Identity Resolution as a ServiceOften, when people ask what we do, we throw out a few phrases: marketing data science, machine learning and identity resolution. While people generally understand what data science and machine learning are all about, sometimes the phrase Identity Resolution takes some addition explaining.
Simply stated, Identity Resolution (IR) is the data management process of linking different data related to individuals in order to obtain a consistent view. That is, since your name, address, email address and other information about you exists in any number of databases, IR is the process of identifying you as the same person in each of those. This allows marketers to combine information from multiple sources in order to better target consumers.
The Chalenges of IR
This process is simple if there is a consistent identifier for you in each of those databases, such as a social security number. But we usually don't have access to such universal identifiers, so marketers need to rely the more difficult process of IR. And this is not as simple as it may seem for a number of reasons:
- Databases often have partial or incomplete information. This can be especially true with physical addresses.
- People use different names in different situations, such as William vs. Bill.
- Sometimes information is simply wrong due to a data entry error.
- Data standards can be inconsistent. For example, one database might have a concatenated "full name" while another will store "first name" and "last name" values.
- There are many ways you might identify an individual, including any combination of name, address, email address, birth date and even IP addresses and web cookies. And at matchbox.io, we are pioneering new methods in IR with machine learning (ML) and data science.
A number of different techniques and algorithms are used in IR. And recently, more and more sophisticated techniques are being deployed to better identify matches. We here at matchbox.io have been employing Machine Learning (ML) in order to improve our own Identity Resolution process. You can find a great blog on the subject at liliendahl.com.
Identity Resolution Software
IR goes by several names, including Customer Data Integration (CDI) and Master Data Management (MDM), Entity Resolution (ER), but often we simply call it "matching". There are a number of Identity Resolution Software providers, such as Informatica. At Matchbox.io, we offer Identity Resolution as a Service.
- To learn more about different software providers, read our post: 5 Identity Resolution Software Solutions.
- For a more detailed discussion of Identity Resolution, see A framework of identity resolution: evaluating identity attributes and matching algorithms.
- To download a list of IR algorithms we use here at Matchbox.io, see Identity Resolution Algorithms.