicates with typos, misspelled names and phone numbers saved in different formats, and when to use it in a CRM. - URL: https://dedupe.ly/glossary/similar-matching - Section: glossary - Site index: https://dedupe.ly/llms.txt - Any page: append `.md` to its URL for this plain-text form --- [Glossary](https://dedupe.ly/glossary) Similar Matching ## Similar Matching Similar matching is a record-matching method that pairs values that differ only by small typos or by mixed or missing letters, such as "gmail" and "gmal," or "Johnson" and "Jonhson." It's a common way to find duplicates with typos: records for the same person or company that an exact comparison keeps apart because one character is off. On phone numbers, similar matching ignores formatting and compares the core of the number, so "+1 (555) 010-2000" and "555-010-2000" can pair. That mix of typo tolerance and phone number normalization makes it a natural next step after exact matching, on names, email addresses, company names and phone numbers. Many tools call this kind of comparison fuzzy matching. Because it allows some difference between values, similar matching works best alongside a stronger field that confirms the match. See how Dedupely finds duplicates with typos in your CRM [Try for free](https://v2.app.dedupe.ly/signup) [Talk to sales](https://meetings.dedupe.ly/meetings/gaby-salinas/talk-to-sales) Overview Similar matching pairs records whose values differ only by a small typo or a mixed or missing letter, like "gmail" and "gmal." On phone numbers, it ignores formatting and compares the core of the number. It catches near matches exact matching misses and works best with a stronger field. On this page - What Is Similar Matching? - How Similar Matching Works - Why Similar Matching Matters for CRM Teams - Getting Started With Similar Matching - Key Takeaways: Similar Matching - Frequently Asked Questions About Similar Matching - Related Terms and Concepts See how Dedupely finds duplicates with typos in your CRM [Try for free](https://v2.app.dedupe.ly/signup) [Talk to sales](https://meetings.dedupe.ly/meetings/gaby-salinas/talk-to-sales) ### What Is Similar Matching? Similar matching, also called typo-tolerant matching, compares two values and allows a small amount of difference between them. It's built for the misspelled duplicates that come from fast typing, form fills, imports and manual entry: a dropped letter, two letters swapped or one letter typed in place of another. It works on contacts in HubSpot, Nimble and Google Contacts, accounts and leads in Salesforce, people and organizations in Pipedrive, and leads in Close. Take "Katherine Jonhson, katherine@gmal.com, +1 (555) 010-2000" and "Katherine Johnson, katherine@gmail.com, 555-010-2000." Exact matching on last name, email or phone keeps the two apart, because each value differs by a letter or by formatting. Similar matching on those fields pairs them as near matches. Tools differ in how they measure the difference, and there are two common approaches for text. Some count the letters that have to be added, dropped, swapped or changed to turn one value into the other. Their strength is an allowance that's easy to predict. Others score how alike two values are and pair them above a threshold. Their strength is an allowance that adapts to the length of the value. Phone numbers are normalized first: punctuation and country codes are stripped. Some tools then compare the whole remaining number, which keeps the comparison precise. Others compare its core digits, such as the last seven, which also pairs numbers saved with and without an area code. Resources: - [Match Options](https://dedupe.ly/knowledge-base/match-options) (KB) - [What is a duplicate?](https://dedupe.ly/blog/what-is-a-duplicate) ### How Similar Matching Works Similar matching is the comparison step for one field inside a duplicate search. It works in a few steps: - **Prepare the values.** Differences that don't change the meaning, such as capitalization, are evened out first. On phone numbers, normalization strips parentheses, hyphens, periods and country codes, so only the digits that identify the number are compared. - **Measure the difference.** The two values are checked for how far apart they are: a missing letter ("Jonson"), an extra one ("Johnsson"), two letters swapped ("Jonhson") or one letter changed ("Johnsen"). - **Apply the allowance.** Values within the allowed difference count as a match. Short values need extra care, because one changed letter can turn one real name into another, such as "Dan" and "Don." - **Combine with other fields.** Similar matching on one field is often paired with exact matching on another, such as similar on last name and exact on email, so a near match on one value doesn't decide the pair alone. - **Group and review.