Does Data Matching Qualify as Identity Resolution?
By Douglas Wood, Infoglide Senior Vice President
Identity Resolution and Entity Analytics are names given to advanced operational business intelligence tools that allow organizations to glide across disparate – possibly remote – sources of data, with a view to understanding possible matches and non-obvious relationships across that data. In order to accomplish this feat, an identity resolution engine must utilize attribute-specific matching algorithms that take into account variances in structures and spellings of different attributes, AND have the capabilities to point to subtle relationships or clusters within those data sources.
Any true identity resolution engine must not consist simply of a mathematical matching equation, nor must it rely upon data-deterioration processes such as ETL and data warehousing. Math cannot tell a user that Ian is an acceptable nickname for John, nor can simple mathematical equations point to similarities between the colors grey and charcoal. Only an identity resolution engine – not a generic matching tool – has the capabilities required for mission-critical functionality within the enterprise.
Entity analytics tools that require movement and/or cleansing of data are equally ineffective. Once attributes are moved, modified or mitigated, the forensic value is lost forever. True identity resolution engines never destroy data, nor do they assume one equation solves all matching problems.
As a leader in the development and deployment of identity resolution and entity analytics software, we at Infoglide have begun to see an interesting repositioning of simple data-matching tools as “identity resolution” software. While we welcome the growth of our industry with open arms, it’s important to understand the differences between a matching tool versus a robust identity resolution engine for entity analytics and non-obvious relationship analysis.
Subscribe to this blog
January 14th, 2009 at 1:39 pm
Interesting post, Doug.
I got very familiar with D&B’s matching capabilities during the three years I was there, and it always struck me that “matching” was one of those simple words that meant different things to different people.
I think you’re right that the simpler data matching tools are trying to position themselves “up market” as full-fledged identity resolution engines, but there’s a world of difference between a match engine (some of which can have very simple capabilities) and the Infoglide identity resolution engine.
What’s interesting to me now is, when does a simple matching tool (such as the ones built into the typical MDM hubs) suffice, and when do you have to put that aside for the more robust capabilities such as those that Infoglide provides.
One application would be for anything mission critical - a Terrorist Watch List, or a Enterprise Master Patient Index - that type of thing.
But I think another indicator is sheer volume. If a simpler match engine fails to match even 10% of your records, and you’ve got a million records total, then you’ve got to deal (somehow) with 100,000 exceptions. That can be pretty daunting!
So people in the MDM space should be on the lookout for the indicators that a simpler match engine won’t suffice, and be aware of more sophisticated products like Infoglide.
Best regards — Dan
January 20th, 2009 at 11:11 pm
Interesting. So what Identity Resolution consists of is a bunch of data standardization tables and a matching tool? Seems like a name equivalence table and a color equivalence table and any of the off the shelf matching tools would solve your problem. That is a pretty trivial solution to a pretty complicated problem. Thanks for you insight
January 21st, 2009 at 2:08 pm
There’s too much to say here, so I’m going to respond to this in another post.