Are Name Matching and Entity Resolution Equivalent? No.
Wednesday, April 30th, 2008By Robert Barker, Infoglide Senior Vice President & Chief Marketing Officer
The rumble about Informatica’s acquisition of Identity Systems continues. Vincent McBurney’s recent post from Australia illustrates the global impact of such industry events and makes some interesting points that deserve a response.
First of all, I take issue with the comment that entity resolution is “a fancy name for name matching.” Name matching is necessary but not sufficient for entity resolution, i.e. entity resolution (aka identity resolution) includes name matching analytics plus quite a bit more.
In fact, we talk about four layers that constitute a complete entity resolution engine - similarity searching (in our parlance), identity resolution/relationship detection, decisioning and rules processing, and business processing. In a recent post, we pointed out that Ovum, Bloor Research, and other bloggers note that the number of vendors possessing a complete solution is small.
Informatica’s move to add Identity Systems’ products to their data quality suite of capabilities looks like a good one, and I’ll bet they quickly integrate it in a way that will benefit their customers. Adding name matching to enhance de-duping, however, does not equal introducing a complete entity resolution solution. Name matching technology has existed for some time. Entity resolution and analysis is a wholly new area of business intelligence (and we think Gartner would back us on this since they recently added it to their hype cycle as a separate category), which has important implications for both government and commercial markets.
