LinkedIn's 'Fuzzy Matching'

Options
Leif Parcell
Leif Parcell Posts: 6 6senser
in All Discussions

Overview 

LinkedIn uses a technique called ‘Fuzzy Matching’ which is the ability to compare two pieces of data that are not identical but have some degree of similarity. That means when LinkedIn is best trying to match a company to a search query it sometimes is not completely accurate, and typically tends to favor larger and more well-known companies. For instance, certain searches for Amazon Energy or Amazon Publishing might get mistaken for Amazon, the retail giant. LinkedIn does this to help users increase match rates; it’s a balance between match rate and accuracy they manage as best they can, but that comes with its own challenges. This is a default setting across all of LinkedIn meaning they do not allow users or partners to turn off ‘Fuzzy Matching’.  

Challenge 

Since multiple versions of a name can be used to search for a company on LinkedIn, such as “6sense”, “6sense.com”, 6sense Insights”, “6sense Insights Inc.”, all of which are likely referring to the same company, and with LinkedIn using their Fuzzy Matching technique, this means there are chances a company targeting certain businesses may inadvertently get impressions for the wrong organizations. Typically, this is seen to a higher extent when targeting small to medium-sized businesses and instead getting impressions from larger enterprises that are in some way named similarly.  

6sense Solution 

Currently, when available, 6sense includes an account’s LinkedIn company page URL as an additional identifier to improve targeting accuracy and reduce the use of the fuzzy matching against the business name.  When both name and company page URL are present, we’ve measured a significant reduction in matching errors. However, company page URLs are not available for 100% of all accounts. 

Although there is no perfect solution to this challenge as it is the algorithm LinkedIn uses to match companies, 6sense is continually working with the LinkedIn team to provide more accurate matches and will continue to provide updates as available. 

Categories