Fuzzy matching used to match company names

16 March 2015

Matching company names is indeed a serious issue. It is seen that most businessmen often come up with the question how to fuzzy match company names. Numerous businessmen are there who suffer from this same kind of problem. Linking the company informations together from a disparate dataset is also important. This is mainly to provide a perfect 360 degree view of any customer. This company matching job often requires the integration of the financial sales history either from SAP or Oracle. The CRM data gathered from SalesForce or Siebel is often matched with the information taken from the external sources like Bradsheet and Dun. This in turn helps in the creation of insightful business analysis as well as helps to buy propensity models.

The definition is different in this issue of fuzzy matching. So, let’s check out what is termed as successful. Sometimes it is seen that someone has matched 80% of the companies, but the remaining 20% are the most important ones. It is seen that those 20% unmatched data is a company’s biggest customers and are also responsible for 80% of the revenue generator. At this point of time assessing the importance of this remaining unmatched data is very important. It is also important to determine if this unmatched data contains some prospects or customers who are strategically important for the business.

Matching the data with some external data sources helps in bringing information that will help the business to find out the significance of the prospects as well as the customers. By checking the number of employees, the subsidiaries the match making can help you in prioritizing your efforts. There are some steps that must be followed in order to fuzzy match account and company. The steps that are to be followed in a synchronized fashion are

Data standardization- this helps in addressing issues that are with common abbreviations like Incorporated, Inc, Corporation, Corp, Limited, Ltd etc. This step does not advocate by removing the legal entity information unlike other matching systems.

Probabilistic logic- Probabilistic logic is meant for examining a company name and determining which of the elements are the most relevant one for matching. The keywords that are more important for matching are considered in general.

Fuzzy logic- this is the most important part of a fuzzy matching. Fuzzy logic is combined with Probabilistic logic and the probability of the matching data is ascertained in a much better way. It becomes a more powerful tool after combining with Probabilistic logic as coming together it decreases the number of false matches which is more provided by the fuzzy logic.

Other parts of tools of this matching procedure are an extensive knowledge base, leveraging of corroborative information, iterative approach of matching and powerful visualization. Having sound knowledge is the most important thing to carry out the entire match making process without any fault. The last tool that is the execution of an interface that is powerful and user friendly is the actual game changer.