** Records that match are grouped into duplicate sets, and a sample is checked before anything is merged. The result sits between exact matching and looser methods. Similar matching catches spelling slips and formatting differences, but names that sound alike and are spelled quite differently, such as "Catherine" and "Kathryn," or company names with their words in another order, are often beyond it. Those need phonetic or word-based matching. Resources: - [Finding Duplicates](https://dedupe.ly/knowledge-base/find-duplicates) (KB) ### Why Similar Matching Matters for CRM Teams Misspelled duplicates are the ones exact rules can't see, which is why CRM teams need a way to find duplicates with typos. A contact entered once as "katherine@gmail.com" and once as "katherine@gmal.com" stays as two records: it inflates contact counts and pipeline reports, splits activity history and, in CRMs that bill by record count, gets paid for twice. Two reps can reach out to the same customer without knowing it. Phone numbers raise the same problem in a different form. Web forms, imports, dialers and integrations each save numbers their own way, with or without country codes, dashes or parentheses. Without normalization, the phone field can't confirm a match, so one of the strongest identifiers on a record goes unused. That matters most for leads and contacts with no email address, where the phone number may be the only reliable link between two records. Tolerance has a cost, too. A comparison that allows too much difference pairs records that only look alike, such as "Dan Lee" and "Don Lee," and merging that set mixes two customers' histories, owners and contact details. In many systems a merge is hard or impossible to reverse. Getting the balance right decides whether similar matching cleans the data or adds new errors to it. Resources: - [Why your CRM database is turning into a mess (and how to fix it)](https://dedupe.ly/blog/why-your-crm-database-is-turning-into-a-mess) ### Getting Started With Similar Matching Start with exact matching on your strongest fields and merge those duplicates first. What's left includes the near matches: names with a slip of the keyboard, email domains with a missing letter and phone numbers in different formats. Those are the records similar matching is meant for. Loosen one field at a time and keep at least one strong field strict. Similar matching on last name with exact matching on email, or similar matching on phone with exact matching on last name, catches typos without pairing strangers. Check a sample of each new set of results before merging, and reject the pairs that turn out to be two different records. Look closely at short values and shared values. One letter matters more in a short first name than in a long company name, and placeholder values, such as "N/A" or a dummy phone number, can match across many unrelated records. Leaving those out of the comparison keeps the results clean. In Dedupely, Similar is the Match Option for this approach: it catches minor typos and mixed or missing letters, and on phone numbers it strips formatting and country codes and compares the last seven digits. Typo allowance, one of a Duplicate Search's Enhancements, keeps small differences from blocking matches when stricter Match Options are used. Resources: - [How to Approach CRM Deduplication](https://dedupe.ly/knowledge-base/how-to-approach-crm-deduplication) (KB) - [Duplicate Searches](https://dedupe.ly/knowledge-base/search-pads) (KB) ### Key Takeaways: Similar Matching Similar matching pairs values that differ only by small typos or by mixed or missing letters, and on phone numbers it ignores formatting and compares the core of the number. It finds the misspelled duplicates and phone format differences that exact matching misses. Because it allows some difference, it works best backed by a strong field and checked on short or shared values. In Dedupely, Similar is one of the Match Options in a Duplicate Search, and on phone numbers it compares the last seven digits. Typo allowance, one of the search's Enhancements, applies to the whole match and also helps stricter Match Options. Rejected Matches keeps false pairs out of future results, and Auto Match and Auto Merge keep trusted searches running. Resources: - [Dedupely product overview](https://dedupe.ly/product) ### Frequently Asked Questions About Similar Matching #### How do you find duplicates with typos in names and email addresses? Compare those fields with a typo-tolerant method instead of an exact one, and back it with a second field. Typo-tolerant matching pairs values that differ by a missing, extra, swapped or changed letter, such as "Jonhson" and "Johnson," or "gmal.com" and "gmail.com." On its own, a near match on a name can pair different people, so combine it with a field that confirms the match, such as an exact email, a phone number or a company name. Start with the exact matches, merge them, and then run the typo-tolerant comparison on what's left, so the new results are easier to check. In Dedupely, set the field to the Similar Match Option in a Duplicate Search. It catches minor typos and mixed or missing letters. You can also turn on Typo allowance, one of the search's Enhancements, so small differences don't block a match when stricter Match Options are used. #### How do you match phone numbers saved in different formats? Normalize the numbers before comparing them: strip the punctuation and country codes, and compare only the digits that identify the number. "+1 (555) 010-2000," "555.010.2000" and "555-010-2000" are three different values to an exact comparison, but the same number once the formatting is gone. Some tools compare the whole remaining number, and others compare its core digits, which also pairs numbers saved with and without an area code. Phone matching is strongest when combined with a name field, because a shared office line or a reused number can belong to several people. In Dedupely, the Similar Match Option strips formatting and country codes from phone numbers and compares the last seven digits. Each field also has a "Send incorrect formats to manual review" checkbox, off by default. When it's checked, phone numbers in the wrong format and filler values, such as blank or N/A, go to Manual Review automatically. #### When should you use similar matching instead of exact matching? Use similar matching once the exact matches are merged and duplicates you can see are still left. Exact matching finds the clearest duplicates with the fewest mistakes, so it's the right first pass. The duplicates it leaves behind often differ by a typo or by formatting, and that's where similar matching helps. Switch only the field that breaks the match, such as last name or phone, and keep a strong field like email on exact matching, so false matches stay low. Then check a sample of the new sets before merging any. In Dedupely, each field in a Duplicate Search gets its own Match Option, so you can move one field from Exact to Similar and leave the rest. You can clone a search to try the change without touching the active one, then click "Scan" to update the results. If a search compares three fields, all three must meet their criteria to flag a duplicate. #### Why does typo-tolerant matching pair records that aren't duplicates? Because some different values are only a letter or two apart. Short names like "Dan" and "Don," or two different company names a letter apart, can fall within the allowed difference, and placeholder values such as "N/A" can match across many unrelated records. The risk grows when a typo-tolerant field is the only field in the rule. To keep false matches low, pair the loose field with a strong one, such as email or phone, leave filler values out of the comparison, and check a sample of results before merging. Reject the pairs that turn out to be two different records, so they don't come back. In Dedupely, Rejected Matches removes a set from current and future Duplicate Searches. Safety Settings are on by default and hold back matches with more than 20 records or built on poor or example data, such as John Doe, in an Excluded tab. Up to 50 Ignored Terms per field keep common words out of the comparison. #### How do you catch misspelled duplicates in a CRM such as HubSpot, Salesforce or Pipedrive? Run a duplicate search that compares names, emails and phone numbers with a typo-tolerant method, not only an exact one. Built-in duplicate checks differ from one CRM to another, in which fields they compare and how strictly, so the CRM's own documentation is the place to confirm what yours catches. Records that differ by a letter or by phone formatting often need a dedicated search. Choose the fields that identify a record, loosen the ones where typos appear, and keep one strong field strict. The same approach works in Close, Nimble, Google Contacts and in a spreadsheet. In Dedupely, you connect your CRM or upload a CSV, and each field in a Duplicate Search gets its own Match Option, including Similar for typos and phone formats. Matching works the same way for spreadsheets and CRM records. The KB recommends combining 2 to 4 key fields, and merges can't be undone, so check the results before merging. ### Related Terms and Concepts #### [Exact matching](https://dedupe.ly/glossary/exact-matching) Exact matching is the strict first pass that similar matching follows. Many searches use both: exact matching on the strongest field and similar matching on the field where typos appear. #### [Fuzzy matching](https://dedupe.ly/glossary/fuzzy-matching) Fuzzy matching pairs names that sound alike, such as "Sean" and "Shawn," while similar matching handles spelling slips and formatting. The two often run as separate passes on name fields. #### [Similar Words matching](https://dedupe.ly/glossary/similar-words-matching) Similar Words matching compares the words inside a value rather than its letters. It covers company names with words in another order, which similar matching often misses. #### [Multi-field matching](https://dedupe.ly/glossary/multi-field-matching) Multi-field matching combines several fields in one rule. It's what makes similar matching safe to use, because a strong field confirms each near match. #### [False matches](https://dedupe.ly/glossary/false-matches) False matches are records paired as duplicates that aren't. Similar matching produces more of them than exact matching, especially on short names and placeholder values. #### [Matching duplicate records](https://dedupe.ly/glossary/matching-duplicate-records) Matching duplicate records is the stage of deduplication where records are compared. Similar matching is one of its methods, and a common second pass after exact matching. Skip to content